US20070277021A1 - Instruction folding for a stack-based machine - Google Patents

Instruction folding for a stack-based machine Download PDF

Info

Publication number
US20070277021A1
US20070277021A1 US11/461,340 US46134006A US2007277021A1 US 20070277021 A1 US20070277021 A1 US 20070277021A1 US 46134006 A US46134006 A US 46134006A US 2007277021 A1 US2007277021 A1 US 2007277021A1
Authority
US
United States
Prior art keywords
stack
instruction
cache
operand
index
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Abandoned
Application number
US11/461,340
Inventor
James O'Connor
Marc Tremblay
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Sun Microsystems Inc
Original Assignee
Sun Microsystems Inc
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Sun Microsystems Inc filed Critical Sun Microsystems Inc
Priority to US11/461,340 priority Critical patent/US20070277021A1/en
Publication of US20070277021A1 publication Critical patent/US20070277021A1/en
Abandoned legal-status Critical Current

Links

Images

Classifications

    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06FELECTRIC DIGITAL DATA PROCESSING
    • G06F9/00Arrangements for program control, e.g. control units
    • G06F9/06Arrangements for program control, e.g. control units using stored programs, i.e. using an internal store of processing equipment to receive or retain programs
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06FELECTRIC DIGITAL DATA PROCESSING
    • G06F15/00Digital computers in general; Data processing equipment in general
    • G06F15/76Architectures of general purpose stored program computers
    • G06F15/78Architectures of general purpose stored program computers comprising a single central processing unit
    • G06F15/7839Architectures of general purpose stored program computers comprising a single central processing unit with memory
    • G06F15/7842Architectures of general purpose stored program computers comprising a single central processing unit with memory on one IC chip (single chip microcontrollers)
    • G06F15/7846On-chip cache and off-chip main memory
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06FELECTRIC DIGITAL DATA PROCESSING
    • G06F12/00Accessing, addressing or allocating within memory systems or architectures
    • G06F12/02Addressing or allocation; Relocation
    • G06F12/08Addressing or allocation; Relocation in hierarchically structured memory systems, e.g. virtual memory systems
    • G06F12/0802Addressing of a memory level in which the access to the desired data or data block requires associative addressing means, e.g. caches
    • G06F12/0875Addressing of a memory level in which the access to the desired data or data block requires associative addressing means, e.g. caches with dedicated cache, e.g. instruction or stack
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06FELECTRIC DIGITAL DATA PROCESSING
    • G06F9/00Arrangements for program control, e.g. control units
    • G06F9/06Arrangements for program control, e.g. control units using stored programs, i.e. using an internal store of processing equipment to receive or retain programs
    • G06F9/22Microcontrol or microprogram arrangements
    • G06F9/26Address formation of the next micro-instruction ; Microprogram storage or retrieval arrangements
    • G06F9/262Arrangements for next microinstruction selection
    • G06F9/264Microinstruction selection based on results of processing
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06FELECTRIC DIGITAL DATA PROCESSING
    • G06F9/00Arrangements for program control, e.g. control units
    • G06F9/06Arrangements for program control, e.g. control units using stored programs, i.e. using an internal store of processing equipment to receive or retain programs
    • G06F9/30Arrangements for executing machine instructions, e.g. instruction decode
    • G06F9/30003Arrangements for executing specific machine instructions
    • G06F9/30007Arrangements for executing specific machine instructions to perform operations on data operands
    • G06F9/30021Compare instructions, e.g. Greater-Than, Equal-To, MINMAX
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06FELECTRIC DIGITAL DATA PROCESSING
    • G06F9/00Arrangements for program control, e.g. control units
    • G06F9/06Arrangements for program control, e.g. control units using stored programs, i.e. using an internal store of processing equipment to receive or retain programs
    • G06F9/30Arrangements for executing machine instructions, e.g. instruction decode
    • G06F9/30098Register arrangements
    • G06F9/3012Organisation of register space, e.g. banked or distributed register file
    • G06F9/30134Register stacks; shift registers
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06FELECTRIC DIGITAL DATA PROCESSING
    • G06F9/00Arrangements for program control, e.g. control units
    • G06F9/06Arrangements for program control, e.g. control units using stored programs, i.e. using an internal store of processing equipment to receive or retain programs
    • G06F9/30Arrangements for executing machine instructions, e.g. instruction decode
    • G06F9/30145Instruction analysis, e.g. decoding, instruction word fields
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06FELECTRIC DIGITAL DATA PROCESSING
    • G06F9/00Arrangements for program control, e.g. control units
    • G06F9/06Arrangements for program control, e.g. control units using stored programs, i.e. using an internal store of processing equipment to receive or retain programs
    • G06F9/30Arrangements for executing machine instructions, e.g. instruction decode
    • G06F9/30145Instruction analysis, e.g. decoding, instruction word fields
    • G06F9/3016Decoding the operand specifier, e.g. specifier format
    • G06F9/30167Decoding the operand specifier, e.g. specifier format of immediate specifier, e.g. constants
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06FELECTRIC DIGITAL DATA PROCESSING
    • G06F9/00Arrangements for program control, e.g. control units
    • G06F9/06Arrangements for program control, e.g. control units using stored programs, i.e. using an internal store of processing equipment to receive or retain programs
    • G06F9/30Arrangements for executing machine instructions, e.g. instruction decode
    • G06F9/3017Runtime instruction translation, e.g. macros
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06FELECTRIC DIGITAL DATA PROCESSING
    • G06F9/00Arrangements for program control, e.g. control units
    • G06F9/06Arrangements for program control, e.g. control units using stored programs, i.e. using an internal store of processing equipment to receive or retain programs
    • G06F9/30Arrangements for executing machine instructions, e.g. instruction decode
    • G06F9/3017Runtime instruction translation, e.g. macros
    • G06F9/30174Runtime instruction translation, e.g. macros for non-native instruction set, e.g. Javabyte, legacy code
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06FELECTRIC DIGITAL DATA PROCESSING
    • G06F9/00Arrangements for program control, e.g. control units
    • G06F9/06Arrangements for program control, e.g. control units using stored programs, i.e. using an internal store of processing equipment to receive or retain programs
    • G06F9/30Arrangements for executing machine instructions, e.g. instruction decode
    • G06F9/30181Instruction operation extension or modification
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06FELECTRIC DIGITAL DATA PROCESSING
    • G06F9/00Arrangements for program control, e.g. control units
    • G06F9/06Arrangements for program control, e.g. control units using stored programs, i.e. using an internal store of processing equipment to receive or retain programs
    • G06F9/30Arrangements for executing machine instructions, e.g. instruction decode
    • G06F9/34Addressing or accessing the instruction operand or the result ; Formation of operand address; Addressing modes
    • G06F9/345Addressing or accessing the instruction operand or the result ; Formation of operand address; Addressing modes of multiple operands or results
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06FELECTRIC DIGITAL DATA PROCESSING
    • G06F9/00Arrangements for program control, e.g. control units
    • G06F9/06Arrangements for program control, e.g. control units using stored programs, i.e. using an internal store of processing equipment to receive or retain programs
    • G06F9/30Arrangements for executing machine instructions, e.g. instruction decode
    • G06F9/38Concurrent instruction execution, e.g. pipeline, look ahead
    • G06F9/3802Instruction prefetching
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06FELECTRIC DIGITAL DATA PROCESSING
    • G06F9/00Arrangements for program control, e.g. control units
    • G06F9/06Arrangements for program control, e.g. control units using stored programs, i.e. using an internal store of processing equipment to receive or retain programs
    • G06F9/44Arrangements for executing specific programs
    • G06F9/448Execution paradigms, e.g. implementations of programming paradigms
    • G06F9/4488Object-oriented
    • G06F9/449Object-oriented method invocation or resolution
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06FELECTRIC DIGITAL DATA PROCESSING
    • G06F9/00Arrangements for program control, e.g. control units
    • G06F9/06Arrangements for program control, e.g. control units using stored programs, i.e. using an internal store of processing equipment to receive or retain programs
    • G06F9/44Arrangements for executing specific programs
    • G06F9/455Emulation; Interpretation; Software simulation, e.g. virtualisation or emulation of application or operating system execution engines
    • G06F9/45504Abstract machines for programme code execution, e.g. Java virtual machine [JVM], interpreters, emulators
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06FELECTRIC DIGITAL DATA PROCESSING
    • G06F2212/00Indexing scheme relating to accessing, addressing or allocation within memory systems or architectures
    • G06F2212/45Caching of specific data in cache memory
    • G06F2212/451Stack data
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06FELECTRIC DIGITAL DATA PROCESSING
    • G06F9/00Arrangements for program control, e.g. control units
    • G06F9/06Arrangements for program control, e.g. control units using stored programs, i.e. using an internal store of processing equipment to receive or retain programs
    • G06F9/44Arrangements for executing specific programs
    • G06F9/455Emulation; Interpretation; Software simulation, e.g. virtualisation or emulation of application or operating system execution engines
    • G06F9/45504Abstract machines for programme code execution, e.g. Java virtual machine [JVM], interpreters, emulators
    • G06F9/45516Runtime code conversion or optimisation

Definitions

  • the present invention relates to instruction decoders for a stack machine, and in particular, to methods and apparati for folding a sequence of multiple instructions into a single folded operation.
  • intranet In addition, to the public carrier network or Internet, many corporations and other businesses are shifting their internal information systems onto an intranet as a way of more effectively sharing information within a corporate or private network.
  • the basic infrastructure for an intranet is an internal network connecting servers and desktops, which may or may not be connected to the Internet through a firewall. These intranets provide services to desktops via standard open network protocols which are well established in the industry. Intranets provide many benefits to the enterprises which employ them, such as simplified internal information management and improved internal communication using the browser paradigm. Integrating Internet technologies with a company's enterprise infrastructure and legacy systems also leverages existing technology investment for the party employing an intranet.
  • intranets and the Internet are closely related, with intranets being used for internal and secure communications within the business and the Internet being used for external transactions between the business and the outside world.
  • the term “networks” includes both the Internet and intranets. However, the distinction between the Internet and an intranet should be born in mind where applicable.
  • JAVA is a trademark of Sun Microsystems of Mountain View, CA.
  • the JAVA programming language resulted from programming efforts which initially were intended to be coded in the C++ programming language; therefore, the JAVA programming language has many commonalities with the C++ programming language.
  • the JAVA programming language is a simple, object-oriented, distributed, interpreted yet high performance, robust yet safe, secure, dynamic, architecture neutral, portable, and multi-threaded language.
  • the JAVA programming language has emerged as the programming language of choice for the Internet as many large hardware and software companies have licensed it from Sun Microsystems.
  • the JAVA programming language and environment is designed to solve a number of problems in modern programming practice.
  • the JAVA programming language omits many rarely used, poorly understood, and confusing features of the C++ programming language. These omitted features primarily consist of operator overloading, multiple inheritance, and extensive automatic coercions.
  • the JAVA programming language includes automatic garbage collection that simplifies the task of programming because it is no longer necessary to allocated and free memory as in the C programming language.
  • the JAVA programming language restricts the use of pointers as defined in the C programming language, and instead has true arrays in which array bounds are explicitly checked, thereby eliminating vulnerability to many viruses and nasty bugs.
  • the JAVA programming language includes objective-C interfaces and specific exception handlers.
  • the JAVA programming language has an extensive library of routines for coping easily with TCP/IP protocol (Transmission Control Protocol based on Internet protocol), HTTP (Hypertext Transfer Protocol) and FTP (File Transfer Protocol).
  • TCP/IP protocol Transmission Control Protocol based on Internet protocol
  • HTTP Hypertext Transfer Protocol
  • FTP File Transfer Protocol
  • the JAVA programming language is intended to be used in networked/distributed environments.
  • the JAVA programming language enabled the construction of virus-free, tamper-free systems.
  • the authentication techniques are based on public-key encryption.
  • a JAVA virtual machine is an stack-oriented abstract computing machine, which like a physical computing machine has an instruction set and uses various storage areas.
  • a JAVA virtual machine need not understand the JAVA programming language; instead it understands a class file format.
  • a class file includes JAVA virtual machine instructions (or bytecodes) and a symbol table, as well as other ancillary information. Programs written in the JAVA programming language (or in other languages) may be compiled to produce a sequence of JAVA virtual machine instructions.
  • instructions typically operate on data at the top of an operand stack.
  • One or more first instructions such as a load from local variable instruction, are executed to push operand data onto the operand stack as a precursor to execution of an instruction which immediately follows such instruction(s).
  • the instruction which follows e.g., an add operation, pops operand data from the top of the stack, operates on the operand data, and pushes a result onto the operand stack, replacing the operand data at the top of the operand stack.
  • a suitably configured instruction decoder allows the folding away of instructions pushing an operand onto the top of a stack merely as a precursor to a second instruction which operates on the top of stack operand.
  • the instruction decoder identifies foldable instruction sequences (typically 2, 3, or 4 instructions) and supplies an execution unit with an equivalent folded operation (typically a single operation) thereby reducing processing cycles otherwise required for execution of multiple operations corresponding to the multiple instructions of the folded instruction sequence.
  • foldable instruction sequences typically 2, 3, or 4 instructions
  • an equivalent folded operation typically a single operation
  • an instruction sequence including a pair of load instructions for loading integer operands from local variables to the top of stack
  • an add instruction for popping the integer operands of the stack, adding them, and placing the result at the top of stack
  • an store instruction for popping the result from the stack and storing the result in a local variable
  • an apparatus in accordance with an embodiment of the present invention, includes an instruction store, an operand stack, a data store, an execution unit, and an instruction decoder.
  • the instruction decoder is coupled to the instruction store to identify a foldable sequence of instructions represented therein.
  • the foldable sequence includes first and second instructions, in which the first instruction is for pushing a first operand value onto the operand stack from the data store merely as a first source operand for a second instruction.
  • the instruction decoder coupled to supply the execution unit with a single folded operation equivalent to the foldable sequence and including a first operand address identifier selective for the first operand value in the data store, thereby obviating an explicit operation corresponding to the first instruction.
  • the instruction decoder supplies the execution unit with an operation identifier and operand address identifier corresponding to the first instruction only.
  • the instruction decoder further identifies a third instruction in the foldable sequence.
  • This third instruction is for pushing a second operand value onto the operand stack from the data store merely as a second source operand for the second instruction.
  • the single folded operation is equivalent to the foldable sequence and includes a second operand address identifier selective for the second operand value in the data store, thereby obviating an explicit operation corresponding to the third instruction.
  • the instruction decoder further identifies a fourth instruction in the foldable sequence.
  • This fourth instruction is for popping a result of the second instruction from the operand stack and storing the result in a result location of the data store.
  • the single folded operation is equivalent to the foldable sequence and includes a destination address identifier selective for the result location in the data store, thereby obviating an explicit operation corresponding to the fourth instruction.
  • the instruction decoder includes normal and folded decode paths and switching means.
  • the switching means are responsive to the folded decode path for selecting operation, operand, and destination identifiers from the folded decode path in response to a fold indication therefrom, and for otherwise selecting operation, operand, and destination identifiers from the normal decode path.
  • the apparatus is for a virtual machine instruction processor wherein instructions generally source operands from, and target a result to, uppermost entries of an operand stack.
  • the virtual machine instruction processor is a hardware virtual machine instruction processor and the instruction decoder includes decode logic.
  • the virtual machine instruction processor includes a just-in-time compiler implementation and the instruction decoder includes software executable on a hardware processor.
  • the hardware processor includes the execution unit.
  • the virtual machine instruction processor includes a bytecode interpreter implementation and the instruction decoder including software executable on a hardware processor.
  • the hardware processor includes the execution unit.
  • a method includes (a) determining if a first instruction of a virtual machine instruction sequence is an instruction for pushing a first operand value onto the operand stack from a data store merely as a first source operand for a second instruction; and if the result of the (a) determining is affirmative, supplying an execution unit with a single folded operation equivalent to a foldable sequence comprising the first and second instructions.
  • the single folded operation includes a first operand identifier selective for the first operand value, thereby obviating an explicit operation corresponding to the first instruction.
  • the method includes supplying, if the result of the (a) determining is negative, the execution unit with an operation equivalent to the first instruction in the virtual machine instruction sequence.
  • the method includes (b) determining if a third instruction of the virtual machine instruction sequence is an instruction for popping a result value of the second instruction from the operand stack and storing the result value in a result location of the data store and, if the result of the (b) determining is affirmative, further including a result identifier selective for the result location with the equivalent single folded operation, thereby further obviating an explicit operation corresponding to the third instruction.
  • the method includes including, if the result of the (b) determining is negative, a result identifier selective for a top location of the operand stack with the equivalent single folded operation.
  • the (a) determining and the (b) determining are performed substantially in parallel.
  • a stack-based virtual machine implementation includes a randomly-accessible operand stack representation, a randomly-accessible local variable storage representation, and a virtual machine instruction decoder for selectively decoding virtual machine instructions and folding together a selected sequence thereof to eliminate unnecessary temporary storage of operands on the operand stack.
  • the stack-based virtual machine implementation (1) is a hardware virtual machine instruction processor including a hardware stack cache, a hardware instruction decoder, and an execution unit or (2) includes software encoded in a computer readable medium and executable on a hardware processor.
  • the hardware virtual machine instruction processor embodiment (a) the randomly-accessible operand stack local variable storage representations at least partially reside in the hardware stack cache, and (b) the virtual machine instruction decoder includes the hardware instruction decoder coupled to provide the execution unit with opcode, operand, and result identifiers respectively selective for a hardware virtual machine instruction processor operation and for locations in the hardware stack cache as a single hardware virtual machine instruction processor operation equivalent to the selected sequence of virtual machine instructions.
  • the randomly-accessible operand stack local variable storage representations at least partially reside in registers of the hardware processor
  • the virtual machine instruction decoder is at least partially implemented in the software
  • the virtual machine instruction decoder is coupled to provide opcode, operand, and result identifiers respectively selective for a hardware processor operation and for locations in the registers as a single hardware processor operation equivalent to the selected sequence of virtual machine instructions.
  • a hardware virtual machine instruction decoder includes a normal decode path, a fold decode path, and switching means.
  • the fold decode path is for decoding a sequence of virtual machine instructions and, if the sequence is foldable, supplying (a) a single operation identifier, (b) one or more operand identifiers; and (c) a destination identifier, which are together equivalent to the sequence of virtual machine instructions.
  • the switching means is responsive to the folded decode path for selecting operation, operand, and destination identifiers from the folded decode path in response to a fold indication therefrom, and otherwise selecting operation, operand, and destination identifiers from the normal decode path.
  • FIG. 1 (depicted as FIGS. 1A and 1B ) is a block diagram of one embodiment of a virtual machine hardware processor that includes an instruction decoder for providing instruction folding in accordance with some embodiments of the present invention.
  • FIG. 2 is an process flow diagram for generation of virtual machine instructions that are used in one embodiment of this invention.
  • FIG. 3 illustrates an instruction pipeline implemented in the hardware processor of FIG. 1 .
  • FIG. 4A is an illustration of the one embodiment of the logical organization of a stack structure where each method frame includes a local variable storage area, an environment storage area, and an operand stack utilized by the hardware processor of FIG. 1 .
  • FIG. 4B is an illustration of an alternative embodiment of the logical organization of a stack structure where each method frame includes a local variable storage area and an operand stack on the stack, and an environment storage area for the method frame is included on a separate execution environment stack.
  • FIG. 4C is an illustration of an alternative embodiment of the stack management unit for the stack and execution environment stack of FIG. 4B .
  • FIG. 4D is an illustration of one embodiment of the local variables look-aside cache in the stack management unit of FIG. 1 .
  • FIG. 5 illustrates several possible add-one to the hardware processor of FIG. 1 .
  • FIG. 6 illustrates a block diagram of one embodiment of a stack cache management unit in accordance with this invention.
  • FIG. 7 illustrates the memory architecture of one embodiment of a stack cache in accordance with this invention.
  • FIG. 8 illustrates the contents of a register or memory location of one embodiment of a stack cache in accordance with this invention.
  • FIG. 9 illustrates a block diagram of one embodiment of a dribble manager unit in accordance with this invention.
  • FIG. 10A illustrates a block diagram of another embodiment of a dribble manager unit in accordance with this invention.
  • FIG. 10B illustrates a block, diagram of another embodiment of a dribble manager unit in accordance with this invention.
  • FIG. 11 illustrates a block diagram of a portion of an embodiment of a dribble manager unit in accordance with this invention.
  • FIG. 12 illustrates an pointer generation circuit for one embodiment of a stack cache in accordance with this invention.
  • FIG. 13 is an illustration, in the context of a stack data structure, of data flows associated with a pair of stack instructions, wherein the first stack instruction pushes a data item onto the top of the stack only to be consumed by the second stack instruction which pops the top two stack entries off the stack and pushes their sum onto the top of the stack.
  • FIG. 14 is a contrasting illustration of folded execution of first and second stack instructions such as those depicted in FIG. 13 , wherein the first (push data item onto the top of the stack) operation is obviated in accordance with an exemplary embodiment of the present invention.
  • FIG. 15 is a block diagram depicting relationships between operand stack, local variable storage, and constant pool portions of memory storage together with register variables for access thereof in accordance with an exemplary embodiment of the present invention.
  • FIGS. 16 A-D illustrate an iload (integer load)/iadd (integer add)/istore (integer store) instruction sequence operating on an operand stack and local variable storage.
  • FIGS. 16A, 16B , 16 C, and 16 D respectively depict operand stack contents before iload instructions, after iload instructions but before an iadd instruction, after the iadd instruction but before an istore instruction, and after the istore instruction.
  • Intermediate stages depicted in FIGS. 16B and 16C are eliminated by instruction folding in accordance with an exemplary embodiment of the present invention.
  • FIGS. 17A, 17B and 17 C illustrate an aload (object reference load)/arraylength (integer add) instruction sequence operating on the operand stack and local variable storage.
  • FIGS. 17A, 17B , and 17 C respectively depict operand stack contents before the load instruction, after the aload instruction but before an arraylength instruction (without instruction folding), and after the arraylength instruction.
  • the intermediate stage depicted in FIG. 17B is eliminated by instruction folding in accordance with an exemplary embodiment of the present invention.
  • FIG. 18 is a functional block diagram of a stack based processor including an instruction decoder providing instruction folding in accordance with an exemplary embodiment of the present invention.
  • FIG. 19 is a functional block diagram depicting an instruction decoder in accordance with an exemplary embodiment of the present invention and coupled to supply an execution unit with a folded operation, with operand addresses into an operand stack, local variable storage or a constant pool, and with a destination address into the operand stack or local variable storage, wherein the single operation and addresses supplied are equivalent to a sequence of unfolded instructions.
  • FIG. 20 is a functional block diagram of an instruction decoder supporting instruction folding in accordance with an exemplary embodiment of the present invention.
  • FIG. 21 is a functional block diagram of a fold decode portion of an instruction decoder supporting instruction folding in accordance with an exemplary embodiment of the present invention.
  • FIG. 22 is a flow chart depicting an exemplary sequence of operations for identifying a foldable instruction sequence in accordance with an exemplary embodiment of the present invention.
  • FIG. 23 is a functional block diagram depicting component operand and destination address generators of a fold address generator in accordance with an exemplary embodiment of the present invention.
  • FIG. 24 is a functional block diagram depicting an exemplary structure for an operand address generator in accordance with an exemplary embodiment of the present invention.
  • FIG. 1 illustrates one embodiment of a virtual machine instruction hardware processor 100 , hereinafter hardware processor 100 , that includes an instruction decoder 135 for folding a sequence of multiple instructions into a single folded operation in accordance with some embodiments of the present invention, and that directly executes virtual machine instructions that are processor architecture independent.
  • the performance of hardware processor 100 in executing JAVA virtual machine instructions is much better than high-end CPUs, such as the Intel PENTIUM microprocessor or the Sun Microsystems ULTRASPARC processor, (ULTRASPARC is a trademark of Sun Microsystems of Mountain View, Calif.), and PENTIUM is a trademark of Intel Corp.
  • Hardware processor 100 provides similar advantages for other virtual machine stack-based architectures as well as for virtual machines utilizing features such as garbage collection, thread synchronization, etc.
  • a system based on hardware processor 100 presents attractive price for performance characteristics, if not the best overall performance, as compared with alternative virtual machine execution environments including software interpreters and just-in-time compilers.
  • the present invention is not limited to virtual machine hardware processor embodiments, and encompasses any suitable stack-based, or non-stack-based machine implementations, including implementations emulating the JAVA virtual machine as a software interpreter, compiling JAVA virtual machine instructions (either in batch or just-in-time) to machine instruction native to a particular hardware processor, or providing hardware implementing the JAVA virtual machine in microcode, directly in silicon, or in some combination thereof.
  • hardware processor 100 has the advantage that the 250 Kilobytes to 500 Kilobytes (Kbytes) of memory storage, e.g., read-only memory or random access memory, typically required by a software interpreter, is eliminated.
  • a simulation of hardware processor 100 showed that hardware processor 100 executes virtual machine instructions twenty times faster than a software interpreter running on a variety of applications on a PENTIUM processor clocked at the same clock rate as hardware processor 100 , and executing the same virtual machine instructions.
  • Another simulation of hardware processor 100 showed that hardware processor 100 executes virtual machine instructions five times faster than a just-in-time compiler running on a PENTIUM processor running at the same clock rate as hardware processor 100 , and executing the same virtual machine instructions.
  • hardware processor 100 is advantageous. These applications include, for example, an Internet chip for network appliances, a cellular telephone processor, other telecommunications integrated circuits, or other low-power, low-cost applications such as embedded processors, and portable devices.
  • the present invention increases the speed of method invocation by using an execution environment memory 440 in conjunction with stack 400 B.
  • the execution environment of various method calls are stored in execution environment memory 440 while the operands, variables and parameters of the method calls are stored in stack 400 B.
  • Both execution environment memory 440 and stack 400 B can include a stack management unit 150 that utilizes a stack cache 155 to accelerate data transfers for execution unit 140 .
  • stack management unit 150 can be an integral part of hardware processor 100 as shown in FIG. 1 , many embodiments of stack management unit 150 are not integrated with a hardware processor since stack management in accordance with the present invention can be adapted for use with any stack-based computing system.
  • stack management unit 150 includes a stack cache 155 , a dribble manager unit 151 , and a stack control unit 152 .
  • dribble manager unit 151 transfers data from the bottom of stack cache 155 to stack 400 through data cache unit 160 , so that the top portion of stack 400 remains in stack cache 155 .
  • dribble manager unit 151 transfers data from stack 400 to the bottom of stack cache 155 so that the top portion of stack 400 is maintained in stack cache 155 .
  • Instruction decoder 135 allows the folding away of JAVA virtual machine instructions pushing an operand onto the top of a stack merely as a precursor to a second JAVA virtual machine instruction which operates on the top of stack operand.
  • Such an instruction decoder identifies foldable instruction sequences and supplies an execution unit with a single equivalent folded operation thereby reducing processing cycles otherwise required for execution of multiple operations corresponding to the multiple instructions of the folded instruction sequence.
  • Instruction decoder embodiments described herein provide for folding of two, three, four, or more instruction folding. For example, in one instruction decoder embodiment described herein, two load instructions and a store instruction can be folded into execution of operation corresponding to an instruction appearing therebetween in the instruction sequence.
  • a virtual machine is an abstract computing machine that, like a real computing machine, has an instruction set and uses various memory areas.
  • a virtual machine specification defines a set of processor architecture independent virtual machine instructions that are executed by a virtual machine implementation, e.g., hardware processor 100 .
  • Each virtual machine instruction defines a specific operation that is to be performed.
  • the virtual computing machine need not understand the computer language that is used to generate virtual machine instructions or the underlying implementation of the virtual machine. Only a particular file format for virtual machine instructions needs to be understood.
  • the-virtual machine instructions are JAVA virtual machine instructions.
  • Each JAVA virtual machine instruction includes one or more bytes that encode instruction identifying information, operands, and any other required information.
  • Section I which is incorporated herein by reference in its entirety, includes an illustrative set of the JAVA virtual machine instructions. The particular set of virtual machine instructions utilized is not an essential aspect of this invention. In view of the virtual machine instructions in Section I and this disclosure, those of skill in the art can modify the invention for a particular set of virtual machine instructions, or for changes to the JAVA virtual machine specification.
  • a JAVA compiler JAVAC ( FIG. 2 ) that is executing on a computer platform, converts an application 201 written in the JAVA computer language to an architecture neutral object file format encoding a compiled instruction sequence 203 , according to the JAVA Virtual Machine Specification, that includes a compiled instruction set.
  • JAVA Virtual Machine Specification that includes a compiled instruction set.
  • only a source of virtual machine instructions and related information is needed. The method or technique used to generate the source of virtual machine instructions and related information is not essential to this invention.
  • Compiled instruction sequence 203 is executable on hardware processor 100 as well as on any computer platform that implements the JAVA virtual machine using, for example, a software interpreter or just-in-time compiler.
  • hardware processor 100 provides significant performance advantages over the software implementations.
  • hardware processor 100 processes the JAVA virtual machine instructions, which include bytecodes.
  • Hardware processor 100 executes directly most of the bytecodes. However, execution of some of the bytecodes is implemented via microcode.
  • firmware means microcode stored in ROM that when executed controls the operations of hardware processor 100 .
  • hardware-processor 100 includes an I/O bus and memory interface unit 110 , an instruction cache unit 120 including instruction cache 125 , an instruction decode unit 130 , a unified execution unit 140 , a stack management unit 150 including stack cache 155 , a data cache unit 160 including a data cache 165 , and program counter and trap control logic 170 . Each of these units is described more completely below.
  • each unit includes several elements.
  • the interconnections between elements within a unit are not shown in FIG. 1 , However, in view of the following description, those of skill in the art will understand the interconnections and cooperation between the elements in a unit and between the various units.
  • the pipeline stages implemented using the units illustrated in FIG. 1 include fetch, decode, execute, and write-back stages. If desired, extra stages for memory access or exception resolution are provided in hardware processor 100 .
  • FIG. 3 is an illustration of a four stage pipeline for execution of instructions in the exemplary embodiment of processor 100 .
  • fetch stage 301 a virtual machine instruction is fetched and placed in instruction buffer 124 ( FIG. 1 ).
  • the virtual machine instruction is fetched from one of (i) a fixed size cache line from instruction-cache 125 or (ii) microcode ROM 141 in execution unit 140 .
  • each virtual machine instruction is between one and five bytes long. Thus, to keep things simple, at least forty bits are required to guarantee that all of a given instruction is contained in the fetch.
  • Another alternative is to always fetch a predetermined number of bytes, for example, four bytes, starting with the opcode. This is sufficient for 95% of JAVA virtual machine instructions (See Section I). For an instruction requiring more than three bytes of operands, another cycle in the front end must be tolerated if four bytes are fetched. In this case, the instruction execution can be started with the first operands fetched even if the full set of operands is not yet available.
  • decode stage 302 the virtual machine instruction at the front of instruction buffer 124 ( FIG. 1 ) is decoded and instruction folding is performed if possible.
  • Stack cache 155 is accessed only if needed by the virtual machine instruction.
  • Register OPTOP that contains a pointer OPTOP to a top of a stack 400 ( FIG. 4 ), is also updated in decode stage 302 ( FIG. 3 ).
  • a register to store a pointer is illustrative only of one embodiment.
  • the pointer may be implemented using hardware register, a hardware counter, a software counter, a software pointer, or other equivalent embodiments known to those of skill in the art.
  • the particular implementation selected is not essential to the invention, and typically is made based on a price to performance trade-off.
  • execute stage 303 the virtual machine instruction is executed for one or more cycles.
  • an ALU in integer unit 142 FIG. 1
  • DCU data cache unit
  • traps are prioritized and taken at the end of execute stage 303 ( FIG. 3 ).
  • the branch address is calculated in execute stage 303 , as well as the condition upon which the branch is dependent.
  • Cache stage 304 is a non-pipelined stage.
  • Data cache 165 ( FIG. 1 ) is accessed if needed during execution stage 303 ( FIG. 3 ).
  • stage 304 is non-pipelined is because hardware processor 100 is a stack-based machine.
  • the instruction following a load is almost always dependent on the value returned by the load. Consequently, in this embodiment, the pipeline is held for one cycle for a data cache access. This reduces the pipeline stages, and the die area taken by the pipeline for the extra registers and bypasses.
  • Write-back stage 305 is the last stage in the pipeline. In stage 305 , the calculated data is written back to stack cache 155 .
  • Hardware processor 100 directly implements a stack 400 ( FIG. 4A ) that supports the JAVA virtual machine stack-based architecture (See Section I). Sixty-four entries on stack 400 are contained on stack cache 155 in stack management unit 150 . Some entries in stack 400 may be duplicated on stack cache 155 . Operations on data are performed through stack cache 155 .
  • Stack 400 of hardware processor 100 is primarily used as a repository of information for methods. At any point in time, hardware processor 100 is executing a single method. Each method has memory space, i.e., a method frame on stack 400 , allocated for a set of local variables, an operand stack, and an execution environment structure.
  • a new method frame e.g., method frame two 410
  • Current frame 410 FIG. 4A
  • the other method frames may contain a part of or all of the following six entities, depending-on various method invoking situations:
  • the invoker's method context is included in execution environment 422 , sometimes called frame state, that in turn includes: a return program counter value 431 that is the address of the virtual machine instruction, e.g., JAVA opcode, next to the method invoke instruction; a return frame 432 that is the location of the calling method's frame; a return constant pool pointer 433 that is a pointer to the calling method's constant pool table; a current method vector 434 that is the base address of the current method's vector table; and a current monitor address 435 that is the address of the current method's monitor.
  • a return program counter value 431 that is the address of the virtual machine instruction, e.g., JAVA opcode, next to the method invoke instruction
  • a return frame 432 that is the location of the calling method's frame
  • a return constant pool pointer 433 that is a pointer to the calling method's constant pool table
  • a current method vector 434 that is the base address of the current method's
  • the object reference is an indirect pointer to an object-storage representing the object being targeted for the method invocation.
  • JAVA compiler JAVAC See FIG. 2 .
  • This object reference is accessible as local variable zero during the execution of the method.
  • This indirect pointer is not available for a static method invocation as there is no target-object defined for a static method invocation.
  • the list of incoming arguments transfers information from the calling method to the invoked method. Like the object reference, the incoming arguments are pushed onto stack 400 by JAVA compiler generated instructions and may be accessed as local variables.
  • JAVA compiler JAVAC (See FIG. 2 .) statically generates a list of arguments for current method 410 ′ ( FIG. 4A ), and hardware processor 100 determines the number of arguments from the list.
  • the object reference is present in the frame for a non-static method invocation, the first argument is accessible as local variable one. For a static method invocation, the first argument becomes local variable zero.
  • the upper 32-bits i.e., the 32 most significant bits, of a 64-bit entity are placed on the upper location of stack 400 , i.e., pushed on the stack last.
  • the upper 32-bit portion of the 64-bit entity is on the top of the stack, and the lower 32-bit portion of the 64-bit entity is in the-storage location immediately adjacent to the top of stack 400 .
  • the local variable area on stack 400 ( FIG. 4A ) for current method 410 represents temporary variable storage space which is allocated and remains effective during invocation of method 410 .
  • JAVA compiler JAVAC FIG. 2 ) statically determines the required number of local variables and hardware processor 100 allocates temporary variable storage space accordingly.
  • the local variables When a method is executing on hardware processor 100 , the local variables typically reside in stack cache 155 and are addressed as offsets from pointer VARS ( FIGS. 1 and 4 A), which points to the position of the local variable zero. Instructions are provided to load the values of local variables onto operand stack 423 and store values from operand stack into local variables area 421 .
  • the information in execution environment 422 includes the invoker's method context.
  • hardware processor 100 pushes the invoker's method context onto newly allocated frame 410 , and later utilizes the information to restore the invoker's method context before returning.
  • Pointer FRAME. ( FIGS. 1 and 4 A) is a pointer to the execution environment of the current method.
  • each register in register set 144 ( FIG. 1 ) is 32-bits wide.
  • Operand stack 423 is allocated to support the execution of the virtual machine instructions within the current method.
  • Program counter register PC ( FIG. 1 ) contains the address of the next instruction, e.g., opcode, to be executed.
  • Locations on operand stack 423 ( FIG. 4A ) are used to store the operands of virtual machine instructions, providing both source and target storage locations for instruction execution.
  • the size of operand stack 423 is statically determined by JAVA compiler JAVAC ( FIG. 2 ) and hardware processor 100 allocates space for operand stack 423 accordingly.
  • Register OPTOP ( FIGS. 1 and 4 A) holds a pointer to a top of operand stack 423 .
  • the invoked method may return its execution result onto the invoker's top of stack, so that the invoker can access the return value with operand stack references.
  • the return value is placed on the area where an object reference or an argument is pushed before a method invocation.
  • Hardware processor 100 One way to speed up this process is for hardware processor 100 to load the execution environment in the background and indicate what has been loaded so far, e.g., simple one bit scoreboarding. Hardware processor 100 tries to execute the bytecodes of the called method as soon as possible, even though stack 400 is not completely loaded. If accesses are made to variables already loaded, overlapping of execution with loading of stack 400 is achieved, otherwise a hardware interlock occurs and hardware processor 100 just waits for the variable or variables in the execution to be loaded.
  • FIG. 4B illustrates another way to accelerate method invocation.
  • the execution environment of each method frame is stored separately from the local variable area and the operand stack of the method frame.
  • stack 400 B contains modified method frames, e.g. modified method frame 410 B having only local variable area 421 and operand stack 423 .
  • Execution environment 422 of the method frame is stored in execution environment memory 440 . Storing the execution environment in execution environment memory 440 reduces the amount of data in stack cache 155 . Therefore, the size of stack cache 155 can be reduced. Furthermore, execution environment memory 440 and stack cache 155 can be accessed simultaneously. Thus, method invocation can be accelerated by loading or storing the execution environment in parallel with loading or storing data onto stack 400 B.
  • the memory architecture of execution environment memory 440 is also a stack. As modified method frames are pushed onto stack 400 b through stack cache 155 , corresponding execution environments are pushed onto execution environment memory 440 . For example, since modified method frames 0 to 2 , as shown in FIG. 48 , are in stack 400 B, execution environments (EE) 0 to 2 , respectively, are stored in execution environment memory circuit 440 .
  • an execution environment cache can be added to improve the speed of saving and retrieving the execution environment during method invocation.
  • the architecture described more completely below for stack cache 155 , dribbler manager unit 151 , and stack control unit 152 for caching stack 400 can also be applied to caching execution environment memory 440 .
  • FIG. 4C illustrates an embodiment of stack management unit 150 modified to support both stack 400 b and execution environment memory 440 .
  • the embodiment of stack management unit 150 in FIG. 4C adds an execution environment stack cache 450 , an execution environment dribble manager unit 460 , and an execution environment stack control unit 470 .
  • execution dribble manager unit 460 transfers an entire execution environment between execution environment cache 450 and execution environment memory 440 during a spill operation or a fill operation.
  • I/O bus and memory ,interface unit 110 ( FIG. 1 ), sometimes called interface unit 110 , implements an interface between hardware processor 100 and a memory hierarchy which in an exemplary embodiment includes external memory and may optionally include memory storage and/or interfaces on the same die as hardware processor 100 .
  • I/O controller 111 interfaces with external I/O devices and memory controller 112 interfaces with external memory.
  • external memory means memory external to hardware processor 100 .
  • external memory either may be included on the same die as hardware processor 100 , may be external to the die containing hardware processor 100 , or may include both on- and off-die portions.
  • requests to I/O devices go through memory controller 112 which maintains an address map of the entire system including hardware processor 100 .
  • hardware processor 100 is the only master and does not have to arbitrate to use the memory bus.
  • alternatives for the input/output bus that interfaces with I/O bus and memory interface unit 110 include supporting memory-mapped schemes, providing direct support for PCI, PCMCIA, or other standard busses.
  • Fast graphics w/VIS or other technology
  • I/O bus and memory interface unit 110 generates read and write requests to external memory.
  • interface unit 110 provides an interface for instruction cache and data cache controllers 121 and 161 to the external memory.
  • Interface unit 110 includes arbitration logic for internal requests from instruction cache controller 121 and data cache controller 161 to access external memory and in response to a request initiates either a read or a write request on the memory bus to the external memory.
  • a request from data cache controller 161 is always treated as higher priority relative to a request from instruction cache controller 121 .
  • Interface unit 110 provides an acknowledgment signal to the requesting instruction cache controller 121 , or data cache controller 161 on read cycles so that the requesting controller can latch the data. On write cycles, the acknowledgment signal from interface unit 110 is used for flow control so that the requesting instruction cache controller 121 or data cache controller 161 does not generate a new request when there is one pending. Interface unit 110 also handles errors generated on the memory bus to the external memory.
  • Instruction cache unit (ICU) 120 fetches virtual machine instructions from instruction cache 125 and provides the instructions to instruction decode unit 130 .
  • instruction cache controller 121 upon a instruction cache hit, instruction cache controller 121 , in one cycle, transfers an instruction from instruction cache 125 to instruction buffer 124 where the instruction is held until integer execution unit IEU, that is described more completely below, is ready to process the instruction. This separates the rest of pipeline 300 ( FIG. 3 ) in hardware processor 100 from fetch stage 301 . If it is undesirable to incur the complexity of supporting an instruction-buffer type of arrangement, a temporary one instruction register is sufficient for most purposes. However, instruction fetching, caching, and buffering should provide sufficient instruction bandwidth to support instruction folding as described below.
  • the front end of hardware processor 100 is largely separate from the rest of hardware processor 100 . Ideally, one instruction per cycle is delivered to the execution pipeline.
  • the instructions are aligned on an arbitrary eight-bit boundary by byte aligner circuit 122 in response to a signal from instruction decode unit 130 .
  • the front end of hardware processor 100 efficiently deals with fetching from any byte position.
  • hardware processor 100 deals with the problems of instructions that span multiple cache lines of cache 125 .
  • the opcode is the first byte, the design is able to tolerate an extra cycle of fetch latency for the operands.
  • a very simple de-coupling between the fetching and execution of the bytecodes is possible.
  • instruction cache controller 121 In case of an instruction cache miss, instruction cache controller 121 generates an external memory request for the missed instruction to I/O bus and memory interface unit 110 . If instruction buffer 124 is empty, or nearly empty, when there is an instruction cache miss, instruction decode unit 130 is stalled, i.e., pipeline 300 is stalled. Specifically, instruction cache controller 121 generates a stall signal upon a cache miss which is used along with an instruction buffer empty signal to determine whether to stall pipeline 300 . Instruction cache 125 can be invalidated to accommodate self-modifying code, e.g., instruction cache controller 121 can invalidate a particular line in instruction cache 125 .
  • instruction cache controller 121 determines the next instruction to be fetched, i.e., which instruction in instruction cache 125 needs to accessed, and generates address, data and control signals for data and tag RAMs in instruction cache 125 .
  • instruction cache controller 121 determines the next instruction to be fetched, i.e., which instruction in instruction cache 125 needs to accessed, and generates address, data and control signals for data and tag RAMs in instruction cache 125 .
  • On a cache hit four bytes of data are fetched from instruction cache 125 in a single cycle, and a maximum of four bytes can be written into instruction buffer 124 .
  • Byte aligner circuit 122 aligns the data out of the instruction cache RAM and feeds the aligned data to instruction buffer 124 . As explained more completely below, the first two bytes in instruction buffer 124 are decoded to determine the length of the virtual machine instruction. Instruction buffer 124 tracks the valid instructions in the queue and updates the entries, as explained more completely below.
  • Instruction cache controller 121 also provides the data path and control for handling instruction cache misses. On an instruction cache miss, instruction cache controller 121 generates a cache fill request to I/O bus and memory interface unit 110 .
  • instruction cache controller 121 On receiving data from external memory, instruction cache controller 121 writes the data into instruction cache 125 and the data are also bypassed into instruction buffer 124 . Data are bypassed to instruction buffer 124 as soon as the data are available from external memory, and before the completion of the cache fill.
  • Instruction cache controller 121 continues fetching sequential data until instruction buffer 124 is full or a branch or trap has taken place.
  • instruction buffer 124 is considered full if there are more than eight bytes of valid entries in buffer 124 .
  • eight bytes of data are written into instruction cache 125 from external memory in response to the cache fill request sent to interface unit 110 by instruction cache unit 120 . If there is a branch or trap taken while processing an instruction cache miss, only after the completion of the miss processing is the trap or branch executed.
  • a fault indication is generated and stored into instruction buffer 124 along with the virtual machine instruction, i.e., a fault bit is set.
  • the line is not written into instruction cache 125 .
  • the erroneous cache fill transaction acts like a non-cacheable transaction except that a fault bit is set.
  • Instruction cache controller 121 also services non-cacheable instruction reads.
  • An instruction cache enable (ICE) bit in a processor status register in register set 144 , is used to define whether a load can be cached. If the instruction cache enable bit is cleared, instruction cache unit 120 treats all loads as non-cacheable loads.
  • Instruction cache controller 121 issues a non-cacheable request to interface unit 110 for non-cacheable instructions. When the data are available on a cache fill bus for the non-cacheable instruction, the data are bypassed into instruction buffer 124 and are not written into instruction cache 125 .
  • instruction cache 125 is a direct-mapped, eight-byte line size cache. Instruction cache 125 has a single cycle latency.
  • the cache size is configurable to 0K, 1K, 2K, 4K, 8K and 16K byte sizes where K means kilo.
  • the default size is 4K bytes.
  • Each line has a cache tag entry associated with the line. Each cache tag contains a twenty bit address tag field and one valid bit for the-default 4K byte size.
  • Instruction buffer 124 which, in an exemplary embodiment, is a twelve-byte deep first-in, first-out (FIFO) buffer, de-links fetch stage 301 ( FIG. 3 ) from the rest of pipeline 300 for performance reasons.
  • Each instruction in buffer 124 ( FIG. 1 ) has an associated valid bit and an error bit. When the valid bit is set, the instruction associated with that valid bit is a valid instruction. When the error bit is set, the fetch of the instruction associated with that error bit was an erroneous transaction.
  • Instruction buffer 124 includes an instruction buffer control circuit (not shown) that generates signals to pass data to and from instruction buffer 124 and that keeps track of the valid entries in instruction buffer 124 , i.e., those with valid bits set.
  • instruction buffer 124 in an exemplary embodiment, four bytes can be received into instruction buffer 124 in a given cycle. Up to five bytes, representing up to two virtual machine instructions, can be read out of instruction buffer 124 in a given cycle. Alternative embodiments, particularly those providing folding of multi-byte virtual machine instructions and/or those providing folding of more than two virtual machine instructions, provide higher input and output bandwidth. Persons of ordinary skill in the art will recognize a variety of suitable instruction buffer designs including, for example, alignment logic, circular buffer design, etc. When a branch or trap is taken, all the entries in instruction buffer 124 are nullified and the branch/trap data moves to the top of instruction buffer 124 .
  • a unified execution unit 140 is shown.
  • instruction decode unit 130 , integer unit 142 , and stack management unit 150 are considered a single integer execution unit
  • floating point execution unit 143 is a separate optional unit.
  • the various elements in the execution unit may be implemented using the execution unit of another processor.
  • the various elements included in the various units of FIG. 1 are exemplary only of one embodiment. Each unit could be implemented with all or some of the elements shown. Again, the decision is largely dependent upon a price vs. performance trade-off.
  • virtual machine instructions are decoded in decode stage 302 ( FIG. 3 ) of 35 pipeline 300 .
  • two bytes that can correspond to two virtual machine instructions, are fetched from instruction buffer 124 ( FIG. 1 ).
  • the two bytes are decoded in parallel to determine if the two bytes correspond to two virtual machine instructions, e.g., a first load top of stack, instruction and a second add top two stack entries instruction, that can be folded into a single equivalent operation. Folding refers to supplying a single equivalent operation corresponding to two or more virtual machine instructions.
  • a single-byte first instruction can be folded with a second instruction.
  • alternative embodiments provide folding of more than two virtual machine instructions, e.g., two to four virtual machine instructions, and of multi-byte virtual machine instructions, though at the cost of instruction decoder complexity and increased instruction bandwidth. See U.S. patent application Ser. No. 08/786,351, entitled “INSTRUCTION FOLDING FOR A STACK-BASED MACHINE” naming Marc Tremblay and James Michael O'Connor as inventors, assigned to the assignee of this application, and filed on Jan. 23, 1997 with Attorney Docket No. SP2036, now U.S. Pat. No. 6,026,485, which is incorporated herein by reference in its entirety.
  • the first byte which corresponds to the first virtual machine instruction
  • the first and second instructions are not folded.
  • An optional current object loader folder 132 exploits instruction folding, such as that described above, and in greater detail in U.S. patent application Ser. No. 08/786,351, entitled “INSTRUCTION FOLDING FOR A STACK-BASED MACHINE” naming Marc Tremblay and James Michael O'Connor as inventors, assigned to the assignee, of this application, and filed on Jan. 23, 1997, now U.S. Pat. No. 6,026,485, which is incorporated herein by reference in its entirety, in virtual machine instruction sequences which simulation results have shown to be particularly frequent and therefore a desirable target for optimization.
  • a method invocation typically loads an object reference for the corresponding object onto the operand stack and fetches a field from the object.
  • Instruction folding allows this extremely common virtual machine instruction sequence to be executed using an equivalent folded operation.
  • Quick variants are not part of the virtual machine instruction set (See Chapter 3 of Section I), and are invisible outside of a JAVA virtual machine implementation. However, inside a virtual machine implementation, quick variants have proven to be an effective optimization. (See Section A in Section I; which is an integral part of this specification.) Supporting writes for updates of various instructions to quick variants in a non-quick to quick translator cache 131 changes the normal virtual machine instruction to a quick virtual machine instruction to take advantage of the large benefits bought from the quick variants. In particular, as described in more detail in U.S. patent application Ser. No.
  • 08/788,805 entitled “NON-QUICK INSTRUCTION ACCELERATOR INCLUDING INSTRUCTION IDENTIFIER AND DATA SET STORAGE AND METHOD OF IMPLEMENTING SAME” naming Marc Tremblay and James Michael O'Connor as inventors, assigned to the assignee of this application, and filed on Jan. 23, 1997 with Attorney Docket No. SP2039, now U.S. Pat. No. 6,065,108, which is incorporated herein by reference in its entirety, when the information required to initiate execution of an instruction has been assembled for the first time, the information is stored in a cache along with the value of program counter PC as a tag in non-quick to quick translator cache 131 and the instruction is identified as a quick-variant. In one embodiment, this is done with self-modifying-code.
  • instruction decode unit 130 Upon a subsequent call of that instruction, instruction decode unit 130 detects that the instruction is identified as a quick-variant and simply retrieves the information needed to initiate execution of the instruction from non-quick to quick translator cache 131 .
  • Non-quick to quick translator cache is an optional feature of hardware processor 100 .
  • branch predictor circuit 133 With regard to branching, a very short pipe with quick branch resolution is sufficient for most implementations. However, an appropriate simple branch prediction mechanism can alternatively be introduced, e.g., branch predictor circuit 133 . Implementations for branch predictor circuit 133 include branching based on opcode, branching based on offset, or branching based on a two-bit counter mechanism.
  • Operand stack 423 contains a reference to an object and some number of arguments when this instruction is executed.
  • Index bytes one and two are used to generate an index into the constant pool of the current class.
  • the item in the constant pool at that index points to a complete method signature and class. Signatures are defined in Section I and that description is incorporated herein by reference.
  • the method signature a short, unique identifier for each method, is looked up in a method table of the class indicated.
  • the result of the lookup is a method block that indicates the type of method and the number of arguments for the method.
  • the object reference and arguments are popped off this method's stack and become initial values of the local variables of the new method.
  • the execution then resumes with the first instruction of the new method.
  • instructions invokevirtual, opcode 182 , and invokestatic, opcode 184 invoke processes similar to that just described. In each case, a pointer is used to lookup a method block.
  • a method argument cache 134 that also is an optional feature of hardware processor 100 , is used, in a first embodiment, to store the method block of a method for use after the first call to the method, along with the pointer to the method block as a tag.
  • Instruction decode unit 130 uses index bytes one and two to generate the pointer and then uses the pointer to retrieve the method block for that pointer in cache 134 . This permits building the stack frame for the newly invoked method more rapidly in the background in subsequent invocations of the method.
  • Alternative embodiments may use a program counter or method identifier as a reference into cache 134 . If there is a cache miss, the instruction is executed in the normal fashion and cache 134 is updated accordingly.
  • the particular process used to determine which cache entry is overwritten is not an essential aspect of this invention. A least-recently used criterion could be implemented, for example.
  • method argument cache 134 is used to store the pointer to the method block, for use after the first call to the method, along with the value of program counter PC of the method as a tag.
  • Instruction decode unit 130 uses the value of program counter PC to access cache 134 . If the value of program counter PC is equal to one of the tags in cache 134 , cache 134 supplies the pointer stored with that tag to instruction decode unit 130 . Instruction decode unit 130 uses the supplied painter to retrieve the method block for the method.
  • Wide index forwarder 136 which is an optional element of hardware processor 100 , is a specific embodiment of instruction folding for instruction wide. Wide index forwarder 136 handles an opcode encoding an extension of an index operand for an immediately subsequent virtual machine instruction. In this way, wide index forwarder 136 allows instruction decode unit 130 to provide indices into local variable storage 421 when the number of local variables exceeds that addressable with a single byte index without incurring a separate execution cycle for instruction wide.
  • instruction decoder 135 particularly instruction folding, non-quick to quick translator cache 131 , current object loader folder 132 , branch predictor 133 , method argument cache 134 , and wide index forwarder 136 are also useful in implementations that utilize a software interpreter or just-in-time compiler, since these elements can be used to accelerate the operation of the software interpreter or just-in-time compiler.
  • the virtual machine instructions are translated to an instruction for the processor executing the interpreter or compiler, e.g., any one of a Sun processor, a DEC processor, an Intel processor, or a Motorola processor, for example, and the operation of the elements is modified to support execution on that processor.
  • the translation from the virtual machine instruction to the other processor instruction can be done either with a translator in a ROM or a simple software translator.
  • dual instruction set processors see U.S. patent application Ser. No. 08/787,618, entitled “A PROCESSOR FOR EXECUTING INSTRUCTION SETS RECEIVED FROM A NETWORK OR FROM A LOCAL MEMORY” naming Marc Tremblay and James Michael O'Connor as inventors, now U.S. Pat. No. 5,925,123, assigned to the assignee of this application, and filed on Jan. 23, 1997 with Attorney Docket No. SP2042, which is incorporated herein by reference in its entirety.
  • processor 100 implements instruction folding to enhance the performance of processor 100 .
  • instruction folding in accordance with the present invention can be used in any of a stack-based virtual machine implementation, including, e.g., in a hardware processor implementation, in a software interpreter implementation, in a just-in-time compiler implementation, etc.
  • a hardware processor implementation e.g., a hardware processor implementation
  • a software interpreter implementation e.g., a software interpreter implementation
  • just-in-time compiler implementation etc.
  • FIG. 14 illustrates folded execution of first and second stack instructions, according to the principles of this invention.
  • a first operand for an addition instruction resides in top-of-stack (TOS) entry 141 la of stack 1410 .
  • a second operand resides in entry 1412 of stack 1410 . Notice that entry 1412 is not physically adjacent to top-of stack entry 141 la and in fact, is in the interior of stack 1410 .
  • An instruction stream includes a load top-of-stack instruction for pushing the second operand onto the top of stack (see description of instruction iload in Section I) and an addition instruction for operating on the first and second operands residing in the top two entries of stack 1410 (see description of instruction iadd in Section I).
  • the load top-of-stack and addition instructions are folded into a single operation whereby the explicit sequential execution of the load top-of-stack instruction and the associated execution cycle are eliminated. Instead, a folded operation corresponding to the addition instruction operates on the first and second operands, which reside in TOS entry 141 la and entry 1412 of stack 1410 . The result of the folded operation is pushed onto stack 1410 at TOS entry 141 lb.
  • folding according to the principles of this invention enhances performance compared to an unfolded method for executing the same sequence of instructions.
  • a first operand for an addition instruction resides in top-of-stack (TOS) entry 1311 a of stack 1310 (see FIG. 13 ).
  • a second operand resides in entry 1312 of stack 1310 .
  • a load to top-of-stack instruction pushes the second operand onto the top of stack 1310 and typically requires an execution cycle. The push results in the second and first operands residing in TOS entry 1311 b and (TOS-1) entry 1313 , respectively.
  • the addition instruction operates, in another execution cycle, on the first and second operands which properly reside in the top two entries, i.e., TOS entry 1311 b and (TOS-1) entry 1313 , of stack 1310 in accordance with the semantics of a stack architecture.
  • the result of the addition instruction is pushed onto stack 1310 at TOS entry 1311 c and after the addition instruction is completed, it is as if the first and second operand data were never pushed onto stack 1310 .
  • folding reduces the execution cycles required to complete the addition and so enhances the speed of execution of the instruction stream. More complex folding, e.g., folding including store instructions and folding including larger numbers of instructions, is described in greater detail below.
  • instruction decoder unit 130 examines instructions in a stream of instructions. Instruction decoder unit 130 folds first and second adjacent instructions together and provides a single equivalent operation for execution by execution unit 140 when instruction decoder unit 130 detects that the first and second instructions have neither structural nor resource dependencies and the second instruction operates on data provided by the first instruction. Execution of the single operations obtains the same result as execution of an operation corresponding to the first instruction followed by execution an operation corresponding to the second instruction, except that an execution cycle has been eliminated.
  • the JAVA virtual machine is stack-oriented and specifies an instruction set, a register set, an operand stack, and an execution environment.
  • the present invention is described in relation to the JAVA Virtual Machine, those of skill in the art will appreciate that the invention is not limited to embodiments implementing or related to the JAVA virtual machine and, instead, encompasses systems, articles, methods, and apparati for a wide variety of stack machine environments, both virtual and physical.
  • each method has storage allocated for an operand stack and a set of local variables.
  • a series of method frames e.g., method frame 1501 and method frame 1502 on stack 1503 , each include an operand stack instance, local variable storage instance, and frame state information instance for respective methods invoked along the execution path of a JAVA program.
  • a new frame is created and becomes current each time a method is invoked and is destroyed after the method completes execution.
  • a frame ceases to be current if its method invokes another method.
  • the current frame passes back the result of its method invocation, if any, to the previous frame via stack 1503 .
  • Folding in accordance with the present invention is not dependent upon a particular process used to allocate or define memory space for a method, such as a frame, and can, in general, be used in any stack based architecture.
  • This series of method frames may be implemented in any of a variety of suitable memory hierarchies, including for example register/cache/memory hierarchies.
  • an operand stack instance 1512 ( FIG. 15 ) is implemented in randomly-accessible storage 1510 , i.e., at least some of the entries in operand stack instance 1512 can be accessed from locations other than the top most locations of operand stack instance 1512 in contrast with a conventional stack implementation in which only the top entry or topmost entries of the stack can be accessed.
  • register OPTOP stores a pointer that identifies the top of operand stack instance 1512 associated with the current method. The value stored in register OPTOP is maintained to identify the top entry of an operand stack instance corresponding to the current method.
  • local variables for the current method are represented in randomly-accessible storage 1510 .
  • a pointer stored in register VARS identifies the starting address of local variable storage instance 1513 associated with the current method.
  • the value in register VARS is maintained to identify a base address of the local variable storage instance corresponding to the current method.
  • Entries in operand stack instance 1512 and local variable storage instance 1513 are referenced by indexing off of values represented in registers OPTOP and VARS, respectively, that in the embodiment of FIG. 1 are included in register set 144 , and in the embodiment of FIG. 15 are included in pointer registers 1522 .
  • Pointer registers 1522 may be represented in physical registers of a processor implementing the JAVA Virtual Machine, or optionally, in randomly-accessible storage 1510 .
  • commonly used offsets OPTOP-1, OPTOP-2, VARS+1, VARS+2, and VARS+3 are derived from the values in registers OPTOP and VARS, respectively.
  • the additional offsets could be stored in registers of pointer registers 1522 .
  • Operand stack instance 1512 and local variable storage instance 1513 associated with the current method are preferably represented in a flat 64-entry cache, e.g., stack cache 155 (see FIG. 1 ) whose contents are kept updated so that a working set of operand stack and local variable storage entries are cached.
  • a flat 64-entry cache e.g., stack cache 155 (see FIG. 1 )
  • the current frame including operand stack instance 1512 and local variable storage instance 1513 may be fully or partially represented in the cache.
  • Operand stack and local variable storage entries for frames other than the current frame may also be represented in the cache if space allows.
  • a suitable representation of a cache suitable for use with the folding of this invention is described in greater detail in U.S.
  • a constant area 1514 is provided in the address space of a processor implementing the JAVA virtual machine for commonly-used constants, e.g., constants specified by JAVA virtual machine instructions such as instruction iconst.
  • an operand source is represented as an index into constant area 1514 .
  • constant area 1514 is represented in randomly-accessible storage 1510 .
  • entries of constant area 1514 could also be cached, e.g., in stack cache 155 .
  • operand stack 1512 and local variable storage 1513 are referred to simply as operand stack 1512 and local variable storage 1513 .
  • operand stack 1512 and local variable storage 1513 refer to any instances of an operand stack and variable storage associated with the current method, including representations which maintain separate instances for each method and representations which combine instances into a composite representation.
  • Operand sources and result targets for JAVA Virtual Machine instructions typically identify entries of operand stack instance 1512 or local variable storage instance 1513 , i.e., identify entries of the operand stack and local variable storage for the current method.
  • representative JAVA virtual machine instructions are described in Chapter 3 of The JAVA Virtual Machine Specification which is included at Section I.
  • JAVA virtual machine instructions rarely explicitly designate both the source of the operand, or operands, and the result destination. Instead, either the source or the destination is implicitly the top of operand stack 1512 .
  • Some JAVA bytecodes explicitly designate neither a source nor a destination.
  • instruction iconst — 0 pushes a constant integer zero onto operand stack 1512 .
  • the constant zero is implicit in the instruction, although the instruction may actually be implemented by a particular JAVA virtual machine implementation using a representation of the value zero from a pool of constants, such as constant area 1514 , as the source for the zero operand.
  • An instruction decoder for a JAVA virtual machine implementation that implements instruction iconst — 0 in this way could generate, as the source address, the index of the entry in constant area 1514 where the constant zero is represented.
  • the JAVA virtual machine integer add instruction, iadd generates the sum of first and second integer operands, referred to as operand1 and operand2, respectively, that are at the top two locations of operand stack 1512 .
  • the top two locations are identified, at the time of instruction iadd execution, by pointer OPTOP in register OPTOP and by pointer OPTOP-1.
  • the result of the execution of instruction iadd, i.e., the sum of first and second integer operands, is pushed onto operand stack 1512 .
  • FIG. 16A shows the state of operand stack 1512 and local variable storage 1513 that includes first and second values, referred to as value1 and value2, before execution of a pair of JAVA virtual machine integer load instructions iload.
  • pointer OPTOP has the value AAC0h.
  • FIG. 16B shows operand stack 1512 after execution of the pair of instructions iload that load integer values from local variable storage 1513 onto operand stack 1512 , pushing (i.e., copying) values value1 and value2 from locations identified by pointer VARS in register VARS and by pointer VARS+2 onto operand stack 1512 as operand1 at location AAC4h and operand2 at location AAC8h, and updating pointer OPTOP in the process to value AAC8h.
  • FIG. 16C shows operand stack 1512 after instruction iadd has been executed.
  • Execution of instruction iadd pops operands operand1 and operand2 off operand stack 1512 , calculates the sum of operands operand1 and operand2, and pushes that sum onto operand stack 1512 at location AAC4h.
  • pointer OPTOP has the value AAC0h and points to the operand stack 1512 entry storing the sum.
  • FIG. 16D shows operand stack 1512 after an instruction istore has been executed. Execution of instruction istore pops the sum off operand stack 1512 and stores the sum in the local variable storage 1513 entry at the location identified by pointer VARS+2.
  • FIGS. 17 A-C The folding example of FIGS. 17 A-C is analogous to that illustrated with reference to FIGS. 16 A-D, though with only load folding illustrated.
  • Execution of JAVA virtual machine length of array instruction arraylength determines the length of an array whose object reference pointer objectref is at the top of operand stack 1512 , and pushes the length onto operand stack 1512 .
  • FIG. 17A shows the state of operand stack 1512 and local variable storage 1513 before execution of JAVA virtual machine reference load instruction aload that is used to load an object reference from local variable storage 1513 onto the top of operand stack 1512 .
  • pointer OPTOP has the value AAC0h.
  • FIG. 17B shows operand stack 1512 after execution of instruction aload pushes, i.e., copies, object reference pointer objectref onto the top of operand stack 1512 and updates pointer OPTOP to AAC4h in the process.
  • FIG. 17C shows operand stack 1512 after instruction arraylength has been executed.
  • Execution of instruction arraylength pops object reference pointer objectref off operand stack 1512 , calculates the length of the array referenced thereby, and pushes that length onto operand stack 1512 .
  • Suitable implementations of the instruction arraylength may supply object reference pointer objectref to an execution unit, e.g., execution unit 140 , which subsequently overwrites the object reference pointer objectref with the value length.
  • execution unit 140 e.g., execution unit 140
  • pointer OPTOP has the value AAC4h and points to the operand stack 1512 entry storing the value length.
  • FIG. 18 illustrates a processor 1800 wherein loads, such as those illustrated in FIGS. 16A and 16B and in FIGS. 17A and 17B , are folded into execution of subsequent instructions, e.g., into execution of subsequent instruction iadd, or instruction arraylength.
  • loads such as those illustrated in FIGS. 16A and 16B and in FIGS. 17A and 17B
  • FIG. 18 illustrates a processor 1800 wherein loads, such as those illustrated in FIGS. 16A and 16B and in FIGS. 17A and 17B , are folded into execution of subsequent instructions, e.g., into execution of subsequent instruction iadd, or instruction arraylength.
  • loads such as those illustrated in FIGS. 16A and 16B and in FIGS. 17A and 17B
  • FIG. 18 illustrates a processor 1800 wherein loads, such as those illustrated in FIGS. 16A and 16B and in FIGS. 17A and 17B , are folded into execution of subsequent instructions, e.g., into execution of subsequent instruction iadd, or instruction arraylength.
  • hardware processor 1800 includes other embodiments that do not include the various optimizations of hardware processor 100 .
  • the folding processes described below could be implemented in a software interpreter or a included within a just-in-time compiler.
  • stores such as that illustrated in FIG. 16D are folded into execution of prior instructions, e.g., in FIG. 16D , into execution of the immediately prior instruction iadd.
  • Instruction decoder 1818 retrieves fetched instructions from instruction buffer 1816 and depending upon the nature of instructions in the fetched instruction sequence, supplies execution unit 1820 with decoded operation and operand addressing information implementing the instruction sequence as a single folded operation. Unlike instructions of the JAVA virtual machine instruction set to which the instruction sequence from instruction buffer 1816 conforms, decoded operations supplied to execution unit 1820 by instruction decoder 1818 operate on operand values represented in entries of local variable storage 1513 , operand stack 1512 , and constant area 1514 .
  • valid operand sources include local variable storage 1513 entries identified by pointers VARS, VARS+1, VARS+2, and VARS+3, as well as operand stack 1512 entries identified by pointers OPTOP, OPTOP-1 and OPTOP-2.
  • valid result targets include local variable storage 1513 entries identified by operands VARS, VARS+1, VARS+2, and VARS+3.
  • Embodiments in accordance with FIG. 18 may also provide for constant area 1514 entries as valid operand sources as well as other locations in operand stack 1512 and local variable storage 1513 .
  • instruction buffer 1816 is organized as a shift register for JAVA bytecodes.
  • One or more bytecodes are decoded by instruction decoder 1818 during each cycle and operations are supplied to execution unit 1820 in the form of a decoded operation on instruction decode bus instr_dec and associated operand source and result destination addressing information on instruction address bus instr_addr.
  • Instruction decoder 1818 also provides an instruction valid signal instr_valid to execution unit 1820 . When asserted, signal instr_valid indicates that the information on instruction decode bus instr_dec specifies a valid operation.
  • One or more bytecodes are shifted out of instruction buffer 1816 to instruction decode unit 1818 each cycle in correspondence with the supply of decoded operations and operand addressing information to execution unit 1820 , and subsequent undecoded bytecodes are shifted into instruction buffer 1816 .
  • a single instruction is shifted out of instruction buffer 1816 and decoded by instruction decode unit 1818 , and a single corresponding operation is executed by execution unit 1820 during each instruction cycle.
  • multiple instructions e.g., a group of instructions
  • instruction decode unit 1818 In response to the multiple instructions, instruction decode unit 1818 generates a single equivalent folded operation that in turn is executed by execution unit 1820 during each instruction cycle.
  • instruction decoder 1818 selectively decodes bytecodes associated with four JAVA virtual machine instructions:
  • both instructions iload and the instruction istore are folded by instruction decoder 1818 into an add operation corresponding to instruction iadd.
  • instruction decoder 1818 is illustrated using a foldable sequence of four instructions, those of skill in the art will appreciate that the invention is not limited to four instructions. Foldable sequences of two, three, four, five, or more instructions are envisioned. For example, more than one instruction analogous to the instruction istore and more than two instructions analogous to the instructions iload may be included in foldable sequences.
  • Instruction decoder 1818 supplies decoded operation information over bus instr_dec and associated operand source and result destination addressing information over bus instr_addr specifying that execution unit 1820 is to add the contents of local variable storage 1513 location 0 , this is identified by pointer VARS, and local variable storage 1513 location 2 , that is identified by pointer VARS+2, and store the result in local variable storage 1513 location 2 , that is identified by pointer VARS+2.
  • the two load instructions are folded into execution of an operation corresponding to instruction iadd. Two instruction cycles and the intermediate data state illustrated in FIG. 16B are eliminated.
  • instruction istore is also folded into execution of the operation corresponding to instruction iadd, eliminating another instruction cycle, for a total of three, and the intermediate data state illustrated in FIG. 16C .
  • instruction folding in accordance with the present invention may eliminate loads, stores, or both loads and stores.
  • FIG. 20 depicts an exemplary embodiment of an instruction decoder 1818 providing both folded and unfolded decoding of bytecodes. Selection of a folded or unfolded operating mode for instruction decoder 1818 is based on the particular sequence of bytecodes fetched into instruction buffer 1816 and subsequently accessed by instruction decoder 1818 .
  • a normal decode portion 2002 and a fold decode portion 2004 of instruction decoder 1818 are configured in parallel to provide support for unfolded and folded execution, respectively.
  • fold decode portion 2004 detects opportunities for folding execution of bytecodes in the bytecode sequence fetched into instruction buffer 1816 .
  • a detection of such a foldable sequence triggers selection of the output of fold decode portion 2004 , rather than normal decode portion 2002 , for provision to execution unit 1820 .
  • selection of folded or unfolded decoding is transparent to execution unit 1820 , which simply receives operation information over bus instr_dec and associated operand source and result destination addressing information over bus instr_addr, and which need not know whether the information corresponds to a single instruction or a folded instruction sequence.
  • Normal decode portion 2002 functions to inspect a single bytecode from instruction buffer 1816 during each instruction cycle, and generates the following indications in response thereto:
  • n_instr_dec which specifies an operation, e.g., integer addition, corresponding to the decoded instruction, is provided to a first set of input terminals of switch 2006 ;
  • an instruction valid signal instr_valid which indicates whether normal instruction decode signal n_instr_dec specifies a valid operation, is provided to a first input terminal of switch 2008 .
  • fold decode portion 2004 of instruction decoder 1818 inspects sequences of bytecodes from the instruction buffer 1816 and determines whether operations corresponding to these sequences (e.g., the sequence iload value1 from local variable 0, iload value2 from local variable 2, iadd, and istore sum to local variable 2) can be folded together to eliminate unnecessary temporary storage of instruction operands and/or results on the operand stack.
  • fold decode portion 2004 determines that a sequence of bytecodes in instruction buffer 1816 can be folded together, fold decode portion 2004 generates the following indications:
  • a folded instruction decode signal f_instr_dec which specifies an equivalent operation, e.g., integer addition corresponding to the folded instruction sequence, is provided to a second set of input terminals of switch 2006 ;
  • a folded instruction valid signal f_Valid which indicates whether folded instruction decode signal f_instr dec specifies a valid operation, is provided to a second input terminal of switch 2008 .
  • Fold decode portion 2004 also generates a signal on fold line f/nf which indicates whether a sequence of bytecodes in instruction buffer 1816 can be folded together.
  • the signal on fold line f/nf is provided to control inputs of switches 2006 , 2010 and 2008 . If a sequence of bytecodes in instruction buffer 1816 can be folded together, the signal on fold line f/nf causes switches 2006 , 2010 and 2008 to select respective second inputs for provision to execution unit 1820 , i.e., to source folded instruction decode signal f_instr_dec, folded address signal f_adr, and folded instruction valid signal f_valid from fold decode portion 2004 .
  • the signal on fold line f/nf causes switches 2006 , 2010 and 2008 to select respective first inputs for provision to execution unit 1820 , i.e., to source normal instruction decode signal n_instr_dec, normal address signal n_adr, and normal instruction valid signal n valid from fold decode portion 2004 .
  • the operation of fold decode portion 2004 is suppressed in response to an active suppress folding signal suppress_fold supplied from outside instruction decoder 1818 .
  • an asserted suppress folding signal suppress_fold see FIG. 21
  • the signal on fold line f/nf remains in a state selective for respective first inputs of switches 2006 , 2010 and 2008 even if the particular bytecode sequence presented by instruction buffer 1816 would otherwise trigger folding.
  • suppress folding signal suppress_fold is asserted when the local variable storage 1513 entry identified by pointer VARS is not cached, e.g., when entries in operand stack 1512 have displaced local variable storage 1513 from a stack cache 155 .
  • a stack cache and cache control mechanism representing at least a portion of operand stack 1512 and local variable storage 1513 may advantageously assert suppress folding signal suppress_fold if fold-relevant entries of local variable storage 1513 or operand stack 1512 are not represented in stack cache 155 .
  • FIG. 21 illustrates fold decode portion 2004 of instruction decoder 1818 in greater detail.
  • a fold determination portion 2104 selectively inspects the sequence of bytecodes in instruction buffer 1816 . If the next bytecode and one or more subsequent bytecodes represent a foldable sequence of operations (as discussed below with respect to FIG. 22 ), then fold determination portion 2104 supplies a fold-indicating signal on fold line f/nf and a folded instruction decode signal f_instr_dec that specifies an equivalent folded operation. Folded instruction decode signal f_instr_dec is supplied to execution unit 1820 as the decoded instruction instr_dec.
  • a foldable sequence of operations includes those associated with 2, 3, or 4 bytecodes from instruction decoder 1818 (up to 2 bytecodes loading operands onto operand stack 1512 , a bytecode popping the operand(s), operating thereupon, and pushing a result onto operand stack 1512 , and a bytecode popping the result from operand stack 1512 and storing the result.
  • the equivalent folded operation which is encoded by the folded instruction decode signal f_instr_dec, specifies an operation, that when combined with folded execution addressing information obviates the loads to, and stores from, operand stack 1512 .
  • Alternative embodiments may fold only two instructions, e.g., an instruction iload into an instruction iadd or an instruction istore back into an immediately prior instruction iadd.
  • Other alternative embodiments may fold only instructions that push operands onto the operand stack, e.g., one or more instructions iload folded into an instruction iadd, or only instructions that pop results from the operand stack, e.g., an instruction istore back into an immediately prior instruction iadd.
  • Further alternative embodiments may fold larger numbers of instructions that push operands onto the operand stack and/or instructions that pop results from the operand stack instructions in accordance with instructions of a particular virtual machine instruction set. In such alternative embodiments, the above described advantages over normal decoding and execution of instruction sequences are still obtained.
  • Fold determination portion 2104 generates a series of fold address index composite signal f_adr_ind including component first operand index signal first_adr_ind, second operand index signal second_adr_mind, and destination index signal dest_adr_mind, which are respectively selective for a first operand address, a second operand address, and a destination address for the equivalent folded operation.
  • Fold determination portion 2104 provides the composite signal f_adr_ind to fold address generator 2102 for use in supplying operand and destination addresses for the equivalent folded operation.
  • Fold determination portion 2104 asserts a fold-indicating signal on fold line f/nf to control the switches 2006 , 2010 and 2008 (see FIG.
  • fold determination portion 2104 begins an instruction decode cycle and transfers processing to initialize index 2202 .
  • initialize index 2202 an instruction index instr_index into instruction buffer 1816 is initialized to identify the next bytecode of a bytecode sequence in instruction buffer 1816 .
  • instruction index instr_index is initialed to one (1) and the next bytecode is the first bytecode in instruction buffer 1816 since prior bytecodes have already been shifted out of instruction buffer 1816 , although a variety of other indexing and instruction buffer management schemes would also be suitable.
  • initialize index 2202 transfers processing to first instruction check 2204 .
  • first instruction check 2204 determines whether the instruction identified by index instr_index, i.e., the first bytecode, corresponds to an operation that pushes a value, e.g., an integer value, a floating point value, a reference value, etc., onto operand stack 1512 .
  • first instruction check 2204 determines whether the instruction identified by index instr_index is one that the JAVA virtual machine specification (see Section I) defines as for pushing a first data item onto the operand stack. If so, first operand index signal first_adr_mind is asserted (at first operand address setting 2206 ) to identify the source of the first operand value.
  • first operand index signal first_adr_mind is selective for one of OPTOP, OPTOP-1, OPTOP-2, VARS, VARS+1, VARS+2, and VARS+3, although alternative embodiments may encode larger, smaller, or different sets of source addresses, including for example, source addresses in constant area 1514 .
  • this first bytecode may correspond to an operation which can be folded into the execution of a subsequent operation.
  • fold determination portion 2104 supplies a nonfold-indicating signal on fold line f/nf, whereupon indications from normal decode portion 2002 provide the decoding.
  • index instr_index is incremented (at incrementing 2208 ) to point to the next bytecode in instruction buffer 1816 .
  • fold determination portion 2104 determines whether instruction identified by index instr_index, i.e., the second bytecode, corresponds to an operation that pushes a value, e.g., an integer value, a floating point value, a reference value, etc., onto operand stack 1512 .
  • second instruction check 2210 determines whether the instruction identified by index instr_index is one that the JAVA virtual machine specification (see Section I) defines as for pushing a first data item onto the operand stack. If so, second operand index signal second_adr_mind is asserted (at second operand address setting 2212 ) to indicate the source of the second operand value and index instr_index is incremented (at incrementing 2214 ) to point to the next bytecode in instruction buffer 1816 .
  • second operand index signal second_adr_mind is selective for one of OPTOP, OPTOP-1, OPTOP-2, VARS, VARS+1, VARS+2, and VARS+3, although alternative embodiments are also suitable.
  • Fold determination portion 2104 continues at third instruction check 2216 with index instr_index pointing to either the second or third bytecode in instruction buffer 1816 .
  • fold determination portion 2104 determines whether the instruction identified by index instr_index, i.e., either the second or third bytecode, corresponds to an operation that operates on an operand value or values, e.g., integer value (s), floating point value(s), reference value(s), etc., from the uppermost entries of operand stack 1512 , effectively popping such operand values from operand stack 1512 and pushing a result value onto operand stack 1512 .
  • Popping of operand values may be explicit or merely a net effect of writing the result value to an upper entry of operand stack 1512 and updating pointer OPTOP to identify that entry as the top of operand stack 1512 .
  • third instruction check 2216 determines whether the instruction identified by index instr_index corresponds to an operation that the JAVA virtual machine specification (see Section I) defines as for popping a data item (or items) from the operand stack, for operating on the popped data item(s), and for pushing a result of the operation onto the operand stack. If so, index instr_index is incremented (at incrementing 2218 ) to point to the next bytecode in instruction buffer 1816 . If not, folding is not appropriate and fold determination portion 2104 supplies a nonfold-indicating signal on fold line f/nf, whereupon normal decode portion 2002 provides decoding.
  • fold determination portion 2104 determines whether the instruction identified by index instr_index, i.e., either the third or fourth bytecode, corresponds to an operation that pops a value from operand stack 1512 and stores the value in a data store such as local variable storage 1513 .
  • fourth instruction check 2220 determines whether the instruction identified by index instr_index corresponds to an operation that the JAVA virtual machine specification (see Section I) defines as for popping the result data item from the operand stack. If so, index signal dest_adr_mind is asserted (at destination address setting 2222 ) to identify the destination of the result value of the equivalent folded operation.
  • index signal dest_adr_mind is asserted (at destination address setting 1824 ) to identify the top of operand stack 1512 .
  • index signal dest_adr_mind is asserted (at destination address setting 1824 ) to identify the pointer OPTOP, whether the top of operand stack 1512 or a store operation destination is selected, the folded instruction valid signal f_valid is asserted (at valid fold asserting 1826 ) and a fold-indicating signal on line f/nf is supplied to select fold decode inputs of switches 2006 , 2008 , and 2010 for supply to execution unit 1820 .
  • Fold determination portion 2104 ends an instruction decode cycle at finish 2250 .
  • an instruction decoder for hardware processor 100 may limit fold decoding to instruction sequences of two instructions and/or to sequences of single bytecode instructions.
  • instruction decoder 135 may limit fold decoding to instruction sequences of two instructions and/or to sequences of single bytecode instructions.
  • FIG. 23 shows fold address generator 2102 including three component address generators, first operand address generator 2302 , second operand address generator 2304 , and destination address generator 2306 , respectively supplying a corresponding first operand, second operand, and destination address based on indices supplied thereto and pointer VARS and pointer OPTOP values from pointer registers 1522 .
  • first operand address generator 2302 , second operand address generator 2304 , and destination address generator 2306 supply addresses in randomly-accessible storage 1510 corresponding to a subset of operand stack 1512 and local variable storage 1513 entries.
  • Alternative embodiments may supply identifiers selective for storage other than random access memory, e.g., physical registers, which in a particular JAVA virtual machine implementation provide underlying operand stack and local variable storage.
  • First operand address generator 2302 receives first operand index signal first_adr_mind from fold determination portion 2104 and, using pointer VARS and pointer OPTOP values from pointer registers 1522 , generates a first operand address signal first_op_adr for a first operand for the equivalent folded operation.
  • the operation of second operand address generator 2304 and destination address generator 2306 is analogous.
  • Second operand address generator 2304 receives first operand index signal first_adr_ind and generates a second operand address signal second_op_adr for a second operand (if any) for the equivalent folded operation.
  • Destination address generator 2306 receives the destination index signal dest_ad_ind and generates the destination address signal dest_adr for the result of the equivalent folded operation.
  • first operand address signal first_op —adr, second operand address signal second _op_adr, and destination address signal dest_adr are collectively supplied to switch 2010 as fold address signal f_adr for supply to execution unit 1820 as the first operand, second operand, and destination addresses for the equivalent folded operation.
  • FIG. 24 illustrates an exemplary embodiment of first operand address generator 2302 .
  • Second operand address generator 2304 and destination address generator 2306 are analogous.
  • first operand address signal first_op_adr is selected from a subset of locations in local variable storage 1513 and operand stack 1512 .
  • Alternative embodiments may generate operand and destination addresses from a larger, smaller, or different subset of operand stack 1512 and local variable storage 1513 locations or from a wider range of locations in randomly-accessible storage 1510 .
  • alternative embodiments may generate addresses selective for location in constant area 1514 . Suitable modifications to the exemplary embodiment of FIG. 24 will be apparent to those of skill in the art.
  • First operand address generator 2302 , second operand address generator 2304 , and destination address generator 2306 may advantageously define differing sets of locations. For example, whereas locations in constant area 1514 and in the interior of operand stack 1512 are valid as operand sources, they are not typically appropriate result targets. For this reason, the set of locations provided by an exemplary embodiment of destination address generator 2306 is restricted to local variable storage 1513 entries and uppermost entries of operand stack 1512 , although alternative sets are also possible. Referring to FIG. 24 , pointer OPTOP is supplied to register 2402 , which latches the value and provides the latched value to a first input of a data selector 2450 .
  • pointer OPTOP is supplied to registers 2404 and 2406 , which latch the value minus one and minus two, respectively, and provide the latched values to second and third inputs of data selector 2450 .
  • addresses identified by values OPTOP, OPTOP-1, and OPTOP-2 are available for selection by data selector 2450 .
  • pointer VARS is supplied to a series of registers 2408 , 2410 , 2412 and 2414 , which respectively latch the values VARS, VARS+1, VARS+2, and VARS+3 for provision to the fourth, fifth, sixth, and seventh inputs of data selector 2450 .
  • addresses identified by values VARS, VARS+1, VARS+2, and VARS+3 are available for selection by data selector 2450 .
  • offsets from pointer VARS are positive because local variable storage 1513 is addressed from its base (identified by pointer VARS) while offsets to pointer OPTOP are negative because operand stack 1512 is addressed from its top (identified by pointer OPTOP).
  • Data selector 2450 selects from among the latched addresses available at its inputs.
  • load source addresses in local variable storage 1513 other than those addressed by values VARS, VARS+1, VARS+2, and VARS+3 are handled as unfoldable and decoded via normal decode portion 2002 .
  • Second operand address generator 2304 and destination address generator 2306 are of analogous design, although destination address generator 2306 does not provide support for addressing into constant area 1514 .
  • signal RS 1 _D is supplied to the zeroth input of data selector 2450 .
  • additional decode logic allows for direct supply of register identifier information to support an alternate instruction set.
  • Addition decode logic support for such an alternate instruction set is described in greater detail in a U.S. Pat. No. 5,925,123 entitled “A PROCESSOR FOR EXECUTING INSTRUCTION SETS RECEIVED FROM A NETWORK OR FROM A LOCAL MEMORY” naming Marc Tremblay and James Michael O'Connor as inventors, assigned to the assignee of this application, and filed on Jan. 23, 1997, the detailed description of which is incorporated herein by reference.
  • fold determination portion 2104 of fold decode portion 2004 when fold determination portion 2104 of fold decode portion 2004 identifies a foldable bytecode sequence, fold determination portion 2104 asserts a fold-indicating signal on line f/nf, supplies an equivalent folded operation as folded instruction decode signal f_instr_dec, and supplies, based on load and store instructions from the foldable bytecode sequence, indices into latched addresses maintained by first operand address generator 2302 , second operand address generator 2304 , and destination address generator 2306 .
  • Fold decode portion 2004 supplies the addresses so indexed as folded address signal f-adr. Responsive to the signal on line tinf, switches 2006 , 2008 , 2010 supply decode information for the equivalent folded operation to execution unit 1820 .
  • fold decode portion 2004 has been described above in the context of an exemplary four instruction foldable sequence, it is not limited thereto. Based on the description herein, those of skill in the art will appreciate suitable extensions to support folding of additional instructions and longer foldable instruction sequences, e.g., sequences of five or more instructions. By way of example and not of limitation, support for additional operand address signals, e.g., a third operand address signal, and/or for additional destination address signals, e.g., a second destination address signal, could be provided.
  • additional operand address signals e.g., a third operand address signal
  • additional destination address signals e.g., a second destination address signal
  • Integer execution unit IEU that includes instruction decode unit 130 , integer unit 142 , and stack management unit 150 , is responsible for the execution of all the virtual machine instructions except the floating point related instructions.
  • the floating point related instructions are executed in floating point unit 143 .
  • Integer execution unit IEU interacts at the front end with instructions cache unit 120 to fetch instructions, with floating point unit (FPU) 143 to execute floating point instructions, and finally with data cache unit (ECU) 160 to execute load and store related instructions.
  • Integer execution unit IEU also contains microcode ROM 141 which contains instructions to execute certain virtual machine instructions associated with integer operations.
  • Integer execution unit IEU includes a cached portion of stack 400 , i.e., stack cache 155 .
  • Stack cache 155 provides fast storage for operand stack and local variable entries associated with a current method, e.g., operand stack 423 and local variable storage 421 entries.
  • stack cache 155 may provide sufficient storage for all operand stack and local variable entries associated with a current method, depending on the number of operand stack and local variable entries, less than all of local variable entries or less than all of both local variable entries and operand stack entries may be represented in stack cache 155 .
  • additional entries e.g., operand stack and or local variable entries for a calling method, may be represented in stack cache 155 if space allows.
  • Stack cache 155 is a sixty-four entry thirty-two-bit wide array of registers that is physically implemented as a register file in one embodiment.
  • Stack cache 155 has three read ports, two of which are dedicated to integer execution unit IEU and one to dribble manager unit 151 .
  • Stack cache 155 also has two write ports, one dedicated to integer execution unit IEU and one to dribble manager unit 151 .
  • Integer unit 142 maintains the various pointers, which are used to access variables, such as local variables, and operand stack values, in stack cache 155 . Integer unit 142 also maintains pointers to detect whether a stack cache hit has taken place. Runtime exceptions are caught and dealt with by exception handlers that are implemented using information in microcode ROM 141 and circuit 170 .
  • Integer unit 142 contains a 32-bit ALU to support arithmetic operations.
  • the operations supported by the ALU include: add, subtract, shift, and, or, exclusive or, compare, greater than, less than, and bypass.
  • the ALU is also used to determine the address of conditional branches while a separate comparator determines the outcome of the branch instruction.
  • the most common set of instructions which executes cleanly through the pipeline is the group of ALU instructions.
  • the ALU instructions read the operands from the top of stack 400 in decode stage 302 and use the ALU in execution stage 303 to compute the result.
  • the result is written back to stack 400 in write-back stage 305 .
  • a shifter is also present as part of the ALU. If the operands are not available for the instruction in decode stage 302 , or at a maximum at the beginning of execution stage 303 , an interlock holds the pipeline stages before execution stage 303 .
  • the instruction cache unit interface of integer execution unit IEU is a valid/accept interface, where instruction cache unit 120 delivers instructions to instruction decode unit 130 in fixed fields along with valid bits.
  • Instruction decoder 135 responds by signaling how much byte aligner circuit 122 needs to shift, or how many bytes instruction decode unit 130 could consume in decode stage 302 .
  • the instruction cache unit interface also signals to instruction cache unit 120 the branch mis-predict condition, and the branch address in execution stage 303 . Traps, when taken, are also similarly indicated to instruction cache unit 120 .
  • Instruction cache unit 120 can hold integer unit 142 by not asserting any of the valid bits to instruction decode unit 130 .
  • Instruction decode unit 130 can hold instruction cache unit 120 by not asserting the shift signal to byte aligner circuit 122 .
  • the data cache interface of integer execution unit IEU also is a valid-accept interface, where integer unit 142 signals, in execution stage 303 , a load or store operation along with its attributes, e.g., non-cached, special stores etc., to data cache controller 161 in data cache unit 160 .
  • Data cache unit 160 can return the data on a load, and control integer unit 142 using a data control unit hold signal. On a data cache hit, data cache-unit 160 returns the requested data, and then releases the pipeline.
  • integer unit 142 On store operations, integer unit 142 also supplies the data along with the address in execution stage 303 .
  • Data cache unit 160 can hold the pipeline in cache stage 304 if data cache unit 160 is busy, e.g., doing a line fill etc.
  • Instruction decoder 135 fetches and decodes floating point unit 143 related instructions. Instruction decoder 135 sends the floating point operation operands for execution to floating point unit 142 in decode state 302 . While floating point unit 143 is busy executing the floating point operation, integer unit 142 halts the pipeline and waits until floating point unit 143 signals to integer unit 142 that the result is available.
  • a floating point ready signal from floating point unit 143 indicates that execution stage 303 of the floating point operation has concluded.
  • the result is written back into stack cache 155 by integer unit 142 , Floating point load and stores are entirely handled by integer execution unit IEU, since the operands for both floating point unit 143 and integer unit 142 are found in stack cache 155 .
  • a stack management unit 150 stores information, and provides operands to execution unit 140 .
  • Stack management unit 150 also takes care of overflow and underflow conditions of stack cache 155 .
  • stack management unit 150 includes stack cache 155 that, as described above, a three read port, two write port register file in one embodiment; a stack control unit 152 which provides the necessary control signals for two read ports and one write port that are used to retrieve operands for execution unit 140 and for storing data back from a write-back register or data cache 165 into stack cache 155 ; and a dribble manager 151 which speculatively dribbles data in and out of stack cache 155 into memory whenever there is an overflow or underflow in stack cache 155 .
  • memory includes data cache 165 and any memory storage interfaced by memory interface unit 110 .
  • memory includes any suitable memory hierarchy including caches, addressable read/write memory storage, secondary storage, etc.
  • Dribble manager 151 also provides the necessary control signals for a single read port and a single write port of stack cache 155 which are used exclusively for background dribbling purposes.
  • stack cache 155 is managed as a circular buffer which ensures that the stack grows and shrinks in a predictable manner to avoid overflows or overwrites.
  • the saving and restoring of values to and from data cache 165 is controlled by dribbler manager 151 using high- and low-water marks, in one embodiment.
  • Stack management unit 150 provides execution unit 140 with two 32-bit operands in a given cycle. Stack management unit 150 can store a single 32-bit result in a given cycle.
  • Dribble manager 151 handles spills and fills of stack cache 155 by speculatively dribbling the data in and out of stack cache 155 from and to data cache 165 .
  • Dribble manager 151 generates a pipeline stall signal to stall the pipeline when a stack overflow or underflow condition is detected.
  • Dribble manager 151 also keeps track of requests sent to data cache unit 160 .
  • a single request to data cache unit 160 is a 32-bit consecutive load or store request.
  • stack cache 155 The hardware organization of stack cache 155 is such that, except for long operands (long integers and double precision floating-point numbers), implicit operand fetches for opcodes do not add latency to the execution of the opcodes.
  • the number of entries in operand stack 423 ( FIG. 4A ) and local variable storage 421 that are maintained in stack cache 155 represents a hardware/performance tradeoff. At least a few operand stack 423 and local variable storage 421 entries are required to get good performance. In the exemplary embodiment of FIG. 1 , at least the top three entries of operand stack 423 and the first four local variable storage 421 entries are preferably represented in stack cache 155 .
  • stack cache 155 ( FIG. 1 ) is to emulate a register file where access to the top two registers is always possible without extra cycles.
  • a small hardware stack is sufficient if the proper intelligence is provided to load/store values from/to memory in the background, therefore preparing stack cache 155 for incoming virtual machine instructions.
  • stack 400 As indicated above, all items on stack 400 (regardless of size) are placed into a 32-bit word. This tends to waste space if many small data items are used, but it also keeps things relatively simple and free of lots of tagging or muxing.
  • An entry in stack 400 thus represents a value and not a number of bytes.
  • Long integer and double precision floating-point numbers require two entries. To keep the number of read and write ports low, two cycles to read two long integers or two double precision floating point numbers are required.
  • the mechanism for filling and spilling the operand stack from stack cache 155 out to memory by dribble manager 151 can assume one of several alternative forms.
  • One register at a time can be filled or spilled, or a block of several registers filled or spilled at once.
  • a simple scoreboarded method is appropriate for stack management.
  • a single bit indicates if the register in stack cache 155 is currently valid.
  • some embodiments of stack cache 155 use a single bit to indicate whether the data content of the register is saved to stack 400 , i.e., whether the register is dirty.
  • a high-water mark/low-water mark heuristic determines when entries are saved to and restored from stack 400 , respectively ( FIG. 4A ).
  • stack management unit 150 when the top-of-the-stack becomes close to bottom 401 of stack cache 155 by a fixed, or alternatively, a programmable number of entries, the hardware starts loading registers from stack 400 into stack cache 155 .
  • stack management unit 150 and dribble manager unit 151 are described below and in U.S. patent application Ser. No. 08/787,736, entitled “STACK MANAGEMENT UNIT AND METHOD FOR A PROCESSOR HAVING A STACK” naming Marc Tremblay and James Michael O'Connor as inventors, assigned to the assignee of this application, and filed on Jan. 23, 1997 with Attorney Docket No. SP2037, now U.S. Pat. No. 6,038,643, which is incorporated herein by reference in its entirety.
  • stack management unit 150 also includes an optional local variable look-aside cache 153 .
  • Cache 153 is most important in applications where both the local variables and operand stack 423 ( FIG. 4A ) for a method are not located on stack cache 155 . In such instances when cache 153 is not included in hardware processor 100 , there is a miss on stack cache 155 when a local variable is accessed, and execution unit 140 accesses data cache unit 160 , which in turn slows down execution. In contrast, with cache 153 , the local variable is retrieved from cache 153 and there is no delay in execution.
  • FIG. 4D One embodiment of local variable, look-aside cache 153 is illustrated in FIG. 4D for method 0 to 2 on stack 400 .
  • Local variables zero to M, where M is an integer, for method 0 are stored in plane 421 A_ 0 of cache 153 and plane 421 A_ 0 is accessed when method number 402 is zero.
  • Local variables zero to N, where N is an integer, for method 1 are stored in plane 421 A_ 1 of cache 153 and plane 421 A_ 1 is accessed when method number 402 is one.
  • Local variables zero to P, where P is an integer, for method 1 are stored in plane 421 A_ 2 of cache 153 and plane 421 A_ 2 is accessed when method number 402 is two. Notice that the various planes of cache 153 may be different sizes, but typically each plane of the cache has a fixed size that is empirically determined.
  • a new plane 421 A_ 2 in cache 153 is loaded with the local variables for that method, and method number register 402 , which in one embodiment is a counter, is changed, e.g., incremented, to point to the plane of cache 153 containing the local variables for the new method.
  • method number register 402 which in one embodiment is a counter, is changed, e.g., incremented, to point to the plane of cache 153 containing the local variables for the new method.
  • the local variables are ordered within a plane of cache 153 so that cache 153 is effectively a direct-mapped cache.
  • the variable is accessed directly from the most recent plane in cache 153 , i.e., the plane identified by method number 402 .
  • method number register 402 is changed, e.g., decremented, to point at previous plane 421 A_ 1 of cache 153 .
  • Cache 153 can be made as wide and as deep as necessary.
  • Data cache unit 160 manages all requests for data in data cache 165 .
  • Data cache requests can come from dribbling manager 151 or execution unit 140 .
  • Data cache controller 161 arbitrates between these requests giving priority to the execution unit requests.
  • data cache controller 161 In response to a request, data cache controller 161 generates address, data and control signals for the data and tags RAMs in data cache 165 . For a data cache hit, data cache controller 161 reorders the data RAM output to provide the right data.
  • Data cache controller 161 also generates requests to I/O bus and memory interface unit 110 in case of data cache misses, and in case of non-cacheable loads and stores. Data cache controller 161 provides the data path and control logic for processing non-cacheable requests, and the data path and data path control functions for handling cache misses.
  • data cache unit 160 For data cache hits, data cache unit 160 returns data to execution unit 140 in one cycle for loads. Data cache unit 160 also takes one cycle for write hits. In case of a cache miss, data cache unit 160 stalls the pipeline until the requested data is available from the external memory. For both non-cacheable loads and stores, data cache 165 is bypassed and requests are sent to I/O bus and memory interface unit 110 . Non-aligned loads and stores to data cache 165 trap in software.
  • Data cache 165 is a two-way set associative, write back, write allocate, 16-byte line cache.
  • the cache size is configurable to 0, 1, 2, 4, 8, 16 Kbyte sizes. The default size is 6 Kbytes.
  • Each line has a cache tag store entry associated with the line. On a cache miss, 16 bytes of data are written into cache 165 from external memory.
  • Each data cache tag contains a 20-bit address tag field, one valid bit, and one dirty bit. Each cache tag is also associated with a least recently used bit that is used for replacement policy. To support multiple cache sizes, the width of the tag fields also can be varied. If a cache enable bit in processor service register is not set, loads and stores are treated like non-cacheable instructions by data cache controller 161 .
  • a single sixteen-byte write back buffer is provided for writing back dirty cache lines which need to be replaced.
  • Data cache unit 160 can provide a maximum of four bytes on a read and a maximum of four bytes of data can be written into cache 165 in a single cycle. Diagnostic reads and writes can be done on the caches.
  • data cache unit 160 includes a memory allocation accelerator 166 .
  • memory allocation accelerator 166 When a new object is created, fields for the object are fetched from external memory, stored in data cache 165 and then the field is cleared to zero. This is a time consuming process that is eliminated by memory allocation accelerator 166 .
  • memory allocation accelerator 166 When a new object is created, no fields are retrieved from external memory. Rather, memory allocation accelerator 166 simply stores line of zeros in data cache 165 and marks that line of data cache 165 as dirty.
  • Memory allocation accelerator 166 is particularly advantageous with a write-back cache. Since memory allocation accelerator 166 eliminates the external memory access each time a new object is created, the performance of hardware processor 100 is enhanced.
  • Floating point unit (FPU) 143 includes a microcode sequencer, input/output section with input/output registers, a floating point adder, i.e., an ALU, and a floating point multiply/divide unit.
  • the microcode sequencer controls the microcode flow and microcode branches.
  • the input/output section provides the control for input/output data transactions, and provides the input data loading and output data unloading registers. These registers also provide intermediate result storage.
  • the floating point adder ALU includes the combinatorial logic used to perform the floating point adds, floating point subtracts, and conversion operations.
  • the floating point multiply/divide unit contains the hardware for performing multiply/divide and remainder.
  • Floating point unit 143 is organized as a microcoded engine with a 32-bit data path. This data path is often reused many times during the computation of the result. Double precision operations require approximately two to four times the number of cycles as single precision operations.
  • the floating point ready signal is asserted one-cycle prior to the completion of a given floating point operation. This allows integer unit 142 to read the floating point unit output registers without any wasted interface cycles. Thus, output data is available for reading one cycle after the floating point ready signal is asserted.
  • JAVA Virtual Machine Specification of Section I is hardware independent, the virtual machine instructions are not optimized for a particular general type of processor, e.g., a complex instruction-set computer (CISC) processor, or a reduced instruction set computer (RISC) processor. In fact, some virtual machine instructions have a CISC nature and others a RISC nature. This dual nature complicates the operation and optimization of hardware processor 100 .
  • CISC complex instruction-set computer
  • RISC reduced instruction set computer
  • the JAVA virtual machine specification defines opcode 171 for an instruction lookupswitch, which is a traditional switch statement.
  • the datastream to instruction cache unit 320 includes an opcode 171 , identifying the N-way switch statement, that is followed zero to three bytes of padding. The number of bytes of padding is selected so that first operand byte begins at an address that is a multiple of four.
  • datastream is used generically to indicate information that is provided to a particular element, block, component, or unit.
  • a first operand in the first pair is the default offset for the switch statement that is used when the argument, referred to as an integer key, or alternatively, a current match value, of the switch statement is not equal to any of the values of the matches in the switch statement.
  • the second operand in the first pair defines the number of pairs that follow in the datastream.
  • Each subsequent operand pair in the datastream has a first operand that is a match value, and a second operand that is an offset. If the integer key is equal to one of the match values, the offset in the pair is added to the address of the switch statement to define the address to which execution branches. Conversely, if the integer key is unequal to any of the match values, the default offset in the first pair is added to the address of the switch statement to define the address to which execution branches. Direct execution of this virtual machine instruction requires many cycles.
  • look-up switch accelerator 145 is included in hardware processor 100 .
  • Look-up switch accelerator 145 includes an associative memory which stores information associated with one or more lookup switch statements. For each lookup switch statement, i.e., each instruction lookupswitch, this information includes a lookup switch identifier value, i.e., the program counter value associated with the lookup switch statement, a plurality of match values and a corresponding plurality of jump offset values.
  • Lookup switch accelerator 145 determines whether a current instruction received by hardware processor 100 corresponds to a lookup switch statement stored in the associative memory. Lookup switch accelerator 145 further determines whether a current match value associated with the current instruction corresponds with one of the match values stored in the associative memory. Lookup switch accelerator 145 accesses a jump offset value from the associative memory when the current instruction corresponds to a lookup switch statement stored in the memory and the current match value corresponds with one of the match values stored in the memory wherein the accessed jump offset value corresponds with the current match value.
  • Lookup switch accelerator 145 further includes circuitry for retrieving match and jump offset values associated with a current lookup switch statement when the associative memory does not already contain the match and jump offset values associated with the current lookup switch statement.
  • Lookup switch accelerator 145 is described in more detail in U.S. patent application Ser. No. 08/788,811, entitled “LOOK-UP SWITCH ACCELERATOR AND METHOD OF OPERATING SAME” naming Marc Tremblay and James Michael O'Connor as inventors, assigned to the assignee of this application, and filed on Jan. 23, 1997 with Attorney Docket No. SP2040, now U.S. Pat. No. 6,076,141, which is incorporated herein by reference in its entirety.
  • execution unit 140 accesses a method vector to retrieve one of the method pointers in the method vector, i.e., one level of indirection. Execution unit 140 then uses the accessed method pointer to access a corresponding method, i.e., a second level of indirection.
  • each object is provided with a dedicated copy of each of the methods to be accessed by the object.
  • Execution unit 140 then accesses the methods using a single level of indirection. That is, each method is directly accessed by a pointer which is derived from the object. This eliminates a level of indirection, which was previously introduced by the method pointers.
  • the acceleration of execution unit 140 by reducing the levels of indirection experienced by execution unit 140 is described in more detail in U.S.
  • TLB translation lookaside buffer
  • JAVA virtual machine specification defines an instruction putfield, opcode 181 , that upon execution sets a field in an object and an instruction getfield, opcode 180 , that upon execution fetches a field from an object.
  • the opcode is followed by an index byte one and an index byte two.
  • Operand stack 423 contains a reference to an object followed by a value for instruction putfield, but only a reference to an object for instruction getfield.
  • Index bytes one and two are used to generate an index into the constant pool of the current class.
  • the item in the constant pool at that index is a field reference to a class name and a field name.
  • the item is resolved to a field block pointer which has both the field width, in bytes, and the field offset, in bytes.
  • An optional getfield-putfield accelerator 146 in execution unit 140 stores the field block pointer for instruction getfield or instruction putfield in a cache, for use after the first invocation of the instruction, along with the index used to identify the item in the constant pool that was resolved into the field block pointer as a tag. Subsequently, execution unit 140 uses index bytes one and two to generate the index and supplies the index to getfield-putfield accelerator 146 . If the index matches one of the indexes stored as a tag, i.e., there is a hit, the field block pointer associated with that tag is retrieved and used by execution unit 140 . Conversely, if a match is not found, execution unit 140 performs the operations described above. Getfield-putfield accelerator 146 is implemented without using self-modifying code that was used in one embodiment of the quick instruction translation described above.
  • getfield-putfield accelerator 146 includes an associative memory that has a first section that holds the indices that function as tags, and a second section that holds the field block pointers. When an index is applied through an input section to the first section of the associative memory, and there is a match with one of the stored indices, the field block pointer associated with the stored index that matched in input-index is output from the second section of the associative memory.
  • Bounds check unit 147 ( FIG. 1 ) in execution unit 140 is an optional hardware circuit that checks each access to an element of an array to determine whether the access is to a location within the array. When the access is to a location outside the array, bounds check unit 147 issues an active array bound exception signal to execution unit 140 . In response to the active array bound exception signal, execution unit 140 initiates execution of an exception handler stored in microcode ROM 141 that in handles the out of bounds array access.
  • bounds check unit 147 includes an associative memory element in which is stored a array identifier for an array, e.g., a program counter value, and a maximum value and a minimum value for the array.
  • a array identifier for an array e.g., a program counter value
  • a maximum value and a minimum value for the array e.g., the array identifier for that array is applied to the associative memory element, and assuming the array is represented in the associative memory element
  • the stored minimum value is a first input signal to a first comparator element, sometimes called a comparison element
  • the stored maximum value is a first input signal to a second comparator element, sometimes also called a comparison element.
  • a second input signal to the first and second comparator elements is the value associated with the access of the array's element.
  • bounds check unit 147 A more detailed description of one embodiment of bounds check unit 147 is provided in U.S. patent application Ser. No. 08/786,352, entitled “PROCESSOR WITH ACCELERATED ARRAY ACCESS BOUNDS CHECKING” naming Marc Tremblay, James Michael O'Connor, and William N. Joy as inventors, assigned to the assignee of this application, and filed on Jan. 23, 1997 with Attorney Docket No. SP2041, now U.S. Pat. No. 6,014,723, which is incorporated herein by reference in its entirety.
  • the JAVA Virtual Machine Specification defines that certain instructions can cause certain exceptions.
  • the checks for these exception conditions are implemented, and a hardware/software mechanism for dealing with them is provided in hardware processor 100 by information in microcode ROM 141 and program counter and trap control logic 170 .
  • the alternatives include having a trap vector style or a single trap target and pushing the trap type on the stack so that the dedicated trap handler routine determines the appropriate action.
  • FIG. 5 illustrates several possible add-ons to hardware processor 100 to create a unique system. Circuits supporting any of the eight functions shown, i.e., NTSC encoder 501 , MPEG 502 , Ethernet controller 503 , VIS 504 , ISDN 505 , I/O controller 506 , ATM assembly/reassembly 507 , and radio link 508 can be integrated into the same chip as hardware processor 100 of this invention.
  • FIG. 6 is a block diagram of one embodiment of a stack management unit 150 .
  • Stack management unit 150 serves as a high speed buffer between stack 400 and hardware processor 100 .
  • Hardware processor 100 accesses stack management unit 150 as if stack management unit 150 were stack 400 .
  • Stack management unit 150 automatically transfers data between stack management unit 150 and stack 400 as necessary to provide improve the throughput of data between stack 400 and hardware processor 100 .
  • data cache unit 160 retrieves the requested data word and places the requested data word at the top of stack cache 155 .
  • Stack management unit 150 contains a stack cache, memory circuit 610 .
  • Stack cache memory circuit 610 is typically fast memory devices such as a register file or SRAM; however, slower memory devices such as DRAM can also be used.
  • access to stack cache memory circuit 610 is controlled by stack control unit 152 .
  • a write port 630 allows hardware processor 100 to write data on data lines 635 to stack cache memory circuit 610 .
  • Read port 640 and read port 650 allow hardware processor 100 to read data from stack cache memory circuit 610 on data lines 645 and 655 , respectively. Two read ports are provided to increase throughput since many operations of stack-based computing systems require two operands from stack 400 .
  • Other embodiments of stack cache 155 may provide more or less read and write ports.
  • Dribble manager unit 151 controls the transfer of data between stack 400 ( FIG. 4 ( a )) and stack cache memory circuit 610 .
  • the transfer of data between stack 400 and stack cache memory circuit 610 goes through data cache unit 160 .
  • Dribble manager unit 151 includes a fill control unit 694 and a spill control unit 698 .
  • fill control unit 694 and spill control unit 698 function independently.
  • Fill control unit 694 determines if a fill condition exists. If the fill condition exists, fill control unit 694 transfers data words from stack 400 to stack cache memory circuit 610 on data lines 675 through a write port 670 .
  • Spill control unit 698 determines if a spill condition exists. If the spill condition exists, spill control unit 698 transfers data words from stack cache memory circuit 610 to stack 400 through read port 680 on data lines 685 . Write port 670 and read port 680 allows transfers between stack 400 and stack cache memory circuit 610 to occur simultaneously with reads and writes controlled by stack control unit 152 . If contention for read and write ports of stack cache memory circuit 610 is not important, dribble manager unit 151 can share read and write ports with stack control unit 152 .
  • stack management unit 150 is described in the context of buffering stack 400 for hardware processor 100 , stack management unit 150 can perform caching for any stack-based computing system.
  • the details of hardware processor 100 are provided only as an example of one possible stack-based computing system for use with the present invention. Thus, one skilled in the art can use the principles described herein to design a stack management unit in accordance to the present invention for any stack-based computing system.
  • FIG. 7 shows a conceptual model of the memory architecture of stack cache memory circuit 610 for one embodiment of stack cache 155 .
  • stack cache memory circuit 610 is a register file organized in a circular buffer memory architecture capable of holding 64 data words. Other embodiments may contain a different number of data words.
  • the circular memory architecture causes data words in excess of the capacity of stack cache memory circuit 610 to be written to previously used registers. If stack cache memory unit 610 uses a different memory device, such as an SRAM, different registers would correspond to different memory locations.
  • One technique to address registers in a circular buffer is to use pointers containing modulo stack cache size (modulo-SCS) addresses to the various registers of stack cache memory circuit 610 .
  • modulo-N operations have the results of the standard operation mapped to a number between 0 and SCS-1 using a standard MOD N function. Some common modulo operations are defined as follows
  • FIG. 7 One embodiment of the pointer addresses of the registers of stack cache memory circuit 610 are shown in FIG. 7 as numbered 0-63 along the outer edge of stack cache memory circuit 610 .
  • 70 data words (numbered 1 to 70) are written to stack cache memory circuit 610 when stack cache memory circuit 610 is empty, data words 1 to 64 are written to registers 0 to 63 , respectively and data words 65 to 70 are written subsequently to registers 0 to 5 .
  • dribble manager unit 151 Prior to writing data words 65 to 70 , dribble manager unit 151 , as described below, transfers data words 1 to 6 which were in registers 0 to 5 to stack 400 .
  • data words 1 to 6 can be retrieved from stack 400 and placed in memory locations 0 to 5 .
  • a pointer OPTOP contains the location of the top of stack 400 , i.e. the top memory location.
  • pointer OPTOP is a programmable register in execution unit 140 .
  • stack management unit 150 maintain pointer OPTOP in stack control unit 152 . Since pointer OPTOP is often increased by one, decreased by one, or changed by a specific amount, pointer OPTOP, in one embodiment is a programmable up/down counter.
  • pointer OPTOP indicates the register of stack cache memory circuit 610 containing the most recently written data word in stack cache memory circuit 610 , i.e. pointer OPTOP points to the register containing the most recently written data word also called the top register.
  • Some embodiments of stack management unit 150 also contains a pointer OPTOPI (not shown) which points to the register preceding the register pointed to by pointer OPTOP. Pointer OPTOPI can improve the performance of stack management unit 150 since many operations in hardware processor 100 require two data words from stack management unit 150 .
  • Pointer OPTOP and pointer OPTOP1 are incremented whenever a new data word is written to stack cache 155 .
  • Pointer OPTOP and pointer OPTOP1 are decremented whenever a stacked data word, i.e. a data word already in stack 400 , is popped off of stack cache 155 . Since some embodiments of hardware processor 100 may add or remove multiple data words simultaneously, pointer OPTOP and OPTOP1 are implemented, in one embodiment as programmable registers so that new values can be written into the registers rather than requiring multiple increment or decrement cycles.
  • pointer OPTOP1 may also be implemented using a modulo SCS subtractor which modulo-SCS subtracts one from pointer OPTOP.
  • Some embodiments of stack cache 155 may also include pointers OPTOP2 or pointer OPTOP3.
  • stack cache memory circuit 610 Since data words are stored in stack cache memory circuit 610 circularly, the bottom of stack cache memory circuit 610 can fluctuate. Therefore, most embodiments of stack cache memory circuit 610 include a pointer CACHE_BOTTOM to indicate the bottom memory location of stack cache memory circuit 610 .
  • Pointer CACHE_BOTTOM is typically maintained by dribble manager unit 151 . The process to increment or decrement pointer CACHE_BOTTOM varies with the specific embodiment of stack management unit 150 .
  • Pointer CACHE_BOTTOM is typically implemented as a programmable up/down counter.
  • stack management unit 150 also includes other pointers, such as pointer VARS, which points to a memory location of a data word that is often accessed.
  • pointer VARS pointer VARS
  • the method frames often contain local variables that are accessed frequently. Therefore, having pointer.
  • VARS pointed to the first local variable of the active method decreases the access time necessary to read the local variable.
  • Other pointers such as a pointer VARS 1 (not shown) and a pointer VARS 2 (not shown) may point to other often used memory locations such as the next two local variables of the active method in a JAVA Virtual Machine.
  • pointer VARS is stored in a programmable register in execution unit 140 . If stack cache 155 is organized using sequential addressing, pointer VARS 1 may also be implemented using a modulo-SCS adder which modulo-SCS adds one to pointer VARS.
  • stack management unit 150 To determine which data words to transfer between stack cache memory circuit 610 and stack 400 , stack management unit 150 , typically tags, i.e. tracks, the valid data words and the data words which are stored in both stack cache memory circuit 610 and stack 400 .
  • FIG. 8 illustrates one tagging scheme used in some embodiments of stack management unit 150 . Specifically, FIG. 8 shows a register 810 from stack cache memory circuit 610 . The actual data word is stored in data section 812 . A valid bit 814 and a saved bit 816 are used to track the status of register 810 . If valid bit 814 is at a valid logic state, typically logic high, data section 812 contains a valid data word.
  • valid bit 814 is at an invalid logic state, typically logic low, data section 812 does not contain a valid data word. If saved bit 816 is at a saved logic state, typically logic high, the data word contained in data section 812 is also stored in stack 400 . However, if saved bit 816 is at an unsaved logic state, typically logic low, the data word contained in data section 812 is not stored in stack 400 . Typically, when stack management unit 150 is powered up or reset, valid bit 814 of each register is set to the invalid logic state and saved bit 816 of each register is set to the unsaved logic state.
  • stack control unit 152 when stack control unit 152 writes a data word to a register in stack cache memory circuit 610 through write port 630 the valid bit of that register is set to the valid logic state and the saved bit of that register is set to the unsaved logic state.
  • dribble manager unit 151 transfer a data word to a register of stack cache memory circuit 610 through write port 670 , the valid bit of that register is set to the valid logic state and the saved bit of that register is set to the saved logic state since the data word is currently saved in stack 400 .
  • stack cache 155 is generally much smaller than the memory address space of hardware processor 100
  • the pointers used to access stack cache memory circuit 610 are generally much smaller than general memory addresses.
  • the specific technique used to map stack cache 155 into the memory space of hardware processor 100 can vary.
  • the pointers used to access stack cache memory circuit 610 are only the lower bits of general memory pointers, i.e, the least significant bits. For example, if stack cache memory circuit 610 comprises 64 registers, pointers OPTOP, VARS, and CACHE_BOTTOM need only be six bits long. If hardware processor 100 has a 12 bit address space, pointers OPTOP, VARS, and CACHE_BOTTOM could be the lower-six bits of a general memory pointer. Thus stack cache memory circuit 610 is mapped to a specific segment of the address space having a unique upper six bit combination.
  • stack cache management unit 150 may be used with purely stacked based computing system so that there is not a memory address space for the system. In this situation, the pointers for accessing stack cache 155 are only internal to stack cache management unit 150 .
  • stack management unit 150 can improve data accesses of hardware processor 100 while only caching the top portion of stack 400 .
  • hardware processor 100 pushes more data words to stack management unit 150 than stack cache memory circuit 610 is able to store, the data words near the bottom of stack cache memory circuit 610 are transferred to stack 400 .
  • hardware processor 100 pops data words out of stack cache 155 , data words from stack 400 are copied under the bottom of stack cache memory circuit 610 , and pointer CACHE_BOTTOM is decremented to point to the new bottom of stack cache memory circuit 610 .
  • dribble manager unit 151 should transfer data from stack cache memory circuit 610 to stack 400 , i.e. a spill operation, as hardware processor fills stack cache memory circuit 610 . Conversely, dribble manager unit 151 should copy data from stack 400 to stack cache memory circuit 610 , i.e. a fill operation, as hardware processor empties stack cache memory circuit 610 .
  • FIG. 9 shows one embodiment of dribble manager unit 151 in which decisions on transferring data from stack cache memory circuit 610 to stack 400 , i.e. spilling data, are based on the number of free registers in stack cache memory circuit 610 .
  • Free registers includes registers without valid data as well as registers containing data already stored in stack 400 , i.e. registers with saved bit 816 set to the saved logic state.
  • Decisions on transferring data from stack 400 to stack cache memory circuit 610 i.e. filling data, are based on the number of used registers.
  • a used registers contains a valid but unsaved data word in stack cache memory circuit 610 .
  • dribble manager unit 151 further includes a stack cache status circuit 910 and a cache bottom register 920 , which can be a programmable up/down counter.
  • Stack cache status circuit 910 receives pointer CACHE_BOTTOM from cache bottom register 920 and pointer OPTOP to determine the number of free registers FREE and the number of used registers USED.
  • stack cache status circuit 910 can be implemented with a modulo SCS adder/subtractor.
  • the number of used registers USED and the number of free registers FREE can also be generated using a programmable up/down counters. For example, a used register can be incremented whenever a data word is added to stack cache 155 and decremented whenever a data word is removed from stack cache 155 . Specifically, if pointer OPTOP is modulo-SCS incremented by some amount, the used register is incremented by the same amount. If pointer OPTOP is modulo-SCS decremented by some amount, the used register is decremented by the same amount.
  • pointer CACHE_BOTTOM is modulo-SCS incremented by some amount
  • the used register is decremented by the same amount. If pointer CACHE_BOTTOM is modulo-SCS. decremented by some amount, the used register is incremented the same amount.
  • the number of free registers FREE can be generated by subtracting the number of used registers USED from the total number of registers.
  • Spill control unit 694 ( FIGS. 6 and 9 ) includes a cache high threshold register 930 and a comparator 940 .
  • Comparator 940 compares the value in cache high threshold register 930 to the number of free registers FREE. If the number of free registers FREE is less than the value in cache high threshold register 930 , comparator 940 drives a spill signal SPILL to a spill logic level, typically logic high, to indicate that the spill condition exists and one or more data words should be transferred from stack cache memory circuit 610 to stack 400 , i.e. a spill operation should be performed. The spill operation is described in more detail below.
  • cache high threshold register 930 is programmable by hardware processor 100 .
  • Fill control unit 698 includes a cache low threshold register 950 and a comparator 960 .
  • Comparator 960 compares the value in cache low threshold register 950 to the number of used registers USED. If the number of used registers is less than the value in cache low threshold register 950 , comparator 960 drives a fill signal FILL to a fill logic level, typically logic high, to indicate that the fill condition exists and one or more data words should be transferred from stack 400 to stack cache memory circuit 610 , i.e. a fill operation should be performed. The fill operation is described in more detail below.
  • cache low threshold register 950 is programmable by hardware processor 100 .
  • a single cache threshold register can be used.
  • Fill control unit 698 can be modified to use the number of free registers FREE to drive signal FILL to the fill logic level if then number of free registers is greater than the value in cache low threshold 950 , with a proper modification of the value in cache low threshold 950 .
  • spill control unit 694 can be modified to use the number of used registers.
  • FIG. 10A shows another embodiment of dribble manager unit 151 , which uses a high-water mark/low water mark heuristic to determine when a spill condition or a fill condition exists.
  • Spill control unit 694 includes a high water mark register 1010 implemented as a programmable up/down counter.
  • a comparator 1020 in spill control unit 694 compares the value in high water mark register 1010 , i.e. the high water mark, with pointer OPTOP. If pointer OPTOP is greater than the high water mark, comparator 1020 drives spill signal SPILL to the spill logic level to indicate a spill operation should be performed.
  • the high water mark is relative to pointer CACHE_BOTTOM, the high water mark is modulo-SCS incremented and modulo-SCS decremented whenever pointer CACHE_BOTTOM is modulo-SCS incremented or modulo-SCS decremented, respectively.
  • Fill control unit 698 includes a low water mark register 1010 implemented as a programmable up/down counter.
  • a comparator 1030 in fill control unit 698 compares the value in low water mark register 1030 , i.e. the low water mark, with pointer OPTOP. If pointer OPTOP is less than the low water mark, comparator 1040 drives fill signal FILL to the fill logic level to indicate a fill operation should be performed. Since the low water mark is relative to pointer CACHE_BOTTOM, the low water mark register is modulo-SCS incremented and modulo-SCS decremented whenever pointer CACHE_BOTTOM is modulo-SCS incremented or modulo-SCS decremented, respectively.
  • FIG. 10B shows an alternative circuit to generate the high water mark and low water mark.
  • Cache high threshold register 930 typically implemented as a programmable register, contains the number of free registers which should be maintained in stack cache memory circuit 610 .
  • the high water mark is then calculated by modulo-SCS subtractor 1050 by modulo-SCS subtracting the value in cache high threshold register 930 from pointer CACHE_BOTTOM stored in cache bottom register 920 .
  • the low water mark is calculated by doing a modulo-SCS addition.
  • cache low threshold register 950 is programmed to contain the minimum number of used data registers desired to be maintained in stack cache memory circuit 610 .
  • the low water mark is then calculated by modulo-SCS adder 1060 by modulo-SCS adding the value in cache low threshold register 950 with pointer CACHE_BOTTOM stored in cache bottom register 920 .
  • a spill operation is the transfer of one or more data words from stack cache memory circuit 610 to stack 400 .
  • the transfers occurs though data cache unit 160 .
  • the specific interface between stack management unit 150 and data cache unit 160 can vary.
  • stack management unit 150 and more specifically dribble manager unit 151 , sends the data word located at the bottom of stack cache 155 , as indicated by pointer CACHE_BOTTOM from read port 680 to data cache unit 160 .
  • pointer CACHE_BOTTOM is also provided to data cache unit 160 so that data cache unit 160 can address the data word appropriately.
  • the saved bit of the register indicated by pointer CACHE_BOTTOM is set to the saved logic level.
  • pointer CACHE_BOTTOM is modulo-SCS incremented by one.
  • Other registers as described above may also be modulo-SCS incremented by one.
  • high water mark register 1010 FIG. 10A
  • low water mark 1030 would be modulo-SCS incremented by one.
  • Some embodiments of dribble manager unit 157 . transfer multiple words for each spill operation.
  • pointer CACHE_BOTTOM is modulo-SCS incremented by the number words transferred to stack 400 .
  • pointer CACHE_BOTTOM is at the saved logic level, the data word in that data register is already stored in stack 400 . Therefore, the data word in that data register does not need to be copied to stack 400 .
  • pointer CACHE_BOTTOM is still modulo-SCS incremented by one.
  • a fill operation transfers data words from stack 400 to stack cache memory circuit 610 .
  • the transfers occurs though data cache unit 160 .
  • the specific interface between stack management unit 150 and data cache unit 160 can vary.
  • stack management unit 150 and more specifically dribble manager unit 151 , determines whether the data register preceding the data register pointed by CACHE_BOTTOM is free, i.e. either the saved bit is in the saved logic state or the valid bit is in the invalid logic state. If the data register preceding the data register pointed to by pointer CACHE_BOTTOM is free, dribble manager unit 151 requests a data word from stack 400 by sending a request with the value of pointer CACHE_BOTTOM modulo-SCS minus one.
  • pointer CACHE_BOTTOM When the data word is received from data cache unit 160 , pointer CACHE_BOTTOM is modulo-SCS decremented by one and the received data word is written to the data register pointed to by pointer CACHE_BOTTOM through write port 670 .
  • Other registers as described above may also be modulo-SCS decremented.
  • the saved bit and valid bit of the register pointed to by pointer CACHE_BOTTOM are set to the saved logic state and valid logic state, respectively.
  • Some embodiments of dribble manager unit 151 transfer multiple words for each spill operation. For these embodiments, pointer CACHE_BOTTOM is modulo-SCS decremented by the number words transferred to stack 400 .
  • pointer CACHE_BOTTOM is at the saved logic level and the valid logic level, respectively, then the data word in that data register was never overwritten. Therefore, the data word in that data register does not need to be copied from stack 400 .
  • pointer CACHE_BOTTOM is still modulo-SCS decremented by one.
  • FIG. 11 shows a circuit to generate pointers for a typical operation which reads two data words from stack cache 155 and writes one data word to stack cache 155 .
  • the most common stack manipulation for a stack-based computing system is to pop the top two data words off of the stack and to push a data word onto the top of the stack. Therefore, the circuit of FIG. 11 is configured to be able to provide read pointers to the value of pointer OPTOP and the value of pointer OPTOP modulo-SCS minus one, and a write pointer to the current value of OPTOP modulo-SCS minus one.
  • Multiplexer (MUX) 1110 drives a read pointer RP 1 for read port 640 .
  • a select line RS 1 controlled by hardware processor 100 determines whether multiplexer 1110 drives the same value as pointer OPTOP or a read address R_ADDR 1 as provided by hardware processor 100 .
  • Multiplexer 1120 provides a read pointer RP 2 for read port 650 .
  • Modulo adder 1140 modulo-SCS adds negative one to the value of pointer OPTOP and drives the resulting sum to multiplexer 1120 .
  • a select line RS 2 controlled by hardware processor 100 determines whether multiplexer 1120 drives the value from modulo adder 1140 or a read address R_ADDR 2 as provided by hardware processor 100 .
  • Multiplexer 1130 provides a write pointer WP for write port 630 .
  • a modulo adder 1150 modulo-SCS adds one to the value of pointer OPTOP and drives the resulting sum to multiplexer 1130 .
  • Select lines WS controlled by hardware processor 100 determines whether multiplexer 1130 drives the value from modulo-SCS adder 1140 , the value from modulo-SCS adder 1150 , or a write address W_ADDR as provided by hardware processor 100 .
  • FIG. 12 shows a circuit that generates a read pointer R for read port 640 or read port 650 in embodiments allowing accessing stack cache memory circuit using pointer VARS.
  • Multiplexer 1260 drives read pointer R to one of several input values received on input ports 1261 - 1267 as determined by selection signals RS. Selection signals RS are controlled by hardware processor 100 .
  • the value of pointer OPTOP is driven to input port 1261 .
  • Modulo-SCS adder 1210 drives the modulo-SCS sum of the value of pointer OPTOP with negative one to input port 1262 .
  • Modulo-SCS adder 1210 drives the modulo-SCS sum of the value of pointer OPTOP with negative two to input port 1263 .
  • the value of pointer VARS is driven to input port 1264 .
  • Modulo-SCS adder 1230 drives the modulo-SCS sum of the value of pointer VARS with one to input port 1265 .
  • Modulo-SCS adder 1240 drives the modulo-SCS sum of the value of pointer VARS with two to input port 1266 .
  • Modulo adder-SCS 1250 drives the modulo-SCS sum of the value of pointer VARS with three to input port 1263 .
  • Other embodiments may provide other values to the input ports of multiplexer 1260 .
  • a dribbling management unit can efficiently control transfers between the stack cache and the stack. Specifically, the dribbling management unit is able to transfer data out of the stack cache to make room for additional data as necessary and transfer data into the stack cache as room becomes available transparently to the stack-based computing system using the stack management unit.
  • This BETA quality release and related documentation are protected by copyright and distributed under licenses restricting its use, copying, distribution, and decompilation. No part of this release or related documentation may be reproduced in any form by any means without prior written authorization of Sun and its licensors, if any.
  • Portions of this product may be derived from the UNIX® and Berkeley 4.3 BSD systems, licensed from UNIX System Laboratories, Inc. and the University of California, respectively. Third-party font software in this release is protected by copyright and licensed from Sun's Font Suppliers.
  • Sun, Sun Microsystems, Sun Microsystems Computer Corporation, the Sun logo, the Sun Microsystems Computer Corporation logo, WebRunner, JAVA, FirstPerson and the FirstPerson logo and agent are trademarks or registered trademarks of Sun Microsystems, Inc.
  • the “Duke” character is a trademark of Sun Microsystems, Inc. and Copyright (c) 1992-1995 Sun Microsystems, Inc. All Rights Reserved.
  • UNIX® is a registered trademark in the United States and other countries, exclusively licensed through X/Open Company, Ltd.
  • OPEN LOOK is a registered trademark of Novell, Inc. All other product names mentioned herein are the trademarks of their respective owners.
  • SPARC trademarks including the SCD Compliant logo, are trademarks or registered trademarks of SPARC International, Inc.
  • SPARCstation, SPARCserver, SPARCengine, SPARCworks, and SPARCompiler are licensed exclusively to Sun Microsystems, Inc; Products bearing SPARC trademarks are based upon an architecture developed by Sun Microsystems, Inc.
  • the OPEN LOOK® and SunTM Graphical User Interfaces were developed by Sun Microsystems, Inc. for its users and licensees. Sun acknowledges the pioneering efforts of Xerox in researching and developing the concept of visual or graphical user interfaces for the computer industry. Sun holds a non-exclusive license from Xerox to the Xerox Graphical User Interface, which license also covers Sun's licensees who implement OPEN LOOK GUIs and otherwise comply with Sun's written license agreements.
  • X Window System is a trademark and product of the Massachusetts Institute of Technology.
  • This document describes version 1.0 of the JAVA Virtual Machine and its instruction set. We have written this document to act as a specification for both compiler writers, who wish to target the machine, and as a specification for others who may wish to implement a compliant JAVA Virtual Machine.
  • the JAVA Virtual Machine is an imaginary machine that is implemented by emulating it in software on a real machine.
  • Code for the JAVA Virtual Machine is stored in class files, each of which contains the code for at most one public class.
  • Simple and efficient emulations of the JAVA Virtual Machine are possible because the machine's format is compact and efficient bytecodes. Implementations whose native code speed approximates that of compiled C are also possible, by translating the bytecodes to machine code, although Sun has not released such implementations at this time.
  • the virtual machine data types include the basic data types of the JAVA language:
  • the virtual machine doesn't have separate instructions for boolean types. Instead, integer instructions, including integer returns, are used to operate on boolean values; byte arrays are used for arrays of boolean.
  • the virtual machine specifies that floating point be done in IEEE 754 format, with support for gradual underflow. Older computer architectures that do not have support for IEEE format may run JAVA numeric programs very slowly.
  • an object reference is to a handle, which is a pair of pointers: one to a method table for the object, and the other to the data allocated for the object.
  • Other implementations may use inline caching, rather than method table dispatch such methods are likely to be faster on hardware that is emerging between now and the year 2000.
  • Programs represented by JAVA Virtual Machine bytecodes are expected to maintain proper type discipline and an implementation may refuse to execute a bytecode program that appears to violate such type discipline.
  • JAVA Virtual Machines would appear to be limited by the bytecode de ⁇ nition to running on a 32-bit address space machine, it is possible to build a version of the JAVA Virtual Machine that automatically translates the bytecodes into a 64-bit form. A description of this transformation is beyond the scope of the JAVA Virtual Machine Specification.
  • the virtual machine is executing the code of a single method, and the pc register contains the address of the next bytecode to be executed.
  • Each method has memory space allocated for it to hold:
  • Each JAVA method uses a fixed-sized set of local variables. They are addressed as word offsets from the vary register. Local variables are all 32 bits wide.
  • Long integers and double precision floats are considered to take up two local variables but are addressed by the index of the first local variable. (For example, a local variable with index containing a double precision float actually occupies storage at indices n and n+1.)
  • the virtual machine specifcation does not require 64-bit values in local variables to be 64-bit aligned. Implementor are free to decide the appropriate way to divide long integers and double precision floats into two words.
  • Instructions are provided to load the values of local variables onto the operand stack and store values from the operand stack into local variables.
  • the machine instructions all take operands from an operand stack, operate on them, and return results to the stack.
  • the operand stack is 32 bits wide. It is used to pass parameters to methods and receive method results, as well as to supply parameters for operations and save operation results.
  • execution of instruction iadd adds two integers together. It expects that the two integers are the top two words on the operand stack, and were pushed there by previous instructions. Both integers are popped from the stack, added, and their sum pushed back onto the operand stack. Subcomputations may be nested on the operand stack, and result in a single operand that can be used by the is nesting computation.
  • Each primitive data type has specialized instructions that know how to operate on operands of that type.
  • Each operand requires a single location on the stack, except for long and double operands, which require two locations.
  • Operands must be operated on by operators appropriate to their type. It is illegal, for example, to push two integers and then treat them as a long. This restriction is enforced, in the Sun implementation, by the bytecode verifier. However, a small number of operations (the dup opcodes and swap) operate on runtime data areas as raw values of a given width without regard to type.
  • the information contained in the execution environment is used to do dynamic linking, normal method returns, and exception propagation.
  • the execution environment contains references to the interpreter symbol table for the current method and current class, in support of dynamic linking of the method code.
  • the class file code for a method refers to methods to-be called and variables to be accessed symbolically. Dynamic linking translates these symbolic method calls into actual method calls, loading classes as necessary to resolve as-yet-undefined symbols, and translates variable accesses into appropriate offsets in storage structures associated with the runtime location of these variables.
  • the execution environment is used in this case to restore the registers of the caller, with the program counter of the caller appropriately incremented to skip the method call instruction. Execution then continues in the calling method's execution environment.
  • the execution environment may be extended with additional implementation-specified information, such as debugging information.
  • the JAVA heap is the runtime data area from which class instances (objects) are allocated.
  • the JAVA language is designed to be garbage collected—it does not, give the programmer the ability to deallocate objects explicitly.
  • the JAVA language does not presuppose any particular kind of garbage collection; various algorithms may be used depending on system requirements.
  • the method area is analogous to the store for compiled code in conventional languages or the text segment in a UNIX process. It stores method code (compiled JAVA code) and symbol tables. In the current JAVA implementation, method code is not part of the garbage-collected heap, although this is planned for a future release.
  • An instruction in the JAVA instruction set consists of a one-byte opcode specifying the operation to be performed, and zero or more operands supplying parameters or data that will be used by the operation. Many instructions have no operands and consist only of an opcode.
  • the inner loop of the virtual machine execution is effectively: do ⁇ fetch an opcode byte execute an action depending on the value of the opcode ⁇ while (there is more to do);
  • the number and size of the additional operands is determined by the opcode. If an additional operand is more than one byte in size, then it is stored in big-endian order—high order byte first. For example, a 16-bit parameter is stored as two bytes whose value is: first_byte*256+second_byte
  • the bytecode instruction stream is only byte-aligned, with the exception being the tableswitch and lookupswitch instructions, which force alignment to a 4-byte boundary within their instructions.
  • the per-class constant pool has a maximum of 65535 entries. This acts as an internal limit on the total complexity of a single class.
  • the amount of code per method is limited to 65535 bytes by the sizes of the indices in the code in the exception table, the line number table, and the local variable table.
  • Each class file contains the compiled version of either a JAVA class or a JAVA interface.
  • Compliant JAVA interpreters must be capable of dealing with all class files that conform to the following specification.
  • a JAVA class file consists of a stream of 8-bit bytes. All 16-bit and 32-bit quantities are constructed by reading in two or four 8-bit bytes, respectively. The bytes are joined together in network (big-endian) order, where the high bytes come first.
  • This format is supported by the JAVA JAVA.io.DataInput and JAVA.io.DataOutput interfaces, and classes such as JAVA.io.DataInputStream and JAVA.io.DataOutputStream.
  • the class file format is described here using a structure notation. Successive fields in the structure appear in the external representation without padding or alignment. Variable size arrays, often of variable sized elements, are called tables and are commonplace in these structures.
  • the types u1, u2, and u4 mean an unsigned one-, two-, or four-byte quantity, respectively, which are read by method such as readUnsignedByte, readUnsignedShort and readInt of the JAVA.io.DataInput interface.
  • This field must have the value 0 ⁇ CAFEBABE.
  • a change of the major version number indicates a major incompatible change, one that requires a different virtual machine that may not support the old major version in any way.
  • the current major version number is 45; the current minor version number is 3.
  • This field indicates the number of entries in the constant pool in the class file.
  • the constant pool is a table of values. These values are the various string constants, class names, field names, and others that are referred to by the class structure or by the code.
  • constant_pool[0] is always unused by the compiler, and may be used by an implementation for any purpose.
  • Each of the constant_pool entries 1 through constant_pool_count ⁇ 1 is a variable-length entry, whose format is given by the first “tag” byte, as described in section 2.3.
  • This field contains a mask of up to sixteen modifiers used with class, method, and field declarations.
  • the same encoding is used on similar fields in field_info and method_info as described below.
  • Flag Name Value Meaning Used By ACC_PUBLIC 0x0001 Visible to everyone Class, Method, Variable ACC_PRIVATE 0x0002 Visible only to the Method, defining class Variable ACC_PROTECTED 0x0004 Visible to Method, subclasses Variable ACC_STATIC 0x0008 Variable or method Method is static Variable ACC_FINAL 0x0010 No further sub- Class, classing, over- Method, riding, or assign- Variable ment after ini- tialization ACC_SYNCHRONIZED 0x0020 Wrap use in monitor Method lock ACC_VOLATILE 0x0040 Can't cache Variable ACC_TRANSIENT 0x0080 Not to be written Variable or read by a per- sistent object manager ACC
  • This field is an index into the constant pool; constant_pool [this_class] must be a CONSTANT_class.
  • This field gives the number of interfaces that this class implements.
  • Each value in this table is an index into the constant pool. If a table value is nonzero (interfaces[i]
  • This field gives the number of instance variables, both static and dynamic, defined by this class.
  • the fields table includes only those variables that are defined explicitly by this class. It does not include those instance variables that are accessible from this class but are inherited from superclasses.
  • This field indicates the number of methods, both static and dynamic, defined by this class. This table only includes those methods that are explicitly defined by this class. It does not include inherited methods.
  • This field indicates the number of additional attributes about this class.
  • a class can have any number of optional attributes associated with it. Currently, the only class attribute recognized is the “SourceFile” attribute, which indicates the name of the source file from which this class file was compiled. See section 2.6 for more information on the attribute_info structure.
  • a signature is a string representing a type of a method, field or array.
  • ⁇ array_type> ⁇ base_type> :: B
  • the character V indicates that the method returns no value. Otherwise, the signature indicates the type of the return value.
  • a method signature represents the arguments that the method expects, and the value that it returns.
  • ⁇ method_signature> :: ( ⁇ arguments_signature>)
  • Each tag byte is then followed by one or more bytes giving more information about the specific constant.
  • CONSTANT_Class is used to represent a class or an interface.
  • CONSTANT_Class_info ⁇ u1 tag; u2 name_index; ⁇ tag
  • the opcodes anewarray and multianewarray can reference array “classes” via CONSTANT_Class items in the constant pool.
  • the name of the class is its signature. For example, the class name for
  • the tag will have the value CONSTANT_Fieldref, CONSTANT_Methodref, or CONSTANT_InterfaceMethodref.
  • constant_pool [class_index] will be an entry of type CONSTANT_Class giving the name of the class or interface containing the field or method.
  • CONSTANT_Class item For CONSTANT_Fieldref and CONSTANT_Methodref, the CONSTANT_Class item must be an actual class. For CONSTANT_InterfaceMethodref, the item must be an interface which purports to implement the given method.
  • constant_pool [name_and_type_index] will be an entry of type CONSTANT_NameAndType. This constant pool entry indicates the name and signature of the field or method.
  • CONSTANT_String is used to represent constant objects of the built-in type String.
  • CONSTANT_String_info ⁇ u1 tag; u2 string_index; ⁇ tag
  • the tag will have the value CONSTANT_String
  • CONSTANT_Integer andCONSTANT_Float represent four-byte constants. CONSTANT_Integer_info ⁇ u1 tag; u4 bytes; ⁇ CONSTANT_Float_info ⁇ u1 tag; u4 bytes; ⁇ tag
  • the four bytes are the integer value.
  • integers For integers, the four bytes are the integer value.
  • floats they are the IEEE 754 standard representation of the floating point value. These bytes are in network (high byte first) order.
  • CONSTANT_Long andCONSTANT_Double represent eight-byte constants.
  • CONSTANT_Long_info ⁇ u1 tag; u4 high_bytes; u4 low_bytes; ⁇
  • CONSTANT_Double_info ⁇ u1 tag; u4 high_bytes; u4 low_bytes; ⁇
  • the tag will have the value CONSTANT_Long or CONSTANT_Double.
  • the 64-bit value is (high_bytes ⁇ 32)+low_bytes.
  • the 64-bit value;high_bytes and low_bytes together represent the standard IEEE 754 representation of the double-precision floating point number.
  • CONSTANT_NameAndType is used to represent a field or method, without indicating which class it belongs to.
  • CONSTANT_NameAndType_info ⁇ u1 tag; u2 name_index; u2 signature_index; ⁇ tag
  • the tag will have the valueCONSTANT_NameAndType.
  • constant_pool [name_index] is a CONSTANT_Utf8 string giving the name of the field or method.
  • constant_pool [signature_index] is a CONSTANT_Utf8 string giving the signature of the field or method.
  • CONSTANT_Utf8 andCONSTANT_Unicode are used to represent constant string values.
  • CONSTANT_Utf8 strings are “encoded” so that strings containing only non-null ASCII characters, can be represented using only one byte per character, but characters of up to 16 bits can be represented:
  • null character (0 ⁇ 0000) and characters in the range 0 ⁇ 0080 to 0 ⁇ 07FF are represented by a pair of two bytes:
  • Each field is described by a variable length field_info structure.
  • the format of this structure is as follows: field_info ⁇ u2 access_flags; u2 name_index; u2 signature_index; u2 attributes_count; attribute_info attributes[attribute_count]; ⁇ access_flags
  • the possible fields that can be set for a field are ACC_PUBLIC, ACC_PRIVATE, ACC_PROTECTED, ACC_STATIC, ACC_FINAL, ACC_VOLATILE, and ACC_TRANSIENT.
  • At most one of ACC_PUBLIC, ACC_PROTECTED, and ACC_PRIVATE can be set for any method.
  • constant_pool [name_index] is a CONSTANT_Utf8 string which is the name of the field.
  • signature_index is a CONSTANT_Utf8 string which is the signature of the field. See the section “Signatures” for more information on signatures.
  • This value indicates the number of additional attributes about this field.
  • a field can have any number of optional attributes associated with it.
  • the only field attribute recognized is the “ConstantValue” attribute, which indicates that this field is a static numeric constant, and indicates the constant value of that field.
  • method_info ⁇ u2 access_flags; u2 name_index; u2 signature_index; u2 attributes_count; attribute_info attributes [attribute_count]; ⁇ access_flags
  • the possible fields that can be set for a method are ACC_PUBLIC, ACC_PRIVATE, ACC_PROTECTED, ACC_STATIC, ACC_FINAL, ACC_SYNCHRONIZED, ACC_NATIVE, and ACC_ABSTRACT.
  • At most one of ACC_PUBLIC, ACC_PROTECTED, and ACC_PRIVATE can be set for any method.
  • constant_pool[name_index] is a CONSTANT_Utf8 string giving the name of the method.
  • signature_index is a CONSTANT_Utf8 string giving the signature of the field. See the section “Signatures” for more information on signatures.
  • This value indicates the number of additional attributes about this field.
  • a field can have any number of optional attributes associated with it. Each attribute has a name, and other additional information. Currently, the only field attributes recognized are the “Code” and “Exceptions” attributes, which describe the bytecodes. that are executed to perform this method, and the JAVA Exceptions which are declared to result from the execution of the method, respectively.
  • Attributes are used at several different places in the class format. All attributes have the following format: GenericAttribute_info ⁇ u2 attribute_name; u4 attribute_length; u1 info[attribute_length]; ⁇
  • the attribute_name is a 16-bit index into the class's constant pool; the value of constant_pool [attribute_name] is a CONSTANT_Utf8 string giving the name of the attribute.
  • the field attribute length indicates the length of the subsequent information in bytes. This length does not include the six bytes of the attribute_name and attribute_length.
  • SourceFile_attribute ⁇ u2 attribute_name_index; u4 attribute_length; u2 sourcefile_index; ⁇ attribute_name_index
  • SourceFile_attribute The length of a SourceFile_attribute must be 2.
  • constant_pool [sourcefile_index] is a CONSTANT_Utf8 string giving the source file from which this class file was compiled.
  • ConstantValue_attribute ⁇ u2 attribute_name_index; u4 attribute_length; u2 constantvalue_index; ⁇ attribute_name_index
  • ConstantValue_attribute The length of a ConstantValue_attribute must be 2.
  • the constant pool entry must be of a type appropriate to the field, as shown by the following table: long CONSTANT_Long float CONSTANT_Float double CONSTANT_Double int, short, char, byte, boolean CONSTANT_Integer 2.6.3 Code
  • the “Code” attribute has the following format: Code_attribute ⁇ u2 attribute_name_index; u4 attribute_length; u2 max_stack; u2 max_locals; u4 code_length; u1 code[code_length]; u2 exception_table_length; ⁇ u2 start_pc; u2 end_pc; u2 handler_pc; u2 catch_type; ⁇ exception_table[exception_table_length]; u2 attributes_count; attribute_info attributes [attribute_count]; ⁇ attribute_name_index
  • This field indicates the total length of the “Code” attribute, excluding the initial six bytes.
  • Each entry in the exception table describes one exception handler in the code.
  • start_pc and end_pc indicate the ranges in the code at which the exception handler is active.
  • the values of both fields are offsets from the start of the code.start_pc is inclusive.end_pc is exclusive.
  • This field indicates the starting address of the exception handler.
  • the value of the field is an offset from the start of the code.
  • constant_pool [catch_type] will be the class of exceptions that this exception handler is designated to catch. This exception handler should only be called if the thrown exception is an instance of the given class.
  • catch_type is zero, this exception handler should be called for all exceptions.
  • This field indicates the number of additional attributes about code.
  • the “Code” attribute can itself have attributes.
  • a “Code” attribute can have any number of optional attributes associated with it. Each attribute has a name, and other additional information. Currently, the only code attributes defined are the “LineNumberTable” and “LocalVariableTable,” both of which contain debugging information.
  • This field indicates the total length of the Exceptions_attribute, excluding the initial six bytes.
  • This field indicates the number of entries in the following exception index table.
  • the LineNumberTable_attribute has the following format: LineNumberTable_attribute ⁇ u2 attribute_name_index; u4 attribute_length; u2 line_number_table_length; ⁇ u2 start_pc; u2 line_number; ⁇ line_number_table[line — number_table_length]; ⁇ attribute_name_index
  • This field indicates the total length of the LineNumberTable_attribute, excluding the initial six bytes.
  • This field indicates the number of entries in the following line number table.
  • Each entry in the line number table indicates that the line number in the source file changes at a given point in the code.
  • source_pc ⁇ SHOULD THAT BEstart_pc?>> is an offset from the beginning of the code.
  • This attribute is used by debuggers to determine the value of a given local variable during the dynamic execution of a method.
  • the format of the LocalVariableTable_attribute is as follows: LocalVariableTable_attribute ⁇ u2 attribute_name_index; u4 attribute_length; u2 local_variable_table_length; ⁇ u2 start_pc; u2 length; u2 name_index; u2 signature_index; u2 slot; ⁇ local_variable_table[local — variable_table_length]; ⁇ attribute_name_index
  • This field indicates the total length of the LineNumberTable_attribute, excluding the initial six bytes.
  • This field indicates the number of entries in the following local variable table.
  • Each entry in the local variable table indicates a code range during which a local variable has a value. It also indicates where on the stack the value of that variable can be found.
  • the given local variable will have a value at the code between start_pc andstart_pc+length.
  • the two values are both offsets from the beginning of the code.
  • constant_pool [name_index] and constant_pool [signature_index] are CONSTANT_Utf8 strings giving the name and signature of the local variable.
  • the given variable will be the slot th local variable in the method's frame.
  • JAVA Virtual Machine instructions are represented in this document by an entry of the following form.
  • Each line in the syntax table represents a single 8-bit byte.
  • Instruction iload has the short description “Load integer from local variable.” Implicitly, the integer is loaded onto the stack. Instruction iadd is described as “Integer add”; both its source and destination are the stack.
  • Instructions that do not affect the control flow of a computation may be assumed to always advance the virtual machine program counter to the opcode of the following instruction. Only instructions that do affect control flow will explicitly mention the effect they have on the program counter.
  • byte1 is interpreted as a signed 8-bitvalue. This value is expanded to an integer and pushed onto the operand stack.
  • byte1 and byte2 are assembled into a signed 16-bit value. This value is expanded to an integer and pushed onto the operand stack.
  • indexbyte1 is used as an unsigned 8-bit index into the constant pool of the current class. The item at that index is resolved and pushed onto the stack. If a String is being pushed and there isn't enough memory to allocate space for it then an OutOfMemoryError is thrown.
  • indexbyte1 and indexbyte2 are used to construct an unsigned 16-bit index into the constant pool of the current class. The item at that index is resolved and pushed onto the stack. If a String is being pushed and there isn't enough memory to allocate space for it then an OutOfMemoryError is thrown.
  • constant-word1, constant-word2 indexbyte1 and indexbyte2 are used to construct an unsigned 16-bit index into the constant pool of the current class.
  • the two-word constant that index is resolved and pushed onto the stack.
  • This instruction is the same as iload with a vindex of ⁇ n>, except that the operand ⁇ n> is implicit.
  • This instruction is the same as lload with a vindex of ⁇ n>, except that the operand ⁇ n> is implicit.
  • This instruction is the same as fload with a vindex of ⁇ n>, except that the operand ⁇ n> is implicit.
  • This instruction is the same as dload with a vindex of ⁇ n>, except that the operand ⁇ on> is implicit.
  • This instruction is the same as aload with a vindex of ⁇ n>, except that the operand ⁇ n> is implicit.
  • This instruction is the same as istore with a vindex of ⁇ n>, except that the operand ⁇ n> is implicit.
  • This instruction is the same as fstore with a vindex of ⁇ n>, except that the operand ⁇ n> is implicit.
  • This instruction is the same as dstore with a vindex of ⁇ n>, except that the operand ⁇ n> is implicit.
  • This instruction is the same as astore with a vindex of ⁇ n>, except that the operand ⁇ n> is implicit.
  • This bytecode must precede one of the following bytecodes: iload, lload, fload, dload, aload, istore, lstore, fstore, dstore, astore, iinc.
  • the vindex of the following bytecode and vindex2 from this bytecode are assembled into an unsigned 16-bit index to a local variable in the current JAVA frame.
  • the following bytecode operates as normal except for the use of this wider index.
  • size must be an integer. It represents the number of elements in the new array.
  • atype is an internal code that indicates the type of array to allocate. Possible values for atype are as follows: T_BOOLEAN 4 T_CHAR 5 T_FLOAT 6 T_DOUBLE 7 T_BYTE 8 T_SHORT 9 T_INT 10 T_LONG 11
  • a new array of atype, capable of holding size elements, is allocated, and result is a reference to this new object. Allocation of an array large enough to contain size items of atype is attempted. All elements of the array are initialized to zero.
  • indexbyte1 and indexbyte2 are used to construct an index into the constant pool of the current class. The item at that index is resolved. The resulting entry must be a class.
  • a new array of the indicated class type and capable of holding size elements is allocated, and result is a reference to this new object. Allocation of an array large enough to contain size items of the given class type is attempted. All elements of the array are initialized to null.
  • anewarray is used to create a single dimension of an array of object references. For example, to create
  • anewarray can also be used to create the first dimension of a multi-dimensional array.
  • array declaration :
  • Each size must be an integer. Each represents the number of elements in a dimension of the array.
  • indexbyte1 and indexbyte2 are used to construct an index into the constant pool of the current class.
  • the item at that index is resolved.
  • the resulting entry must be an array class of one or more dimensions.
  • the result is a reference to the new array object.
  • objectref must be a reference to an array object. The length of the array is determined and replaces objectref on the top of the stack.
  • arrayref must be a reference to an array of integers.index must be an integer. The integer value at position number index in the array is retrieved and pushed onto the top of the stack.
  • arrayref must be a reference to an array of long integers, index must be an integer.
  • index must be an integer. The long integer value at position number index in the array is retrieved and pushed onto the top of the stack.
  • arrayref must be a reference to an array of single-precision floating point numbers, index must be an integer.
  • index must be an integer. The single-precision floating point number value at position number index in the array is retrieved and pushed onto the top of the stack.
  • arrayref must be a reference to an array of double-precision floating point numbers.
  • index must be an integer. The double-precision floating point number value at position number index in the array is retrieved and pushed onto the top of the stack.
  • arrayref must be a reference to an array of references to objects. index must be an integer. The object reference at position number index in the array is retrieved and pushed onto the top of the stack.
  • arrayref must be a reference to an array of signed bytes.
  • index must be an integer.
  • the signed byte value at position number index in the array is retrieved, expanded to an integer, and pushed onto the top of the stack.
  • arrayref must be a reference to an array of characters. index must be an integer. The character value at position number index in the array is retrieved, zero-extended to an integer, and pushed onto the top of the stack.
  • arrayref must be a reference to an array of short integers.
  • index must be an integer.
  • the ;signed short integer value at position number index in the array is retrieved, expanded to an integer, and pushed onto the top of the stack.
  • arrayref must be a reference to an array of integers, index must be an integer, and value an integer. The integer value is stored at position index in the array.
  • arrayref must be a reference to an array of long integers, index must be an integer, and value a long integer. The long integer value is stored at position index in the array.
  • arrayref must be an array of single-precision floating point numbers, index must be an integer, and value a single-precision floating point number.
  • the single float value is stored at position index in the array.
  • arrayref must be a reference to an array of double-precision floating point numbers, index must be an integer, and value a double-precision floating point number.
  • the double float value is stored at position index in the array.
  • arrayref must be a reference to an array of references to objects, index must be an integer, and value a reference to an object.
  • the object reference value is stored at position index in the array.
  • the actual type of value must be conformable with the actual type of the elements of the array. For example, it is legal to store an instance of class Thread in an array of class Object, but not vice versa. An ArrayStoreException is thrown if an attempt is made to store an incompatible object reference.
  • arrayref must be a reference to an array of signed bytes, index must be an integer, and value an integer. The integer value is stored at position index in the array. If value is too large to be a signed byte, it is truncated.
  • arrayref must be an array of characters, index must be an integer, and value an integer.
  • the integer value is stored at position index in the array. If value is too large to be a character, it is truncated.
  • arrayref must be an array of shorts, index must be an integer, and value an integer.
  • the integer value is stored at position index in the array. If value is too large to be an short, it is truncated.
  • value1 and value 2 must be integers. The values are added and are replaced on the stack by their integer sum.
  • value1 and value 2 must be long integers. The values are added and are replaced on the stack by their long integer sum.

Abstract

An instruction decoder allows the folding away of JAVA virtual machine instructions pushing an operand onto the top of a stack merely as a precursor to a second JAVA virtual machine instruction which operates on the top of stack operand. Such an instruction decoder identifies foldable instruction sequences and supplies an execution unit with a single equivalent folded operation thereby reducing processing cycles otherwise required for execution of multiple operations corresponding to the multiple instructions of the folded instruction sequence. Instruction decoder embodiments described herein provide for folding of two, three, four, or more instruction folding. For example, in one instruction decoder embodiment described herein, two load instructions and a store instruction can be folded into execution of operation corresponding to an instruction appearing therebetween in the instruction sequence.

Description

    RELATED APPLICATIONS
  • This application is a continuation of application Ser. No. 11/096,183, filed Mar. 30, 2005, which is itself a continuation of application Ser. No. 10/346,886, filed Jan. 17, 2003, now U.S. Pat. No. 6,950,923, which is in turn a continuation of application Ser. No. 08/787,617, filed Jan. 23, 1997, now U.S. Pat. No. 6,532,531. application Ser. No. 08/787,617 is itself a continuation-in-part of application Ser. No. 08/647,103, filed May 7, 1996 and a continuation-in-part of application Ser. No. 08/642,253, filed May 2, 1996, both now abandoned. application Ser. Nos. 08/787,617, 08/647,103 and 08/642,253 each claim benefit of U.S. Provisional Application No. 60/010,527, filed Jan. 24, 1996.
  • The present application, as well as each of successive priority application Ser. Nos. 11/096,183, 10/346,886 and 08/787,617, incorporate by reference the entirety of application Ser. No. 08/786,351, filed Jan. 23, 1997, now U.S. Pat. No. 6,026,485.
  • REFERENCE SECTION I
  • A portion of the disclosure of this patent document including Section I, The JAVA Virtual Machine Specification and Section A thereto, contains material which is subject to copyright protection. The copyright owner has no objection to the facsimile reproduction by anyone of the patent document or the patent disclosure, as it appears in the U.S. Patent and Trademark Office patent files or records, but otherwise reserves all copyright rights whatsoever.
  • BACKGROUND OF THE INVENTION
  • 1. Field of the Invention
  • The present invention relates to instruction decoders for a stack machine, and in particular, to methods and apparati for folding a sequence of multiple instructions into a single folded operation.
  • 2. Discussion of Related Art
  • Many individuals and organizations in the computer and communications industries tout the Internet as the fastest growing market on the planet. In the 1990s, the number of users of the Internet appears to be growing exponentially with no end in sight. In June of 1995, an estimated 6,642,000 hosts were connected to the Internet; this represented an increase from an estimated 4,852,000 hosts in January, 1995. The number of hosts appears to be growing at around 75% per year. Among the hosts, there were approximately 120,000 networks and over 27,000 web servers. The number of web servers appears to be approximately doubling every 53 days.
  • In July 1995, with over 1,000,000 active Internet users, over 12,505 usenet news groups, and over 10,000,000 usenet readers, the Internet appears to be destined to explode into a very large market for a wide variety of information and multimedia services.
  • In addition, to the public carrier network or Internet, many corporations and other businesses are shifting their internal information systems onto an intranet as a way of more effectively sharing information within a corporate or private network. The basic infrastructure for an intranet is an internal network connecting servers and desktops, which may or may not be connected to the Internet through a firewall. These intranets provide services to desktops via standard open network protocols which are well established in the industry. Intranets provide many benefits to the enterprises which employ them, such as simplified internal information management and improved internal communication using the browser paradigm. Integrating Internet technologies with a company's enterprise infrastructure and legacy systems also leverages existing technology investment for the party employing an intranet. As discussed above, intranets and the Internet are closely related, with intranets being used for internal and secure communications within the business and the Internet being used for external transactions between the business and the outside world. For the purposes of this document, the term “networks” includes both the Internet and intranets. However, the distinction between the Internet and an intranet should be born in mind where applicable.
  • In 1990, programmers at Sun Microsystems wrote a universal programming language. This language was eventually named the JAVA programming language. (JAVA is a trademark of Sun Microsystems of Mountain View, CA.) The JAVA programming language resulted from programming efforts which initially were intended to be coded in the C++ programming language; therefore, the JAVA programming language has many commonalities with the C++ programming language. However, the JAVA programming language is a simple, object-oriented, distributed, interpreted yet high performance, robust yet safe, secure, dynamic, architecture neutral, portable, and multi-threaded language.
  • The JAVA programming language has emerged as the programming language of choice for the Internet as many large hardware and software companies have licensed it from Sun Microsystems. The JAVA programming language and environment is designed to solve a number of problems in modern programming practice. The JAVA programming language omits many rarely used, poorly understood, and confusing features of the C++ programming language. These omitted features primarily consist of operator overloading, multiple inheritance, and extensive automatic coercions. The JAVA programming language includes automatic garbage collection that simplifies the task of programming because it is no longer necessary to allocated and free memory as in the C programming language. The JAVA programming language restricts the use of pointers as defined in the C programming language, and instead has true arrays in which array bounds are explicitly checked, thereby eliminating vulnerability to many viruses and nasty bugs. The JAVA programming language includes objective-C interfaces and specific exception handlers.
  • The JAVA programming language has an extensive library of routines for coping easily with TCP/IP protocol (Transmission Control Protocol based on Internet protocol), HTTP (Hypertext Transfer Protocol) and FTP (File Transfer Protocol). The JAVA programming language is intended to be used in networked/distributed environments. The JAVA programming language enabled the construction of virus-free, tamper-free systems. The authentication techniques are based on public-key encryption.
  • Many computing systems, including those implementing the JAVA virtual machine, can execute multiple methods each of which has a method frame. Typically, method invocation significantly impacts the performance of the computing system due to the excessive number of memory accesses method invocation requires. Therefore, a method and memory architecture targeted to reduce the latency caused by method invocation is desirable.
  • SUMMARY OF THE INVENTION
  • A JAVA virtual machine is an stack-oriented abstract computing machine, which like a physical computing machine has an instruction set and uses various storage areas. A JAVA virtual machine need not understand the JAVA programming language; instead it understands a class file format. A class file includes JAVA virtual machine instructions (or bytecodes) and a symbol table, as well as other ancillary information. Programs written in the JAVA programming language (or in other languages) may be compiled to produce a sequence of JAVA virtual machine instructions.
  • Typically, in a stack-oriented machine, instructions typically operate on data at the top of an operand stack. One or more first instructions, such as a load from local variable instruction, are executed to push operand data onto the operand stack as a precursor to execution of an instruction which immediately follows such instruction(s). The instruction which follows, e.g., an add operation, pops operand data from the top of the stack, operates on the operand data, and pushes a result onto the operand stack, replacing the operand data at the top of the operand stack.
  • A suitably configured instruction decoder allows the folding away of instructions pushing an operand onto the top of a stack merely as a precursor to a second instruction which operates on the top of stack operand. The instruction decoder identifies foldable instruction sequences (typically 2, 3, or 4 instructions) and supplies an execution unit with an equivalent folded operation (typically a single operation) thereby reducing processing cycles otherwise required for execution of multiple operations corresponding to the multiple instructions of the folded instruction sequence. Using an instruction decoder in accordance with the present invention, multiple load instructions and a store instruction can be folded into execution of an instruction appearing therebetween in the instruction sequence. For example, an instruction sequence including a pair of load instructions (for loading integer operands from local variables to the top of stack), an add instruction (for popping the integer operands of the stack, adding them, and placing the result at the top of stack), and an store instruction (for popping the result from the stack and storing the result in a local variable) can be folded into a single equivalent operation specifying source and destination addresses in stack and local variable storage which are randomly accessible.
  • In accordance with an embodiment of the present invention, an apparatus includes an instruction store, an operand stack, a data store, an execution unit, and an instruction decoder. The instruction decoder is coupled to the instruction store to identify a foldable sequence of instructions represented therein. The foldable sequence includes first and second instructions, in which the first instruction is for pushing a first operand value onto the operand stack from the data store merely as a first source operand for a second instruction. The instruction decoder coupled to supply the execution unit with a single folded operation equivalent to the foldable sequence and including a first operand address identifier selective for the first operand value in the data store, thereby obviating an explicit operation corresponding to the first instruction.
  • In a further embodiment, if the sequence of instructions represented in the instruction buffer is not a foldable sequence, the instruction decoder supplies the execution unit with an operation identifier and operand address identifier corresponding to the first instruction only.
  • In another further embodiment, the instruction decoder further identifies a third instruction in the foldable sequence. This third instruction is for pushing a second operand value onto the operand stack from the data store merely as a second source operand for the second instruction. The single folded operation is equivalent to the foldable sequence and includes a second operand address identifier selective for the second operand value in the data store, thereby obviating an explicit operation corresponding to the third instruction.
  • In yet another further embodiment, the instruction decoder further identifies a fourth instruction in the foldable sequence. This fourth instruction is for popping a result of the second instruction from the operand stack and storing the result in a result location of the data store. The single folded operation is equivalent to the foldable sequence and includes a destination address identifier selective for the result location in the data store, thereby obviating an explicit operation corresponding to the fourth instruction.
  • In still yet another further embodiment, the instruction decoder includes normal and folded decode paths and switching means. The switching means are responsive to the folded decode path for selecting operation, operand, and destination identifiers from the folded decode path in response to a fold indication therefrom, and for otherwise selecting operation, operand, and destination identifiers from the normal decode path.
  • In various further alternative embodiments, the apparatus is for a virtual machine instruction processor wherein instructions generally source operands from, and target a result to, uppermost entries of an operand stack. In one such alternative embodiment, the virtual machine instruction processor is a hardware virtual machine instruction processor and the instruction decoder includes decode logic. In another, the virtual machine instruction processor includes a just-in-time compiler implementation and the instruction decoder includes software executable on a hardware processor. The hardware processor includes the execution unit. In yet another, the virtual machine instruction processor includes a bytecode interpreter implementation and the instruction decoder including software executable on a hardware processor. The hardware processor includes the execution unit.
  • In accordance with another embodiment of the present invention, a method includes (a) determining if a first instruction of a virtual machine instruction sequence is an instruction for pushing a first operand value onto the operand stack from a data store merely as a first source operand for a second instruction; and if the result of the (a) determining is affirmative, supplying an execution unit with a single folded operation equivalent to a foldable sequence comprising the first and second instructions. The single folded operation includes a first operand identifier selective for the first operand value, thereby obviating an explicit operation corresponding to the first instruction.
  • In a further embodiment, the method includes supplying, if the result of the (a) determining is negative, the execution unit with an operation equivalent to the first instruction in the virtual machine instruction sequence.
  • In another further embodiment, the method includes (b) determining if a third instruction of the virtual machine instruction sequence is an instruction for popping a result value of the second instruction from the operand stack and storing the result value in a result location of the data store and, if the result of the (b) determining is affirmative, further including a result identifier selective for the result location with the equivalent single folded operation, thereby further obviating an explicit operation corresponding to the third instruction. In a further embodiment, the method includes including, if the result of the (b) determining is negative, a result identifier selective for a top location of the operand stack with the equivalent single folded operation. In certain embodiments, the (a) determining and the (b) determining are performed substantially in parallel.
  • In accordance with yet another embodiment of the present invention, a stack-based virtual machine implementation includes a randomly-accessible operand stack representation, a randomly-accessible local variable storage representation, and a virtual machine instruction decoder for selectively decoding virtual machine instructions and folding together a selected sequence thereof to eliminate unnecessary temporary storage of operands on the operand stack.
  • In various alternative embodiments, the stack-based virtual machine implementation (1) is a hardware virtual machine instruction processor including a hardware stack cache, a hardware instruction decoder, and an execution unit or (2) includes software encoded in a computer readable medium and executable on a hardware processor. In the hardware virtual machine instruction processor embodiment, (a) the randomly-accessible operand stack local variable storage representations at least partially reside in the hardware stack cache, and (b) the virtual machine instruction decoder includes the hardware instruction decoder coupled to provide the execution unit with opcode, operand, and result identifiers respectively selective for a hardware virtual machine instruction processor operation and for locations in the hardware stack cache as a single hardware virtual machine instruction processor operation equivalent to the selected sequence of virtual machine instructions. In the software embodiment, (a) the randomly-accessible operand stack local variable storage representations at least partially reside in registers of the hardware processor, (b) the virtual machine instruction decoder is at least partially implemented in the software, and (c) the virtual machine instruction decoder is coupled to provide opcode, operand, and result identifiers respectively selective for a hardware processor operation and for locations in the registers as a single hardware processor operation equivalent to the selected sequence of virtual machine instructions.
  • In accordance with still yet another embodiment of the present invention, a hardware virtual machine instruction decoder includes a normal decode path, a fold decode path, and switching means. The fold decode path is for decoding a sequence of virtual machine instructions and, if the sequence is foldable, supplying (a) a single operation identifier, (b) one or more operand identifiers; and (c) a destination identifier, which are together equivalent to the sequence of virtual machine instructions. The switching means is responsive to the folded decode path for selecting operation, operand, and destination identifiers from the folded decode path in response to a fold indication therefrom, and otherwise selecting operation, operand, and destination identifiers from the normal decode path.
  • BRIEF DESCRIPTION OF THE DRAWINGS
  • FIG. 1 (depicted as FIGS. 1A and 1B) is a block diagram of one embodiment of a virtual machine hardware processor that includes an instruction decoder for providing instruction folding in accordance with some embodiments of the present invention.
  • FIG. 2 is an process flow diagram for generation of virtual machine instructions that are used in one embodiment of this invention.
  • FIG. 3 illustrates an instruction pipeline implemented in the hardware processor of FIG. 1.
  • FIG. 4A is an illustration of the one embodiment of the logical organization of a stack structure where each method frame includes a local variable storage area, an environment storage area, and an operand stack utilized by the hardware processor of FIG. 1.
  • FIG. 4B is an illustration of an alternative embodiment of the logical organization of a stack structure where each method frame includes a local variable storage area and an operand stack on the stack, and an environment storage area for the method frame is included on a separate execution environment stack.
  • FIG. 4C is an illustration of an alternative embodiment of the stack management unit for the stack and execution environment stack of FIG. 4B.
  • FIG. 4D is an illustration of one embodiment of the local variables look-aside cache in the stack management unit of FIG. 1.
  • FIG. 5 illustrates several possible add-one to the hardware processor of FIG. 1.
  • FIG. 6 illustrates a block diagram of one embodiment of a stack cache management unit in accordance with this invention.
  • FIG. 7 illustrates the memory architecture of one embodiment of a stack cache in accordance with this invention.
  • FIG. 8 illustrates the contents of a register or memory location of one embodiment of a stack cache in accordance with this invention.
  • FIG. 9 illustrates a block diagram of one embodiment of a dribble manager unit in accordance with this invention.
  • FIG. 10A illustrates a block diagram of another embodiment of a dribble manager unit in accordance with this invention.
  • FIG. 10B illustrates a block, diagram of another embodiment of a dribble manager unit in accordance with this invention.
  • FIG. 11 illustrates a block diagram of a portion of an embodiment of a dribble manager unit in accordance with this invention.
  • FIG. 12 illustrates an pointer generation circuit for one embodiment of a stack cache in accordance with this invention.
  • FIG. 13 is an illustration, in the context of a stack data structure, of data flows associated with a pair of stack instructions, wherein the first stack instruction pushes a data item onto the top of the stack only to be consumed by the second stack instruction which pops the top two stack entries off the stack and pushes their sum onto the top of the stack.
  • FIG. 14 is a contrasting illustration of folded execution of first and second stack instructions such as those depicted in FIG. 13, wherein the first (push data item onto the top of the stack) operation is obviated in accordance with an exemplary embodiment of the present invention.
  • FIG. 15 is a block diagram depicting relationships between operand stack, local variable storage, and constant pool portions of memory storage together with register variables for access thereof in accordance with an exemplary embodiment of the present invention.
  • FIGS. 16A-D illustrate an iload (integer load)/iadd (integer add)/istore (integer store) instruction sequence operating on an operand stack and local variable storage. FIGS. 16A, 16B, 16C, and 16D, respectively depict operand stack contents before iload instructions, after iload instructions but before an iadd instruction, after the iadd instruction but before an istore instruction, and after the istore instruction. Intermediate stages depicted in FIGS. 16B and 16C are eliminated by instruction folding in accordance with an exemplary embodiment of the present invention.
  • FIGS. 17A, 17B and 17C illustrate an aload (object reference load)/arraylength (integer add) instruction sequence operating on the operand stack and local variable storage. FIGS. 17A, 17B, and 17C, respectively depict operand stack contents before the load instruction, after the aload instruction but before an arraylength instruction (without instruction folding), and after the arraylength instruction. The intermediate stage depicted in FIG. 17B is eliminated by instruction folding in accordance with an exemplary embodiment of the present invention.
  • FIG. 18 is a functional block diagram of a stack based processor including an instruction decoder providing instruction folding in accordance with an exemplary embodiment of the present invention.
  • FIG. 19 is a functional block diagram depicting an instruction decoder in accordance with an exemplary embodiment of the present invention and coupled to supply an execution unit with a folded operation, with operand addresses into an operand stack, local variable storage or a constant pool, and with a destination address into the operand stack or local variable storage, wherein the single operation and addresses supplied are equivalent to a sequence of unfolded instructions.
  • FIG. 20 is a functional block diagram of an instruction decoder supporting instruction folding in accordance with an exemplary embodiment of the present invention.
  • FIG. 21 is a functional block diagram of a fold decode portion of an instruction decoder supporting instruction folding in accordance with an exemplary embodiment of the present invention.
  • FIG. 22 is a flow chart depicting an exemplary sequence of operations for identifying a foldable instruction sequence in accordance with an exemplary embodiment of the present invention.
  • FIG. 23 is a functional block diagram depicting component operand and destination address generators of a fold address generator in accordance with an exemplary embodiment of the present invention.
  • FIG. 24 is a functional block diagram depicting an exemplary structure for an operand address generator in accordance with an exemplary embodiment of the present invention.
  • These and other features and advantages of the present invention will be apparent from the Figures, as explained in the Detailed Description of the Invention. Like or similar features are designated by the same reference numeral(a) throughout the drawings and the Detailed Description of the Invention.
  • DETAILED DESCRIPTION OF THE INVENTION
  • FIG. 1 illustrates one embodiment of a virtual machine instruction hardware processor 100, hereinafter hardware processor 100, that includes an instruction decoder 135 for folding a sequence of multiple instructions into a single folded operation in accordance with some embodiments of the present invention, and that directly executes virtual machine instructions that are processor architecture independent. The performance of hardware processor 100 in executing JAVA virtual machine instructions is much better than high-end CPUs, such as the Intel PENTIUM microprocessor or the Sun Microsystems ULTRASPARC processor, (ULTRASPARC is a trademark of Sun Microsystems of Mountain View, Calif.), and PENTIUM is a trademark of Intel Corp. of Sunnyvale, Calif.) interpreting the same virtual machine instructions with a software JAVA interpreter, or with a JAVA just-in-time compiler; is low cost; and exhibits low power consumption. As a result, hardware processor 100 is well suited for portable applications. Hardware processor 100 provides similar advantages for other virtual machine stack-based architectures as well as for virtual machines utilizing features such as garbage collection, thread synchronization, etc.
  • In view of these characteristics, a system based on hardware processor 100 presents attractive price for performance characteristics, if not the best overall performance, as compared with alternative virtual machine execution environments including software interpreters and just-in-time compilers. Nonetheless, the present invention is not limited to virtual machine hardware processor embodiments, and encompasses any suitable stack-based, or non-stack-based machine implementations, including implementations emulating the JAVA virtual machine as a software interpreter, compiling JAVA virtual machine instructions (either in batch or just-in-time) to machine instruction native to a particular hardware processor, or providing hardware implementing the JAVA virtual machine in microcode, directly in silicon, or in some combination thereof.
  • Regarding price for performance characteristics, hardware processor 100 has the advantage that the 250 Kilobytes to 500 Kilobytes (Kbytes) of memory storage, e.g., read-only memory or random access memory, typically required by a software interpreter, is eliminated.
  • A simulation of hardware processor 100 showed that hardware processor 100 executes virtual machine instructions twenty times faster than a software interpreter running on a variety of applications on a PENTIUM processor clocked at the same clock rate as hardware processor 100, and executing the same virtual machine instructions. Another simulation of hardware processor 100 showed that hardware processor 100 executes virtual machine instructions five times faster than a just-in-time compiler running on a PENTIUM processor running at the same clock rate as hardware processor 100, and executing the same virtual machine instructions.
  • In environments in which the expense of the memory required for a software virtual machine instruction interpreter is prohibitive, hardware processor 100 is advantageous. These applications include, for example, an Internet chip for network appliances, a cellular telephone processor, other telecommunications integrated circuits, or other low-power, low-cost applications such as embedded processors, and portable devices.
  • The present invention increases the speed of method invocation by using an execution environment memory 440 in conjunction with stack 400B. The execution environment of various method calls are stored in execution environment memory 440 while the operands, variables and parameters of the method calls are stored in stack 400B. Both execution environment memory 440 and stack 400B can include a stack management unit 150 that utilizes a stack cache 155 to accelerate data transfers for execution unit 140. Although, stack management unit 150 can be an integral part of hardware processor 100 as shown in FIG. 1, many embodiments of stack management unit 150 are not integrated with a hardware processor since stack management in accordance with the present invention can be adapted for use with any stack-based computing system. In one embodiment, stack management unit 150 includes a stack cache 155, a dribble manager unit 151, and a stack control unit 152. When hardware processor 100 is pushing data onto stack 400 (FIG. 4(a)) and stack cache 155 is almost full, dribble manager unit 151 transfers data from the bottom of stack cache 155 to stack 400 through data cache unit 160, so that the top portion of stack 400 remains in stack cache 155. When hardware processor 100 is popping data off of stack 400 and stack cache 155 is almost empty, dribble manager unit 151 transfers data from stack 400 to the bottom of stack cache 155 so that the top portion of stack 400 is maintained in stack cache 155.
  • Instruction decoder 135, as described herein, allows the folding away of JAVA virtual machine instructions pushing an operand onto the top of a stack merely as a precursor to a second JAVA virtual machine instruction which operates on the top of stack operand. Such an instruction decoder identifies foldable instruction sequences and supplies an execution unit with a single equivalent folded operation thereby reducing processing cycles otherwise required for execution of multiple operations corresponding to the multiple instructions of the folded instruction sequence. Instruction decoder embodiments described herein provide for folding of two, three, four, or more instruction folding. For example, in one instruction decoder embodiment described herein, two load instructions and a store instruction can be folded into execution of operation corresponding to an instruction appearing therebetween in the instruction sequence.
  • As used herein, a virtual machine is an abstract computing machine that, like a real computing machine, has an instruction set and uses various memory areas. A virtual machine specification defines a set of processor architecture independent virtual machine instructions that are executed by a virtual machine implementation, e.g., hardware processor 100. Each virtual machine instruction defines a specific operation that is to be performed. The virtual computing machine need not understand the computer language that is used to generate virtual machine instructions or the underlying implementation of the virtual machine. Only a particular file format for virtual machine instructions needs to be understood.
  • In an exemplary embodiment, the-virtual machine instructions are JAVA virtual machine instructions. Each JAVA virtual machine instruction includes one or more bytes that encode instruction identifying information, operands, and any other required information. Section I, which is incorporated herein by reference in its entirety, includes an illustrative set of the JAVA virtual machine instructions. The particular set of virtual machine instructions utilized is not an essential aspect of this invention. In view of the virtual machine instructions in Section I and this disclosure, those of skill in the art can modify the invention for a particular set of virtual machine instructions, or for changes to the JAVA virtual machine specification.
  • A JAVA compiler JAVAC, (FIG. 2) that is executing on a computer platform, converts an application 201 written in the JAVA computer language to an architecture neutral object file format encoding a compiled instruction sequence 203, according to the JAVA Virtual Machine Specification, that includes a compiled instruction set. However, for this invention, only a source of virtual machine instructions and related information is needed. The method or technique used to generate the source of virtual machine instructions and related information is not essential to this invention.
  • Compiled instruction sequence 203 is executable on hardware processor 100 as well as on any computer platform that implements the JAVA virtual machine using, for example, a software interpreter or just-in-time compiler. However, as described above, hardware processor 100 provides significant performance advantages over the software implementations.
  • In this embodiment, hardware processor 100 (FIG. 1) processes the JAVA virtual machine instructions, which include bytecodes. Hardware processor 100, as explained more completely below, executes directly most of the bytecodes. However, execution of some of the bytecodes is implemented via microcode.
  • One strategy for selecting virtual machine instructions that are executed directly by hardware processor 100 is described herein by way of an example. Thirty percent of the JAVA virtual machine instructions are pure hardware translations; instructions implemented in this manner include constant loading and simple stack operations. The next 50% of the virtual machine instructions are implemented mostly, but not entirely, in hardware and require some firmware assistance; these include stack based operations and array instructions. The next 10% of the JAVA virtual machine instructions are implemented in hardware, but require significant firmware support as well; these include function invocation and function return. The remaining 10% of the JAVA virtual machine instructions are not supported in hardware, but rather are supported by a firmware trap and/or microcode; these include functions such as exception handlers. Herein, firmware means microcode stored in ROM that when executed controls the operations of hardware processor 100.
  • In one embodiment, hardware-processor 100 includes an I/O bus and memory interface unit 110, an instruction cache unit 120 including instruction cache 125, an instruction decode unit 130, a unified execution unit 140, a stack management unit 150 including stack cache 155, a data cache unit 160 including a data cache 165, and program counter and trap control logic 170. Each of these units is described more completely below.
  • Also, as illustrated in FIG. 1, each unit includes several elements. For clarity and to avoid distracting from the invention, the interconnections between elements within a unit are not shown in FIG. 1, However, in view of the following description, those of skill in the art will understand the interconnections and cooperation between the elements in a unit and between the various units.
  • The pipeline stages implemented using the units illustrated in FIG. 1 include fetch, decode, execute, and write-back stages. If desired, extra stages for memory access or exception resolution are provided in hardware processor 100.
  • FIG. 3 is an illustration of a four stage pipeline for execution of instructions in the exemplary embodiment of processor 100. In fetch stage 301, a virtual machine instruction is fetched and placed in instruction buffer 124 (FIG. 1). The virtual machine instruction is fetched from one of (i) a fixed size cache line from instruction-cache 125 or (ii) microcode ROM 141 in execution unit 140.
  • With regard to fetching, aside from instructions tableswitch and lookupswitch, (see Section I.) each virtual machine instruction is between one and five bytes long. Thus, to keep things simple, at least forty bits are required to guarantee that all of a given instruction is contained in the fetch.
  • Another alternative is to always fetch a predetermined number of bytes, for example, four bytes, starting with the opcode. This is sufficient for 95% of JAVA virtual machine instructions (See Section I). For an instruction requiring more than three bytes of operands, another cycle in the front end must be tolerated if four bytes are fetched. In this case, the instruction execution can be started with the first operands fetched even if the full set of operands is not yet available.
  • In decode stage 302 (FIG. 3), the virtual machine instruction at the front of instruction buffer 124 (FIG. 1) is decoded and instruction folding is performed if possible. Stack cache 155 is accessed only if needed by the virtual machine instruction. Register OPTOP, that contains a pointer OPTOP to a top of a stack 400 (FIG. 4), is also updated in decode stage 302 (FIG. 3).
  • Herein, for convenience, the value in a register and the register are assigned the same reference numeral. Further, in the following discussion, use of a register to store a pointer is illustrative only of one embodiment. Depending on the specific implementation of the invention, the pointer may be implemented using hardware register, a hardware counter, a software counter, a software pointer, or other equivalent embodiments known to those of skill in the art. The particular implementation selected is not essential to the invention, and typically is made based on a price to performance trade-off.
  • In execute stage 303, the virtual machine instruction is executed for one or more cycles. Typically, in execute stage 303, an ALU in integer unit 142 (FIG. 1) is used either to do an arithmetic computation or to calculate the address of a load or store from data cache unit (DCU) 160. If necessary, traps are prioritized and taken at the end of execute stage 303 (FIG. 3). For control flow instructions, the branch address is calculated in execute stage 303, as well as the condition upon which the branch is dependent.
  • Cache stage 304 is a non-pipelined stage. Data cache 165 (FIG. 1) is accessed if needed during execution stage 303 (FIG. 3). The reason that stage 304 is non-pipelined is because hardware processor 100 is a stack-based machine. Thus, the instruction following a load is almost always dependent on the value returned by the load. Consequently, in this embodiment, the pipeline is held for one cycle for a data cache access. This reduces the pipeline stages, and the die area taken by the pipeline for the extra registers and bypasses.
  • Write-back stage 305 is the last stage in the pipeline. In stage 305, the calculated data is written back to stack cache 155.
  • Hardware processor 100, in this embodiment, directly implements a stack 400 (FIG. 4A) that supports the JAVA virtual machine stack-based architecture (See Section I). Sixty-four entries on stack 400 are contained on stack cache 155 in stack management unit 150. Some entries in stack 400 may be duplicated on stack cache 155. Operations on data are performed through stack cache 155.
  • Stack 400 of hardware processor 100 is primarily used as a repository of information for methods. At any point in time, hardware processor 100 is executing a single method. Each method has memory space, i.e., a method frame on stack 400, allocated for a set of local variables, an operand stack, and an execution environment structure.
  • A new method frame, e.g., method frame two 410, is allocated by hardware processor 100 upon a method invocation in execution stage 303 (FIG. 3) and becomes the current frame, i.e., the frame of the current method. Current frame 410 (FIG. 4A), as well as the other method frames, may contain a part of or all of the following six entities, depending-on various method invoking situations:
      • Object reference;
      • Incoming arguments;
      • Local variables;
      • Invoker's method context;
      • Operand stack; and
      • Return value from method.
  • In FIG. 4A, object reference, incoming arguments, and local variables are included in arguments and local variables area 421. The invoker's method context is included in execution environment 422, sometimes called frame state, that in turn includes: a return program counter value 431 that is the address of the virtual machine instruction, e.g., JAVA opcode, next to the method invoke instruction; a return frame 432 that is the location of the calling method's frame; a return constant pool pointer 433 that is a pointer to the calling method's constant pool table; a current method vector 434 that is the base address of the current method's vector table; and a current monitor address 435 that is the address of the current method's monitor.
  • The object reference is an indirect pointer to an object-storage representing the object being targeted for the method invocation. JAVA compiler JAVAC (See FIG. 2.) generates an instruction to push this pointer onto operand stack 423 prior to generating an invoke instruction. This object reference is accessible as local variable zero during the execution of the method. This indirect pointer is not available for a static method invocation as there is no target-object defined for a static method invocation.
  • The list of incoming arguments transfers information from the calling method to the invoked method. Like the object reference, the incoming arguments are pushed onto stack 400 by JAVA compiler generated instructions and may be accessed as local variables. JAVA compiler JAVAC (See FIG. 2.) statically generates a list of arguments for current method 410′ (FIG. 4A), and hardware processor 100 determines the number of arguments from the list. When the object reference is present in the frame for a non-static method invocation, the first argument is accessible as local variable one. For a static method invocation, the first argument becomes local variable zero.
  • For 64-bit arguments, as well as 64-bit entities in general, the upper 32-bits, i.e., the 32 most significant bits, of a 64-bit entity are placed on the upper location of stack 400, i.e., pushed on the stack last. For example, when a 64-bit entity is on the top of stack 400, the upper 32-bit portion of the 64-bit entity is on the top of the stack, and the lower 32-bit portion of the 64-bit entity is in the-storage location immediately adjacent to the top of stack 400.
  • The local variable area on stack 400 (FIG. 4A) for current method 410 represents temporary variable storage space which is allocated and remains effective during invocation of method 410. JAVA compiler JAVAC (FIG. 2) statically determines the required number of local variables and hardware processor 100 allocates temporary variable storage space accordingly.
  • When a method is executing on hardware processor 100, the local variables typically reside in stack cache 155 and are addressed as offsets from pointer VARS (FIGS. 1 and 4A), which points to the position of the local variable zero. Instructions are provided to load the values of local variables onto operand stack 423 and store values from operand stack into local variables area 421.
  • The information in execution environment 422 includes the invoker's method context. When a new frame is built for the current method, hardware processor 100 pushes the invoker's method context onto newly allocated frame 410, and later utilizes the information to restore the invoker's method context before returning. Pointer FRAME.(FIGS. 1 and 4A) is a pointer to the execution environment of the current method. In the exemplary embodiment, each register in register set 144 (FIG. 1) is 32-bits wide.
  • Operand stack 423 is allocated to support the execution of the virtual machine instructions within the current method. Program counter register PC (FIG. 1) contains the address of the next instruction, e.g., opcode, to be executed. Locations on operand stack 423 (FIG. 4A) are used to store the operands of virtual machine instructions, providing both source and target storage locations for instruction execution. The size of operand stack 423 is statically determined by JAVA compiler JAVAC (FIG. 2) and hardware processor 100 allocates space for operand stack 423 accordingly. Register OPTOP (FIGS. 1 and 4A) holds a pointer to a top of operand stack 423.
  • The invoked method may return its execution result onto the invoker's top of stack, so that the invoker can access the return value with operand stack references. The return value is placed on the area where an object reference or an argument is pushed before a method invocation.
  • Simulation results on the JAVA virtual machine indicate that method invocation consumes a significant portion of the execution time (20-40%). Given this attractive target for accelerating execution of virtual machine instructions, hardware support for method invocation is included in hardware processor 100, as described more completely below.
  • The beginning of the stack frame of a newly invoked method, i.e., the object reference and the arguments passed by the caller, are already stored on stack 400 since the object reference and the incoming arguments come from the top of the stack of the caller. As explained above, following these items on stack 400, the local variables are loaded and then the execution environment is loaded.
  • One way to speed up this process is for hardware processor 100 to load the execution environment in the background and indicate what has been loaded so far, e.g., simple one bit scoreboarding. Hardware processor 100 tries to execute the bytecodes of the called method as soon as possible, even though stack 400 is not completely loaded. If accesses are made to variables already loaded, overlapping of execution with loading of stack 400 is achieved, otherwise a hardware interlock occurs and hardware processor 100 just waits for the variable or variables in the execution to be loaded.
  • FIG. 4B illustrates another way to accelerate method invocation. Instead of storing the entire method frame in stack 400, the execution environment of each method frame is stored separately from the local variable area and the operand stack of the method frame. Thus, in this embodiment, stack 400B contains modified method frames, e.g. modified method frame 410B having only local variable area 421 and operand stack 423. Execution environment 422 of the method frame is stored in execution environment memory 440. Storing the execution environment in execution environment memory 440 reduces the amount of data in stack cache 155. Therefore, the size of stack cache 155 can be reduced. Furthermore, execution environment memory 440 and stack cache 155 can be accessed simultaneously. Thus, method invocation can be accelerated by loading or storing the execution environment in parallel with loading or storing data onto stack 400B.
  • In one embodiment of stack management unit 150, the memory architecture of execution environment memory 440 is also a stack. As modified method frames are pushed onto stack 400 b through stack cache 155, corresponding execution environments are pushed onto execution environment memory 440. For example, since modified method frames 0 to 2, as shown in FIG. 48, are in stack 400B, execution environments (EE) 0 to 2, respectively, are stored in execution environment memory circuit 440.
  • To further enhance method invocation, an execution environment cache can be added to improve the speed of saving and retrieving the execution environment during method invocation. The architecture described more completely below for stack cache 155, dribbler manager unit 151, and stack control unit 152 for caching stack 400, can also be applied to caching execution environment memory 440.
  • FIG. 4C illustrates an embodiment of stack management unit 150 modified to support both stack 400 b and execution environment memory 440. Specifically, the embodiment of stack management unit 150 in FIG. 4C adds an execution environment stack cache 450, an execution environment dribble manager unit 460, and an execution environment stack control unit 470. Typically, execution dribble manager unit 460 transfers an entire execution environment between execution environment cache 450 and execution environment memory 440 during a spill operation or a fill operation.
  • I/O Bus and Memory Interface Unit
  • I/O bus and memory ,interface unit 110 (FIG. 1), sometimes called interface unit 110, implements an interface between hardware processor 100 and a memory hierarchy which in an exemplary embodiment includes external memory and may optionally include memory storage and/or interfaces on the same die as hardware processor 100. In this embodiment, I/O controller 111 interfaces with external I/O devices and memory controller 112 interfaces with external memory. Herein, external memory means memory external to hardware processor 100. However, external memory either may be included on the same die as hardware processor 100, may be external to the die containing hardware processor 100, or may include both on- and off-die portions.
  • In another embodiment, requests to I/O devices go through memory controller 112 which maintains an address map of the entire system including hardware processor 100. On the memory bus of this embodiment, hardware processor 100 is the only master and does not have to arbitrate to use the memory bus.
  • Hence, alternatives for the input/output bus that interfaces with I/O bus and memory interface unit 110 include supporting memory-mapped schemes, providing direct support for PCI, PCMCIA, or other standard busses. Fast graphics (w/VIS or other technology) may optionally be included on the die with hardware processor 100.
  • I/O bus and memory interface unit 110 generates read and write requests to external memory. Specifically, interface unit 110 provides an interface for instruction cache and data cache controllers 121 and 161 to the external memory. Interface unit 110 includes arbitration logic for internal requests from instruction cache controller 121 and data cache controller 161 to access external memory and in response to a request initiates either a read or a write request on the memory bus to the external memory. A request from data cache controller 161 is always treated as higher priority relative to a request from instruction cache controller 121.
  • Interface unit 110 provides an acknowledgment signal to the requesting instruction cache controller 121, or data cache controller 161 on read cycles so that the requesting controller can latch the data. On write cycles, the acknowledgment signal from interface unit 110 is used for flow control so that the requesting instruction cache controller 121 or data cache controller 161 does not generate a new request when there is one pending. Interface unit 110 also handles errors generated on the memory bus to the external memory.
  • Instruction Cache Unit
  • Instruction cache unit (ICU) 120 (FIG. 1) fetches virtual machine instructions from instruction cache 125 and provides the instructions to instruction decode unit 130. In this embodiment, upon a instruction cache hit, instruction cache controller 121, in one cycle, transfers an instruction from instruction cache 125 to instruction buffer 124 where the instruction is held until integer execution unit IEU, that is described more completely below, is ready to process the instruction. This separates the rest of pipeline 300 (FIG. 3) in hardware processor 100 from fetch stage 301. If it is undesirable to incur the complexity of supporting an instruction-buffer type of arrangement, a temporary one instruction register is sufficient for most purposes. However, instruction fetching, caching, and buffering should provide sufficient instruction bandwidth to support instruction folding as described below.
  • The front end of hardware processor 100 is largely separate from the rest of hardware processor 100. Ideally, one instruction per cycle is delivered to the execution pipeline.
  • The instructions are aligned on an arbitrary eight-bit boundary by byte aligner circuit 122 in response to a signal from instruction decode unit 130. Thus, the front end of hardware processor 100 efficiently deals with fetching from any byte position. Also, hardware processor 100 deals with the problems of instructions that span multiple cache lines of cache 125. In this case, since the opcode is the first byte, the design is able to tolerate an extra cycle of fetch latency for the operands. Thus, a very simple de-coupling between the fetching and execution of the bytecodes is possible.
  • In case of an instruction cache miss, instruction cache controller 121 generates an external memory request for the missed instruction to I/O bus and memory interface unit 110. If instruction buffer 124 is empty, or nearly empty, when there is an instruction cache miss, instruction decode unit 130 is stalled, i.e., pipeline 300 is stalled. Specifically, instruction cache controller 121 generates a stall signal upon a cache miss which is used along with an instruction buffer empty signal to determine whether to stall pipeline 300. Instruction cache 125 can be invalidated to accommodate self-modifying code, e.g., instruction cache controller 121 can invalidate a particular line in instruction cache 125.
  • Thus, instruction cache controller 121 determines the next instruction to be fetched, i.e., which instruction in instruction cache 125 needs to accessed, and generates address, data and control signals for data and tag RAMs in instruction cache 125. On a cache hit, four bytes of data are fetched from instruction cache 125 in a single cycle, and a maximum of four bytes can be written into instruction buffer 124.
  • Byte aligner circuit 122 aligns the data out of the instruction cache RAM and feeds the aligned data to instruction buffer 124. As explained more completely below, the first two bytes in instruction buffer 124 are decoded to determine the length of the virtual machine instruction. Instruction buffer 124 tracks the valid instructions in the queue and updates the entries, as explained more completely below.
  • Instruction cache controller 121 also provides the data path and control for handling instruction cache misses. On an instruction cache miss, instruction cache controller 121 generates a cache fill request to I/O bus and memory interface unit 110.
  • On receiving data from external memory, instruction cache controller 121 writes the data into instruction cache 125 and the data are also bypassed into instruction buffer 124. Data are bypassed to instruction buffer 124 as soon as the data are available from external memory, and before the completion of the cache fill.
  • Instruction cache controller 121 continues fetching sequential data until instruction buffer 124 is full or a branch or trap has taken place. In one embodiment, instruction buffer 124 is considered full if there are more than eight bytes of valid entries in buffer 124. Thus, typically, eight bytes of data are written into instruction cache 125 from external memory in response to the cache fill request sent to interface unit 110 by instruction cache unit 120. If there is a branch or trap taken while processing an instruction cache miss, only after the completion of the miss processing is the trap or branch executed.
  • When an error is generated during an instruction cache fill transaction, a fault indication is generated and stored into instruction buffer 124 along with the virtual machine instruction, i.e., a fault bit is set. The line is not written into instruction cache 125. Thus, the erroneous cache fill transaction acts like a non-cacheable transaction except that a fault bit is set. When the instruction is decoded, a trap is taken.
  • Instruction cache controller 121 also services non-cacheable instruction reads. An instruction cache enable (ICE) bit, in a processor status register in register set 144, is used to define whether a load can be cached. If the instruction cache enable bit is cleared, instruction cache unit 120 treats all loads as non-cacheable loads. Instruction cache controller 121 issues a non-cacheable request to interface unit 110 for non-cacheable instructions. When the data are available on a cache fill bus for the non-cacheable instruction, the data are bypassed into instruction buffer 124 and are not written into instruction cache 125.
  • In this embodiment, instruction cache 125 is a direct-mapped, eight-byte line size cache. Instruction cache 125 has a single cycle latency. The cache size is configurable to 0K, 1K, 2K, 4K, 8K and 16K byte sizes where K means kilo. The default size is 4K bytes. Each line has a cache tag entry associated with the line. Each cache tag contains a twenty bit address tag field and one valid bit for the-default 4K byte size.
  • Instruction buffer 124, which, in an exemplary embodiment, is a twelve-byte deep first-in, first-out (FIFO) buffer, de-links fetch stage 301 (FIG. 3) from the rest of pipeline 300 for performance reasons. Each instruction in buffer 124 (FIG. 1) has an associated valid bit and an error bit. When the valid bit is set, the instruction associated with that valid bit is a valid instruction. When the error bit is set, the fetch of the instruction associated with that error bit was an erroneous transaction. Instruction buffer 124 includes an instruction buffer control circuit (not shown) that generates signals to pass data to and from instruction buffer 124 and that keeps track of the valid entries in instruction buffer 124, i.e., those with valid bits set.
  • In an exemplary embodiment, four bytes can be received into instruction buffer 124 in a given cycle. Up to five bytes, representing up to two virtual machine instructions, can be read out of instruction buffer 124 in a given cycle. Alternative embodiments, particularly those providing folding of multi-byte virtual machine instructions and/or those providing folding of more than two virtual machine instructions, provide higher input and output bandwidth. Persons of ordinary skill in the art will recognize a variety of suitable instruction buffer designs including, for example, alignment logic, circular buffer design, etc. When a branch or trap is taken, all the entries in instruction buffer 124 are nullified and the branch/trap data moves to the top of instruction buffer 124.
  • In the embodiment of FIG. 1, a unified execution unit 140 is shown. However, in another embodiment, instruction decode unit 130, integer unit 142, and stack management unit 150 are considered a single integer execution unit, and floating point execution unit 143 is a separate optional unit. In still other embodiments, the various elements in the execution unit may be implemented using the execution unit of another processor. In general, the various elements included in the various units of FIG. 1 are exemplary only of one embodiment. Each unit could be implemented with all or some of the elements shown. Again, the decision is largely dependent upon a price vs. performance trade-off.
  • Instruction Decode Unit
  • As explained above, virtual machine instructions are decoded in decode stage 302 (FIG. 3) of 35 pipeline 300. In an exemplary embodiment, two bytes, that can correspond to two virtual machine instructions, are fetched from instruction buffer 124 (FIG. 1). The two bytes are decoded in parallel to determine if the two bytes correspond to two virtual machine instructions, e.g., a first load top of stack, instruction and a second add top two stack entries instruction, that can be folded into a single equivalent operation. Folding refers to supplying a single equivalent operation corresponding to two or more virtual machine instructions.
  • In an exemplary hardware processor 100 embodiment, a single-byte first instruction can be folded with a second instruction. However, alternative embodiments provide folding of more than two virtual machine instructions, e.g., two to four virtual machine instructions, and of multi-byte virtual machine instructions, though at the cost of instruction decoder complexity and increased instruction bandwidth. See U.S. patent application Ser. No. 08/786,351, entitled “INSTRUCTION FOLDING FOR A STACK-BASED MACHINE” naming Marc Tremblay and James Michael O'Connor as inventors, assigned to the assignee of this application, and filed on Jan. 23, 1997 with Attorney Docket No. SP2036, now U.S. Pat. No. 6,026,485, which is incorporated herein by reference in its entirety. In the exemplary processor 100 embodiment, if the first byte, which corresponds to the first virtual machine instruction, is a multi-byte instruction, the first and second instructions are not folded.
  • An optional current object loader folder 132 exploits instruction folding, such as that described above, and in greater detail in U.S. patent application Ser. No. 08/786,351, entitled “INSTRUCTION FOLDING FOR A STACK-BASED MACHINE” naming Marc Tremblay and James Michael O'Connor as inventors, assigned to the assignee, of this application, and filed on Jan. 23, 1997, now U.S. Pat. No. 6,026,485, which is incorporated herein by reference in its entirety, in virtual machine instruction sequences which simulation results have shown to be particularly frequent and therefore a desirable target for optimization. In particular, a method invocation typically loads an object reference for the corresponding object onto the operand stack and fetches a field from the object. Instruction folding allows this extremely common virtual machine instruction sequence to be executed using an equivalent folded operation.
  • Quick variants are not part of the virtual machine instruction set (See Chapter 3 of Section I), and are invisible outside of a JAVA virtual machine implementation. However, inside a virtual machine implementation, quick variants have proven to be an effective optimization. (See Section A in Section I; which is an integral part of this specification.) Supporting writes for updates of various instructions to quick variants in a non-quick to quick translator cache 131 changes the normal virtual machine instruction to a quick virtual machine instruction to take advantage of the large benefits bought from the quick variants. In particular, as described in more detail in U.S. patent application Ser. No. 08/788,805, entitled “NON-QUICK INSTRUCTION ACCELERATOR INCLUDING INSTRUCTION IDENTIFIER AND DATA SET STORAGE AND METHOD OF IMPLEMENTING SAME” naming Marc Tremblay and James Michael O'Connor as inventors, assigned to the assignee of this application, and filed on Jan. 23, 1997 with Attorney Docket No. SP2039, now U.S. Pat. No. 6,065,108, which is incorporated herein by reference in its entirety, when the information required to initiate execution of an instruction has been assembled for the first time, the information is stored in a cache along with the value of program counter PC as a tag in non-quick to quick translator cache 131 and the instruction is identified as a quick-variant. In one embodiment, this is done with self-modifying-code.
  • Upon a subsequent call of that instruction, instruction decode unit 130 detects that the instruction is identified as a quick-variant and simply retrieves the information needed to initiate execution of the instruction from non-quick to quick translator cache 131. Non-quick to quick translator cache is an optional feature of hardware processor 100.
  • With regard to branching, a very short pipe with quick branch resolution is sufficient for most implementations. However, an appropriate simple branch prediction mechanism can alternatively be introduced, e.g., branch predictor circuit 133. Implementations for branch predictor circuit 133 include branching based on opcode, branching based on offset, or branching based on a two-bit counter mechanism.
  • The JAVA virtual machine specification defines an instruction invokenonvirtual, opcode 183, which, upon execution, invokes methods. The opcode is followed by an index byte one and an index byte two. (See Section I.) Operand stack 423 contains a reference to an object and some number of arguments when this instruction is executed.
  • Index bytes one and two are used to generate an index into the constant pool of the current class. The item in the constant pool at that index points to a complete method signature and class. Signatures are defined in Section I and that description is incorporated herein by reference.
  • The method signature, a short, unique identifier for each method, is looked up in a method table of the class indicated. The result of the lookup is a method block that indicates the type of method and the number of arguments for the method. The object reference and arguments are popped off this method's stack and become initial values of the local variables of the new method. The execution then resumes with the first instruction of the new method. Upon execution, instructions invokevirtual, opcode 182, and invokestatic, opcode 184, invoke processes similar to that just described. In each case, a pointer is used to lookup a method block.
  • A method argument cache 134, that also is an optional feature of hardware processor 100, is used, in a first embodiment, to store the method block of a method for use after the first call to the method, along with the pointer to the method block as a tag. Instruction decode unit 130 uses index bytes one and two to generate the pointer and then uses the pointer to retrieve the method block for that pointer in cache 134. This permits building the stack frame for the newly invoked method more rapidly in the background in subsequent invocations of the method. Alternative embodiments may use a program counter or method identifier as a reference into cache 134. If there is a cache miss, the instruction is executed in the normal fashion and cache 134 is updated accordingly. The particular process used to determine which cache entry is overwritten is not an essential aspect of this invention. A least-recently used criterion could be implemented, for example.
  • In an alternative embodiment, method argument cache 134 is used to store the pointer to the method block, for use after the first call to the method, along with the value of program counter PC of the method as a tag. Instruction decode unit 130 uses the value of program counter PC to access cache 134. If the value of program counter PC is equal to one of the tags in cache 134, cache 134 supplies the pointer stored with that tag to instruction decode unit 130. Instruction decode unit 130 uses the supplied painter to retrieve the method block for the method. In view of these two embodiments, other alternative embodiments will be apparent to those of skill in the art.
  • Wide index forwarder 136, which is an optional element of hardware processor 100, is a specific embodiment of instruction folding for instruction wide. Wide index forwarder 136 handles an opcode encoding an extension of an index operand for an immediately subsequent virtual machine instruction. In this way, wide index forwarder 136 allows instruction decode unit 130 to provide indices into local variable storage 421 when the number of local variables exceeds that addressable with a single byte index without incurring a separate execution cycle for instruction wide.
  • Aspects of instruction decoder 135, particularly instruction folding, non-quick to quick translator cache 131, current object loader folder 132, branch predictor 133, method argument cache 134, and wide index forwarder 136 are also useful in implementations that utilize a software interpreter or just-in-time compiler, since these elements can be used to accelerate the operation of the software interpreter or just-in-time compiler. In such an implementation, typically, the virtual machine instructions are translated to an instruction for the processor executing the interpreter or compiler, e.g., any one of a Sun processor, a DEC processor, an Intel processor, or a Motorola processor, for example, and the operation of the elements is modified to support execution on that processor. The translation from the virtual machine instruction to the other processor instruction can be done either with a translator in a ROM or a simple software translator. For additional examples of dual instruction set processors, see U.S. patent application Ser. No. 08/787,618, entitled “A PROCESSOR FOR EXECUTING INSTRUCTION SETS RECEIVED FROM A NETWORK OR FROM A LOCAL MEMORY” naming Marc Tremblay and James Michael O'Connor as inventors, now U.S. Pat. No. 5,925,123, assigned to the assignee of this application, and filed on Jan. 23, 1997 with Attorney Docket No. SP2042, which is incorporated herein by reference in its entirety.
  • As explained above, one embodiment of processor 100 implements instruction folding to enhance the performance of processor 100. In general, instruction folding in accordance with the present invention can be used in any of a stack-based virtual machine implementation, including, e.g., in a hardware processor implementation, in a software interpreter implementation, in a just-in-time compiler implementation, etc. Thus, while various embodiments of instruction folding are described in the following more detailed description in terms of a hardware processor, those of skill in the art will appreciate, in view of this description, suitable extensions of instruction folding to other stack-based virtual machine implementations.
  • FIG. 14 illustrates folded execution of first and second stack instructions, according to the principles of this invention. In this embodiment, a first operand for an addition instruction resides in top-of-stack (TOS) entry 141 la of stack 1410. A second operand resides in entry 1412 of stack 1410. Notice that entry 1412 is not physically adjacent to top-of stack entry 141 la and in fact, is in the interior of stack 1410. An instruction stream includes a load top-of-stack instruction for pushing the second operand onto the top of stack (see description of instruction iload in Section I) and an addition instruction for operating on the first and second operands residing in the top two entries of stack 1410 (see description of instruction iadd in Section I). However, to speed execution of the instruction stream, the load top-of-stack and addition instructions are folded into a single operation whereby the explicit sequential execution of the load top-of-stack instruction and the associated execution cycle are eliminated. Instead, a folded operation corresponding to the addition instruction operates on the first and second operands, which reside in TOS entry 141 la and entry 1412 of stack 1410. The result of the folded operation is pushed onto stack 1410 at TOS entry 141 lb. Thus, folding according to the principles of this invention enhances performance compared to an unfolded method for executing the same sequence of instructions.
  • Without instruction folding, a first operand for an addition instruction resides in top-of-stack (TOS) entry 1311 a of stack 1310 (see FIG. 13). A second operand resides in entry 1312 of stack 1310. A load to top-of-stack instruction pushes the second operand onto the top of stack 1310 and typically requires an execution cycle. The push results in the second and first operands residing in TOS entry 1311 b and (TOS-1) entry 1313, respectively. Thereafter, the addition instruction operates, in another execution cycle, on the first and second operands which properly reside in the top two entries, i.e., TOS entry 1311 b and (TOS-1) entry 1313, of stack 1310 in accordance with the semantics of a stack architecture. The result of the addition instruction is pushed onto stack 1310 at TOS entry 1311 c and after the addition instruction is completed, it is as if the first and second operand data were never pushed onto stack 1310. As described above, folding reduces the execution cycles required to complete the addition and so enhances the speed of execution of the instruction stream. More complex folding, e.g., folding including store instructions and folding including larger numbers of instructions, is described in greater detail below.
  • In general, instruction decoder unit 130 (FIG. 1) examines instructions in a stream of instructions. Instruction decoder unit 130 folds first and second adjacent instructions together and provides a single equivalent operation for execution by execution unit 140 when instruction decoder unit 130 detects that the first and second instructions have neither structural nor resource dependencies and the second instruction operates on data provided by the first instruction. Execution of the single operations obtains the same result as execution of an operation corresponding to the first instruction followed by execution an operation corresponding to the second instruction, except that an execution cycle has been eliminated.
  • As described above, the JAVA virtual machine is stack-oriented and specifies an instruction set, a register set, an operand stack, and an execution environment. Although, the present invention is described in relation to the JAVA Virtual Machine, those of skill in the art will appreciate that the invention is not limited to embodiments implementing or related to the JAVA virtual machine and, instead, encompasses systems, articles, methods, and apparati for a wide variety of stack machine environments, both virtual and physical.
  • As illustrated in FIG. 4A, according to the JAVA Virtual Machine Specification, each method has storage allocated for an operand stack and a set of local variables. Similarly, in the embodiment of FIG. 15 (see also FIG. 4A), a series of method frames e.g., method frame 1501 and method frame 1502 on stack 1503, each include an operand stack instance, local variable storage instance, and frame state information instance for respective methods invoked along the execution path of a JAVA program. A new frame is created and becomes current each time a method is invoked and is destroyed after the method completes execution. A frame ceases to be current if its method invokes another method. On method return, the current frame passes back the result of its method invocation, if any, to the previous frame via stack 1503. The current frame is then discarded and the previous frame becomes current. Folding in accordance with the present invention, as described more completely below, is not dependent upon a particular process used to allocate or define memory space for a method, such as a frame, and can, in general, be used in any stack based architecture.
  • This series of method frames may be implemented in any of a variety of suitable memory hierarchies, including for example register/cache/memory hierarchies. However, irrespective of the memory hierarchy chosen, an operand stack instance 1512 (FIG. 15) is implemented in randomly-accessible storage 1510, i.e., at least some of the entries in operand stack instance 1512 can be accessed from locations other than the top most locations of operand stack instance 1512 in contrast with a conventional stack implementation in which only the top entry or topmost entries of the stack can be accessed. As described above, register OPTOP stores a pointer that identifies the top of operand stack instance 1512 associated with the current method. The value stored in register OPTOP is maintained to identify the top entry of an operand stack instance corresponding to the current method.
  • In addition, local variables for the current method are represented in randomly-accessible storage 1510. A pointer stored in register VARS identifies the starting address of local variable storage instance 1513 associated with the current method. The value in register VARS is maintained to identify a base address of the local variable storage instance corresponding to the current method.
  • Entries in operand stack instance 1512 and local variable storage instance 1513 are referenced by indexing off of values represented in registers OPTOP and VARS, respectively, that in the embodiment of FIG. 1 are included in register set 144, and in the embodiment of FIG. 15 are included in pointer registers 1522. Pointer registers 1522 may be represented in physical registers of a processor implementing the JAVA Virtual Machine, or optionally, in randomly-accessible storage 1510. In an exemplary embodiment, commonly used offsets OPTOP-1, OPTOP-2, VARS+1, VARS+2, and VARS+3 are derived from the values in registers OPTOP and VARS, respectively. Alternatively, the additional offsets could be stored in registers of pointer registers 1522.
  • Operand stack instance 1512 and local variable storage instance 1513 associated with the current method are preferably represented in a flat 64-entry cache, e.g., stack cache 155 (see FIG. 1) whose contents are kept updated so that a working set of operand stack and local variable storage entries are cached. However, depending on the size of the current frame, the current frame including operand stack instance 1512 and local variable storage instance 1513 may be fully or partially represented in the cache. Operand stack and local variable storage entries for frames other than the current frame may also be represented in the cache if space allows. A suitable representation of a cache suitable for use with the folding of this invention is described in greater detail in U.S. patent application Ser. No. 08/787,736, entitled “METHODS AND APPARATI FOR STACK CACHING” naming Marc Tremblay and James Michael O'Connor as inventors, assigned to the assignee of this application, and filed on Jan. 23, 1997, the detailed description of which is incorporated herein by reference, and in U.S. patent application Ser. No. 08/787,617, entitled “METHOD FRAME STORAGE USING MULTIPLE MEMORY CIRCUITS” naming Marc Tremblay and James Michael O'Connor as inventors, assigned to the assignee of this application, and filed on Jan. 23, 1997, the detailed description of which also is incorporated herein by reference. However, other representations, including separate and/or uncached operand stack and local variable storage areas, are also suitable.
  • In addition to method frames and their associated operand stack and local variable storage instances, a constant area 1514 is provided in the address space of a processor implementing the JAVA virtual machine for commonly-used constants, e.g., constants specified by JAVA virtual machine instructions such as instruction iconst. In some cases, an operand source is represented as an index into constant area 1514. In the embodiment of FIG. 15, constant area 1514 is represented in randomly-accessible storage 1510. Optionally, entries of constant area 1514 could also be cached, e.g., in stack cache 155.
  • Although those of skill in the art will recognize the advantages of maintaining an operand stack and local variable storage instance for each method, as well as the opportunities for passing parameters and results created by maintaining the various instances of operand stack and local variable storage in a stack-oriented structure, in the interest of clarity, the description which follows focuses on the particular instances (operand stack instance 1512 and local variable storage instance 1513) of each associated the current method. Hereafter, these particular instances of an operand stack and local variable storage are referred to simply as operand stack 1512 and local variable storage 1513. Despite this simplification for purposes of illustration, those of skill in the art will appreciate that operand stack 1512 and local variable storage 1513 refer to any instances of an operand stack and variable storage associated with the current method, including representations which maintain separate instances for each method and representations which combine instances into a composite representation.
  • Operand sources and result targets for JAVA Virtual Machine instructions typically identify entries of operand stack instance 1512 or local variable storage instance 1513, i.e., identify entries of the operand stack and local variable storage for the current method. By way of example, and not limitation, representative JAVA virtual machine instructions are described in Chapter 3 of The JAVA Virtual Machine Specification which is included at Section I.
  • JAVA virtual machine instructions rarely explicitly designate both the source of the operand, or operands, and the result destination. Instead, either the source or the destination is implicitly the top of operand stack 1512. Some JAVA bytecodes explicitly designate neither a source nor a destination. For example, instruction iconst 0 pushes a constant integer zero onto operand stack 1512. The constant zero is implicit in the instruction, although the instruction may actually be implemented by a particular JAVA virtual machine implementation using a representation of the value zero from a pool of constants, such as constant area 1514, as the source for the zero operand. An instruction decoder for a JAVA virtual machine implementation that implements instruction iconst 0 in this way could generate, as the source address, the index of the entry in constant area 1514 where the constant zero is represented.
  • Prior to considering the various embodiments of folding in accordance with the present invention, it is informative to consider execution of JAVA virtual machine instructions, such as the iadd instruction and the arraylength instruction, without the folding process. After the operations associated with typical execution of JAVA virtual machine instructions are understood, the advantages of this invention will be more apparent. Further, this understanding will assist those of skill in the art in extending the invention to other stack-based architectures that do not rely upon the JAVA virtual machine instructions.
  • Focusing illustratively on operand stack and local variable storage structures associated with the current method and referring now to FIGS. 16A-D, the JAVA virtual machine integer add instruction, iadd, generates the sum of first and second integer operands, referred to as operand1 and operand2, respectively, that are at the top two locations of operand stack 1512. The top two locations are identified, at the time of instruction iadd execution, by pointer OPTOP in register OPTOP and by pointer OPTOP-1. The result of the execution of instruction iadd, i.e., the sum of first and second integer operands, is pushed onto operand stack 1512.
  • FIG. 16A shows the state of operand stack 1512 and local variable storage 1513 that includes first and second values, referred to as value1 and value2, before execution of a pair of JAVA virtual machine integer load instructions iload. In FIG. 16A, pointer OPTOP has the value AAC0h.
  • FIG. 16B shows operand stack 1512 after execution of the pair of instructions iload that load integer values from local variable storage 1513 onto operand stack 1512, pushing (i.e., copying) values value1 and value2 from locations identified by pointer VARS in register VARS and by pointer VARS+2 onto operand stack 1512 as operand1 at location AAC4h and operand2 at location AAC8h, and updating pointer OPTOP in the process to value AAC8h. FIG. 16C shows operand stack 1512 after instruction iadd has been executed. Execution of instruction iadd pops operands operand1 and operand2 off operand stack 1512, calculates the sum of operands operand1 and operand2, and pushes that sum onto operand stack 1512 at location AAC4h. After execution of instruction iadd, pointer OPTOP has the value AAC0h and points to the operand stack 1512 entry storing the sum.
  • FIG. 16D shows operand stack 1512 after an instruction istore has been executed. Execution of instruction istore pops the sum off operand stack 1512 and stores the sum in the local variable storage 1513 entry at the location identified by pointer VARS+2.
  • Variations for other instructions which push operands onto operand stack 1512 and which operate on values residing at the top of operand stack 1512 will be apparent to those of skill in the art. For example, variations for alternate operations and for data types requiring multiple operand stack 1512 entries, e.g., long integer values, double-precision floating point values, etc., will be apparent to those of skill in the art in view of this disclosure.
  • The folding example of FIGS. 17A-C is analogous to that illustrated with reference to FIGS. 16A-D, though with only load folding illustrated. Execution of JAVA virtual machine length of array instruction arraylength determines the length of an array whose object reference pointer objectref is at the top of operand stack 1512, and pushes the length onto operand stack 1512. FIG. 17A shows the state of operand stack 1512 and local variable storage 1513 before execution of JAVA virtual machine reference load instruction aload that is used to load an object reference from local variable storage 1513 onto the top of operand stack 1512. In FIG. 17A, pointer OPTOP has the value AAC0h.
  • FIG. 17B shows operand stack 1512 after execution of instruction aload pushes, i.e., copies, object reference pointer objectref onto the top of operand stack 1512 and updates pointer OPTOP to AAC4h in the process.
  • FIG. 17C shows operand stack 1512 after instruction arraylength has been executed. Execution of instruction arraylength pops object reference pointer objectref off operand stack 1512, calculates the length of the array referenced thereby, and pushes that length onto operand stack 1512. Suitable implementations of the instruction arraylength may supply object reference pointer objectref to an execution unit, e.g., execution unit 140, which subsequently overwrites the object reference pointer objectref with the value length. Whether the object reference pointer objectref is popped from operand stack 1512 or simply overwritten, after execution of instruction arraylength, pointer OPTOP has the value AAC4h and points to the operand stack 1512 entry storing the value length.
  • FIG. 18 illustrates a processor 1800 wherein loads, such as those illustrated in FIGS. 16A and 16B and in FIGS. 17A and 17B, are folded into execution of subsequent instructions, e.g., into execution of subsequent instruction iadd, or instruction arraylength. In this way, intermediate execution cycles associated with loading operands operand1 and operand2for instruction iadd, or with loading pointer objectref for instruction arraylength onto the top of operand stack 1512 can be eliminated. As a result, single cycle execution of groups of JAVA virtual machine instructions e.g., the group of instructions iload, iload, iadd, and istore, or the group of instructions aload and arraylength, is provided by processor 1800. One embodiment of processor 1800 is presented in FIG. 1 as hardware processor 100. However, hardware processor 1800 includes other embodiments that do not include the various optimizations of hardware processor 100. Further, the folding processes described below could be implemented in a software interpreter or a included within a just-in-time compiler. In the processor 1800 embodiment of FIG. 18, stores such as that illustrated in FIG. 16D, are folded into execution of prior instructions, e.g., in FIG. 16D, into execution of the immediately prior instruction iadd.
  • The instruction folding is provided primarily by instruction decoder 1818. Instruction decoder 1818 retrieves fetched instructions from instruction buffer 1816 and depending upon the nature of instructions in the fetched instruction sequence, supplies execution unit 1820 with decoded operation and operand addressing information implementing the instruction sequence as a single folded operation. Unlike instructions of the JAVA virtual machine instruction set to which the instruction sequence from instruction buffer 1816 conforms, decoded operations supplied to execution unit 1820 by instruction decoder 1818 operate on operand values represented in entries of local variable storage 1513, operand stack 1512, and constant area 1514.
  • In the exemplary embodiment of FIG. 18, valid operand sources include local variable storage 1513 entries identified by pointers VARS, VARS+1, VARS+2, and VARS+3, as well as operand stack 1512 entries identified by pointers OPTOP, OPTOP-1 and OPTOP-2. Similarly, valid result targets include local variable storage 1513 entries identified by operands VARS, VARS+1, VARS+2, and VARS+3. Embodiments in accordance with FIG. 18 may also provide for constant area 1514 entries as valid operand sources as well as other locations in operand stack 1512 and local variable storage 1513.
  • Referring now to FIGS. 18 and 19, a sequence of JAVA virtual machine instructions is fetched from memory and loaded into instruction buffer 1816. Conceptually, instruction buffer 1816 is organized as a shift register for JAVA bytecodes. One or more bytecodes are decoded by instruction decoder 1818 during each cycle and operations are supplied to execution unit 1820 in the form of a decoded operation on instruction decode bus instr_dec and associated operand source and result destination addressing information on instruction address bus instr_addr. Instruction decoder 1818 also provides an instruction valid signal instr_valid to execution unit 1820. When asserted, signal instr_valid indicates that the information on instruction decode bus instr_dec specifies a valid operation.
  • One or more bytecodes are shifted out of instruction buffer 1816 to instruction decode unit 1818 each cycle in correspondence with the supply of decoded operations and operand addressing information to execution unit 1820, and subsequent undecoded bytecodes are shifted into instruction buffer 1816. For normal decode operations, a single instruction is shifted out of instruction buffer 1816 and decoded by instruction decode unit 1818, and a single corresponding operation is executed by execution unit 1820 during each instruction cycle.
  • In contrast, for folded decode operations, multiple instructions, e.g., a group of instructions, are shifted out of instruction buffer 1816 to instruction decode unit 1818. In response to the multiple instructions, instruction decode unit 1818 generates a single equivalent folded operation that in turn is executed by execution unit 1820 during each instruction cycle.
  • Referring illustratively to the instruction sequence described above with reference to FIGS. 16A-16D, instruction decoder 1818 selectively decodes bytecodes associated with four JAVA virtual machine instructions:
  • 1. iload value1;
  • 2. iload value2;
  • 3. iadd; and
  • 4. istore,
  • that were described in the above description of FIGS. 16A-D. As now described, both instructions iload and the instruction istore are folded by instruction decoder 1818 into an add operation corresponding to instruction iadd. Although operation of instruction decoder 1818 is illustrated using a foldable sequence of four instructions, those of skill in the art will appreciate that the invention is not limited to four instructions. Foldable sequences of two, three, four, five, or more instructions are envisioned. For example, more than one instruction analogous to the instruction istore and more than two instructions analogous to the instructions iload may be included in foldable sequences.
  • Instruction decoder 1818 supplies decoded operation information over bus instr_dec and associated operand source and result destination addressing information over bus instr_addr specifying that execution unit 1820 is to add the contents of local variable storage 1513 location 0, this is identified by pointer VARS, and local variable storage 1513 location 2, that is identified by pointer VARS+2, and store the result in local variable storage 1513 location 2, that is identified by pointer VARS+2. In this way, the two load instructions are folded into execution of an operation corresponding to instruction iadd. Two instruction cycles and the intermediate data state illustrated in FIG. 16B are eliminated. In addition, instruction istore is also folded into execution of the operation corresponding to instruction iadd, eliminating another instruction cycle, for a total of three, and the intermediate data state illustrated in FIG. 16C. In various alternative embodiments, instruction folding in accordance with the present invention may eliminate loads, stores, or both loads and stores.
  • FIG. 20 depicts an exemplary embodiment of an instruction decoder 1818 providing both folded and unfolded decoding of bytecodes. Selection of a folded or unfolded operating mode for instruction decoder 1818 is based on the particular sequence of bytecodes fetched into instruction buffer 1816 and subsequently accessed by instruction decoder 1818. A normal decode portion 2002 and a fold decode portion 2004 of instruction decoder 1818 are configured in parallel to provide support for unfolded and folded execution, respectively.
  • In the embodiment of FIG. 20, fold decode portion 2004 detects opportunities for folding execution of bytecodes in the bytecode sequence fetched into instruction buffer 1816. A detection of such a foldable sequence triggers selection of the output of fold decode portion 2004, rather than normal decode portion 2002, for provision to execution unit 1820. Advantageously, selection of folded or unfolded decoding is transparent to execution unit 1820, which simply receives operation information over bus instr_dec and associated operand source and result destination addressing information over bus instr_addr, and which need not know whether the information corresponds to a single instruction or a folded instruction sequence.
  • Normal decode portion 2002 functions to inspect a single bytecode from instruction buffer 1816 during each instruction cycle, and generates the following indications in response thereto:
  • 1. a normal instruction decode signal n_instr_dec, which specifies an operation, e.g., integer addition, corresponding to the decoded instruction, is provided to a first set of input terminals of switch 2006;
  • 2. a normal address signal n_adr, which makes explicit the source and destination addresses, e.g., first operand address=OPTOP, second operand address=OPTOP-1, and destination address=OPTOP-1 for an instruction iadd, for the decoded instruction, is provided to a first bus input of switch 2010;
  • 3. a net change in pointer OPTOP signal n_delta_optop, e.g., for the instruction iadd, net change=−1, which in the embodiment of FIG. 20 is encoded as a component of normal address signal n_adr; and
  • 4. an instruction valid signal instr_valid, which indicates whether normal instruction decode signal n_instr_dec specifies a valid operation, is provided to a first input terminal of switch 2008.
  • In contrast with normal decode portion 2002, and as discussed in greater detail below, fold decode portion 2004 of instruction decoder 1818 inspects sequences of bytecodes from the instruction buffer 1816 and determines whether operations corresponding to these sequences (e.g., the sequence iload value1 from local variable 0, iload value2 from local variable 2, iadd, and istore sum to local variable 2) can be folded together to eliminate unnecessary temporary storage of instruction operands and/or results on the operand stack. When fold decode portion 2004 determines that a sequence of bytecodes in instruction buffer 1816 can be folded together, fold decode portion 2004 generates the following indications:
  • 1. a folded instruction decode signal f_instr_dec, which specifies an equivalent operation, e.g., integer addition corresponding to the folded instruction sequence, is provided to a second set of input terminals of switch 2006;
  • 2. a folded address signal f_adr, which specifies source and destination addresses for the equivalent operation, e.g., first operand address=VARS, second operand address=VARS+2, and destination address=VARS+2, is provided to a second bus input of switch 2010;
  • 3. a net change in pointer OPTOP signal f_delta_optop, e.g., for the above sequence net change=0, which in the embodiment of FIG. 20 is encoded as a component of normal address signal n_adr; and
  • 4. a folded instruction valid signal f_Valid, which indicates whether folded instruction decode signal f_instr dec specifies a valid operation, is provided to a second input terminal of switch 2008.
  • Fold decode portion 2004 also generates a signal on fold line f/nf which indicates whether a sequence of bytecodes in instruction buffer 1816 can be folded together. The signal on fold line f/nf is provided to control inputs of switches 2006, 2010 and 2008. If a sequence of bytecodes in instruction buffer 1816 can be folded together, the signal on fold line f/nf causes switches 2006, 2010 and 2008 to select respective second inputs for provision to execution unit 1820, i.e., to source folded instruction decode signal f_instr_dec, folded address signal f_adr, and folded instruction valid signal f_valid from fold decode portion 2004. If a sequence of bytecodes in instruction buffer 1816 cannot be folded together, the signal on fold line f/nf causes switches 2006, 2010 and 2008 to select respective first inputs for provision to execution unit 1820, i.e., to source normal instruction decode signal n_instr_dec, normal address signal n_adr, and normal instruction valid signal n valid from fold decode portion 2004.
  • In some embodiments in accordance with the present invention, the operation of fold decode portion 2004 is suppressed in response to an active suppress folding signal suppress_fold supplied from outside instruction decoder 1818. In response to an asserted suppress folding signal suppress_fold (see FIG. 21), the signal on fold line f/nf remains in a state selective for respective first inputs of switches 2006, 2010 and 2008 even if the particular bytecode sequence presented by instruction buffer 1816 would otherwise trigger folding. For example, in one such embodiment, suppress folding signal suppress_fold is asserted when the local variable storage 1513 entry identified by pointer VARS is not cached, e.g., when entries in operand stack 1512 have displaced local variable storage 1513 from a stack cache 155. In accordance with the exemplary embodiment described therein, a stack cache and cache control mechanism representing at least a portion of operand stack 1512 and local variable storage 1513 may advantageously assert suppress folding signal suppress_fold if fold-relevant entries of local variable storage 1513 or operand stack 1512 are not represented in stack cache 155.
  • FIG. 21 illustrates fold decode portion 2004 of instruction decoder 1818 in greater detail. A fold determination portion 2104 selectively inspects the sequence of bytecodes in instruction buffer 1816. If the next bytecode and one or more subsequent bytecodes represent a foldable sequence of operations (as discussed below with respect to FIG. 22), then fold determination portion 2104 supplies a fold-indicating signal on fold line f/nf and a folded instruction decode signal f_instr_dec that specifies an equivalent folded operation. Folded instruction decode signal f_instr_dec is supplied to execution unit 1820 as the decoded instruction instr_dec. In an exemplary embodiment, a foldable sequence of operations includes those associated with 2, 3, or 4 bytecodes from instruction decoder 1818 (up to 2 bytecodes loading operands onto operand stack 1512, a bytecode popping the operand(s), operating thereupon, and pushing a result onto operand stack 1512, and a bytecode popping the result from operand stack 1512 and storing the result. The equivalent folded operation, which is encoded by the folded instruction decode signal f_instr_dec, specifies an operation, that when combined with folded execution addressing information obviates the loads to, and stores from, operand stack 1512.
  • Alternative embodiments may fold only two instructions, e.g., an instruction iload into an instruction iadd or an instruction istore back into an immediately prior instruction iadd. Other alternative embodiments may fold only instructions that push operands onto the operand stack, e.g., one or more instructions iload folded into an instruction iadd, or only instructions that pop results from the operand stack, e.g., an instruction istore back into an immediately prior instruction iadd. Further alternative embodiments may fold larger numbers of instructions that push operands onto the operand stack and/or instructions that pop results from the operand stack instructions in accordance with instructions of a particular virtual machine instruction set. In such alternative embodiments, the above described advantages over normal decoding and execution of instruction sequences are still obtained.
  • Fold determination portion 2104 generates a series of fold address index composite signal f_adr_ind including component first operand index signal first_adr_ind, second operand index signal second_adr_mind, and destination index signal dest_adr_mind, which are respectively selective for a first operand address, a second operand address, and a destination address for the equivalent folded operation. Fold determination portion 2104 provides the composite signal f_adr_ind to fold address generator 2102 for use in supplying operand and destination addresses for the equivalent folded operation. Fold determination portion 2104 asserts a fold-indicating signal on fold line f/nf to control the switches 2006, 2010 and 2008 (see FIG. 20) to provide the signals f_instr_dec, f_adr, and f_valid, as signals instr_dec, instr_adr, and instr_valid, respectively. Otherwise respective signals are provided to execution unit 1820 from normal decode portion 2002.
  • The operation of fold determination portion 2104 is now described with reference to the flowchart of FIG. 22. At start 2201, fold determination portion 2104 begins an instruction decode cycle and transfers processing to initialize index 2202. In initialize index 2202, an instruction index instr_index into instruction buffer 1816 is initialized to identify the next bytecode of a bytecode sequence in instruction buffer 1816. In an exemplary embodiment, instruction index instr_index is initialed to one (1) and the next bytecode is the first bytecode in instruction buffer 1816 since prior bytecodes have already been shifted out of instruction buffer 1816, although a variety of other indexing and instruction buffer management schemes would also be suitable. Upon completion, initialize index 2202 transfers processing to first instruction check 2204.
  • In first instruction check 2204, fold determination portion 2104 determines whether the instruction identified by index instr_index, i.e., the first bytecode, corresponds to an operation that pushes a value, e.g., an integer value, a floating point value, a reference value, etc., onto operand stack 1512. Referring illustratively to a JAVA virtual machine embodiment, first instruction check 2204 determines whether the instruction identified by index instr_index is one that the JAVA virtual machine specification (see Section I) defines as for pushing a first data item onto the operand stack. If so, first operand index signal first_adr_mind is asserted (at first operand address setting 2206) to identify the source of the first operand value. In an exemplary embodiment, first operand index signal first_adr_mind is selective for one of OPTOP, OPTOP-1, OPTOP-2, VARS, VARS+1, VARS+2, and VARS+3, although alternative embodiments may encode larger, smaller, or different sets of source addresses, including for example, source addresses in constant area 1514. Depending on the bytecodes which follow, this first bytecode may correspond to an operation which can be folded into the execution of a subsequent operation. However, if the first bytecode does not meet the criteria of first instruction check 2204, folding is not appropriate and fold determination portion 2104 supplies a nonfold-indicating signal on fold line f/nf, whereupon indications from normal decode portion 2002 provide the decoding.
  • Assuming the first bytecode meets the criteria of first instruction check 2204, index instr_index is incremented (at incrementing 2208) to point to the next bytecode in instruction buffer 1816. Then, at second instruction check 2210, fold determination portion 2104 determines whether instruction identified by index instr_index, i.e., the second bytecode, corresponds to an operation that pushes a value, e.g., an integer value, a floating point value, a reference value, etc., onto operand stack 1512. Referring illustratively to a JAVA virtual machine embodiment, second instruction check 2210 determines whether the instruction identified by index instr_index is one that the JAVA virtual machine specification (see Section I) defines as for pushing a first data item onto the operand stack. If so, second operand index signal second_adr_mind is asserted (at second operand address setting 2212) to indicate the source of the second operand value and index instr_index is incremented (at incrementing 2214) to point to the next bytecode in instruction buffer 1816. As before, second operand index signal second_adr_mind is selective for one of OPTOP, OPTOP-1, OPTOP-2, VARS, VARS+1, VARS+2, and VARS+3, although alternative embodiments are also suitable. Fold determination portion 2104 continues at third instruction check 2216 with index instr_index pointing to either the second or third bytecode in instruction buffer 1816.
  • At third instruction check 2216, fold determination portion 2104 determines whether the instruction identified by index instr_index, i.e., either the second or third bytecode, corresponds to an operation that operates on an operand value or values, e.g., integer value (s), floating point value(s), reference value(s), etc., from the uppermost entries of operand stack 1512, effectively popping such operand values from operand stack 1512 and pushing a result value onto operand stack 1512. Popping of operand values may be explicit or merely a net effect of writing the result value to an upper entry of operand stack 1512 and updating pointer OPTOP to identify that entry as the top of operand stack 1512. Referring illustratively to a JAVA virtual machine embodiment, third instruction check 2216 determines whether the instruction identified by index instr_index corresponds to an operation that the JAVA virtual machine specification (see Section I) defines as for popping a data item (or items) from the operand stack, for operating on the popped data item(s), and for pushing a result of the operation onto the operand stack. If so, index instr_index is incremented (at incrementing 2218) to point to the next bytecode in instruction buffer 1816. If not, folding is not appropriate and fold determination portion 2104 supplies a nonfold-indicating signal on fold line f/nf, whereupon normal decode portion 2002 provides decoding.
  • At fourth instruction check 2220, fold determination portion 2104 determines whether the instruction identified by index instr_index, i.e., either the third or fourth bytecode, corresponds to an operation that pops a value from operand stack 1512 and stores the value in a data store such as local variable storage 1513. Referring illustratively to a JAVA virtual machine embodiment, fourth instruction check 2220 determines whether the instruction identified by index instr_index corresponds to an operation that the JAVA virtual machine specification (see Section I) defines as for popping the result data item from the operand stack. If so, index signal dest_adr_mind is asserted (at destination address setting 2222) to identify the destination of the result value of the equivalent folded operation. Otherwise, if the bytecode at instruction buffer 1816 location identified by index instr_index does not match the criterion of fourth instruction check 2220, index signal dest_adr_mind is asserted (at destination address setting 1824) to identify the top of operand stack 1512. Referring illustratively to a JAVA virtual machine embodiment, if the instruction identified by index instr-index does not match the criterion of fourth instruction check 2220, index signal dest_adr_mind is asserted (at destination address setting 1824) to identify the pointer OPTOP, whether the top of operand stack 1512 or a store operation destination is selected, the folded instruction valid signal f_valid is asserted (at valid fold asserting 1826) and a fold-indicating signal on line f/nf is supplied to select fold decode inputs of switches 2006, 2008, and 2010 for supply to execution unit 1820. Fold determination portion 2104 ends an instruction decode cycle at finish 2250.
  • As a simplification, an instruction decoder for hardware processor 100, e.g., instruction decoder 135, may limit fold decoding to instruction sequences of two instructions and/or to sequences of single bytecode instructions. Those of skill in the art will appreciate suitable simplifications to fold decode portion 2004 of instruction decoder 1818.
  • FIG. 23 shows fold address generator 2102 including three component address generators, first operand address generator 2302, second operand address generator 2304, and destination address generator 2306, respectively supplying a corresponding first operand, second operand, and destination address based on indices supplied thereto and pointer VARS and pointer OPTOP values from pointer registers 1522. In an exemplary embodiment, first operand address generator 2302, second operand address generator 2304, and destination address generator 2306 supply addresses in randomly-accessible storage 1510 corresponding to a subset of operand stack 1512 and local variable storage 1513 entries. Alternative embodiments may supply identifiers selective for storage other than random access memory, e.g., physical registers, which in a particular JAVA virtual machine implementation provide underlying operand stack and local variable storage.
  • First operand address generator 2302 receives first operand index signal first_adr_mind from fold determination portion 2104 and, using pointer VARS and pointer OPTOP values from pointer registers 1522, generates a first operand address signal first_op_adr for a first operand for the equivalent folded operation. The operation of second operand address generator 2304 and destination address generator 2306 is analogous. Second operand address generator 2304 receives first operand index signal first_adr_ind and generates a second operand address signal second_op_adr for a second operand (if any) for the equivalent folded operation. Destination address generator 2306 receives the destination index signal dest_ad_ind and generates the destination address signal dest_adr for the result of the equivalent folded operation. In the embodiment of FIGS. 20, 21, and 23, first operand address signal first_op—adr, second operand address signal second_op_adr, and destination address signal dest_adr are collectively supplied to switch 2010 as fold address signal f_adr for supply to execution unit 1820 as the first operand, second operand, and destination addresses for the equivalent folded operation.
  • FIG. 24 illustrates an exemplary embodiment of first operand address generator 2302. Second operand address generator 2304 and destination address generator 2306 are analogous. In the exemplary embodiment of FIG. 24, first operand address signal first_op_adr is selected from a subset of locations in local variable storage 1513 and operand stack 1512. Alternative embodiments may generate operand and destination addresses from a larger, smaller, or different subset of operand stack 1512 and local variable storage 1513 locations or from a wider range of locations in randomly-accessible storage 1510. For example, alternative embodiments may generate addresses selective for location in constant area 1514. Suitable modifications to the exemplary embodiment of FIG. 24 will be apparent to those of skill in the art. First operand address generator 2302, second operand address generator 2304, and destination address generator 2306 may advantageously define differing sets of locations. For example, whereas locations in constant area 1514 and in the interior of operand stack 1512 are valid as operand sources, they are not typically appropriate result targets. For this reason, the set of locations provided by an exemplary embodiment of destination address generator 2306 is restricted to local variable storage 1513 entries and uppermost entries of operand stack 1512, although alternative sets are also possible. Referring to FIG. 24, pointer OPTOP is supplied to register 2402, which latches the value and provides the latched value to a first input of a data selector 2450. Similarly, pointer OPTOP is supplied to registers 2404 and 2406, which latch the value minus one and minus two, respectively, and provide the latched values to second and third inputs of data selector 2450. In this way, addresses identified by values OPTOP, OPTOP-1, and OPTOP-2 are available for selection by data selector 2450. Similarly, pointer VARS is supplied to a series of registers 2408, 2410, 2412 and 2414, which respectively latch the values VARS, VARS+1, VARS+2, and VARS+3 for provision to the fourth, fifth, sixth, and seventh inputs of data selector 2450. In this way, addresses identified by values VARS, VARS+1, VARS+2, and VARS+3 are available for selection by data selector 2450. In the exemplary embodiment described herein, offsets from pointer VARS are positive because local variable storage 1513 is addressed from its base (identified by pointer VARS) while offsets to pointer OPTOP are negative because operand stack 1512 is addressed from its top (identified by pointer OPTOP).
  • Data selector 2450 selects from among the latched addresses available at its inputs. In an embodiment of fold determination portion 2104 in accordance with the FIG. 24 embodiment of first operand address generator 2302, load source addresses in local variable storage 1513 other than those addressed by values VARS, VARS+1, VARS+2, and VARS+3 are handled as unfoldable and decoded via normal decode portion 2002. However, suitable modifications for expanding the set of load addresses supported will be apparent to those of skill in the art. Second operand address generator 2304 and destination address generator 2306 are of analogous design, although destination address generator 2306 does not provide support for addressing into constant area 1514. In one embodiment in accordance with the present invention, signal RS1_D is supplied to the zeroth input of data selector 2450. In this embodiment, additional decode logic (not shown) allows for direct supply of register identifier information to support an alternate instruction set. Addition decode logic support for such an alternate instruction set is described in greater detail in a U.S. Pat. No. 5,925,123 entitled “A PROCESSOR FOR EXECUTING INSTRUCTION SETS RECEIVED FROM A NETWORK OR FROM A LOCAL MEMORY” naming Marc Tremblay and James Michael O'Connor as inventors, assigned to the assignee of this application, and filed on Jan. 23, 1997, the detailed description of which is incorporated herein by reference.
  • Referring back to FIG. 20, when fold determination portion 2104 of fold decode portion 2004 identifies a foldable bytecode sequence, fold determination portion 2104 asserts a fold-indicating signal on line f/nf, supplies an equivalent folded operation as folded instruction decode signal f_instr_dec, and supplies, based on load and store instructions from the foldable bytecode sequence, indices into latched addresses maintained by first operand address generator 2302, second operand address generator 2304, and destination address generator 2306. Fold decode portion 2004 supplies the addresses so indexed as folded address signal f-adr. Responsive to the signal on line tinf, switches 2006, 2008, 2010 supply decode information for the equivalent folded operation to execution unit 1820.
  • Although fold decode portion 2004 has been described above in the context of an exemplary four instruction foldable sequence, it is not limited thereto. Based on the description herein, those of skill in the art will appreciate suitable extensions to support folding of additional instructions and longer foldable instruction sequences, e.g., sequences of five or more instructions. By way of example and not of limitation, support for additional operand address signals, e.g., a third operand address signal, and/or for additional destination address signals, e.g., a second destination address signal, could be provided.
  • Integer Execution Unit
  • Integer execution unit IEU, that includes instruction decode unit 130, integer unit 142, and stack management unit 150, is responsible for the execution of all the virtual machine instructions except the floating point related instructions. The floating point related instructions are executed in floating point unit 143.
  • Integer execution unit IEU interacts at the front end with instructions cache unit 120 to fetch instructions, with floating point unit (FPU) 143 to execute floating point instructions, and finally with data cache unit (ECU) 160 to execute load and store related instructions. Integer execution unit IEU also contains microcode ROM 141 which contains instructions to execute certain virtual machine instructions associated with integer operations.
  • Integer execution unit IEU includes a cached portion of stack 400, i.e., stack cache 155. Stack cache 155 provides fast storage for operand stack and local variable entries associated with a current method, e.g., operand stack 423 and local variable storage 421 entries. Although, stack cache 155 may provide sufficient storage for all operand stack and local variable entries associated with a current method, depending on the number of operand stack and local variable entries, less than all of local variable entries or less than all of both local variable entries and operand stack entries may be represented in stack cache 155. Similarly, additional entries, e.g., operand stack and or local variable entries for a calling method, may be represented in stack cache 155 if space allows.
  • Stack cache 155 is a sixty-four entry thirty-two-bit wide array of registers that is physically implemented as a register file in one embodiment. Stack cache 155 has three read ports, two of which are dedicated to integer execution unit IEU and one to dribble manager unit 151. Stack cache 155 also has two write ports, one dedicated to integer execution unit IEU and one to dribble manager unit 151.
  • Integer unit 142 maintains the various pointers, which are used to access variables, such as local variables, and operand stack values, in stack cache 155. Integer unit 142 also maintains pointers to detect whether a stack cache hit has taken place. Runtime exceptions are caught and dealt with by exception handlers that are implemented using information in microcode ROM 141 and circuit 170.
  • Integer unit 142 contains a 32-bit ALU to support arithmetic operations. The operations supported by the ALU include: add, subtract, shift, and, or, exclusive or, compare, greater than, less than, and bypass. The ALU is also used to determine the address of conditional branches while a separate comparator determines the outcome of the branch instruction.
  • The most common set of instructions which executes cleanly through the pipeline is the group of ALU instructions. The ALU instructions read the operands from the top of stack 400 in decode stage 302 and use the ALU in execution stage 303 to compute the result. The result is written back to stack 400 in write-back stage 305. There are two levels of bypass which may be needed if consecutive ALU operations are accessing stack cache 155.
  • Since the stack cache ports are 32-bits wide in this embodiment, double precision and long data operations take two cycles. A shifter is also present as part of the ALU. If the operands are not available for the instruction in decode stage 302, or at a maximum at the beginning of execution stage 303, an interlock holds the pipeline stages before execution stage 303.
  • The instruction cache unit interface of integer execution unit IEU is a valid/accept interface, where instruction cache unit 120 delivers instructions to instruction decode unit 130 in fixed fields along with valid bits. Instruction decoder 135 responds by signaling how much byte aligner circuit 122 needs to shift, or how many bytes instruction decode unit 130 could consume in decode stage 302. The instruction cache unit interface also signals to instruction cache unit 120 the branch mis-predict condition, and the branch address in execution stage 303. Traps, when taken, are also similarly indicated to instruction cache unit 120. Instruction cache unit 120 can hold integer unit 142 by not asserting any of the valid bits to instruction decode unit 130. Instruction decode unit 130 can hold instruction cache unit 120 by not asserting the shift signal to byte aligner circuit 122.
  • The data cache interface of integer execution unit IEU also is a valid-accept interface, where integer unit 142 signals, in execution stage 303, a load or store operation along with its attributes, e.g., non-cached, special stores etc., to data cache controller 161 in data cache unit 160. Data cache unit 160 can return the data on a load, and control integer unit 142 using a data control unit hold signal. On a data cache hit, data cache-unit 160 returns the requested data, and then releases the pipeline.
  • On store operations, integer unit 142 also supplies the data along with the address in execution stage 303. Data cache unit 160 can hold the pipeline in cache stage 304 if data cache unit 160 is busy, e.g., doing a line fill etc.
  • Floating point operations are dealt with specially by integer execution unit IEU. Instruction decoder 135 fetches and decodes floating point unit 143 related instructions. Instruction decoder 135 sends the floating point operation operands for execution to floating point unit 142 in decode state 302. While floating point unit 143 is busy executing the floating point operation, integer unit 142 halts the pipeline and waits until floating point unit 143 signals to integer unit 142 that the result is available.
  • A floating point ready signal from floating point unit 143 indicates that execution stage 303 of the floating point operation has concluded. In response to the floating point ready signal, the result is written back into stack cache 155 by integer unit 142, Floating point load and stores are entirely handled by integer execution unit IEU, since the operands for both floating point unit 143 and integer unit 142 are found in stack cache 155.
  • Stack Management Unit
  • A stack management unit 150 stores information, and provides operands to execution unit 140. Stack management unit 150 also takes care of overflow and underflow conditions of stack cache 155.
  • In one embodiment, stack management unit 150 includes stack cache 155 that, as described above, a three read port, two write port register file in one embodiment; a stack control unit 152 which provides the necessary control signals for two read ports and one write port that are used to retrieve operands for execution unit 140 and for storing data back from a write-back register or data cache 165 into stack cache 155; and a dribble manager 151 which speculatively dribbles data in and out of stack cache 155 into memory whenever there is an overflow or underflow in stack cache 155. In the exemplary embodiment of FIG. 1, memory includes data cache 165 and any memory storage interfaced by memory interface unit 110. In general, memory includes any suitable memory hierarchy including caches, addressable read/write memory storage, secondary storage, etc. Dribble manager 151 also provides the necessary control signals for a single read port and a single write port of stack cache 155 which are used exclusively for background dribbling purposes.
  • In one embodiment, stack cache 155 is managed as a circular buffer which ensures that the stack grows and shrinks in a predictable manner to avoid overflows or overwrites. The saving and restoring of values to and from data cache 165 is controlled by dribbler manager 151 using high- and low-water marks, in one embodiment.
  • Stack management unit 150 provides execution unit 140 with two 32-bit operands in a given cycle. Stack management unit 150 can store a single 32-bit result in a given cycle.
  • Dribble manager 151 handles spills and fills of stack cache 155 by speculatively dribbling the data in and out of stack cache 155 from and to data cache 165. Dribble manager 151 generates a pipeline stall signal to stall the pipeline when a stack overflow or underflow condition is detected. Dribble manager 151 also keeps track of requests sent to data cache unit 160. A single request to data cache unit 160 is a 32-bit consecutive load or store request.
  • The hardware organization of stack cache 155 is such that, except for long operands (long integers and double precision floating-point numbers), implicit operand fetches for opcodes do not add latency to the execution of the opcodes. The number of entries in operand stack 423 (FIG. 4A) and local variable storage 421 that are maintained in stack cache 155 represents a hardware/performance tradeoff. At least a few operand stack 423 and local variable storage 421 entries are required to get good performance. In the exemplary embodiment of FIG. 1, at least the top three entries of operand stack 423 and the first four local variable storage 421 entries are preferably represented in stack cache 155.
  • One key function provided by stack cache 155 (FIG. 1) is to emulate a register file where access to the top two registers is always possible without extra cycles. A small hardware stack is sufficient if the proper intelligence is provided to load/store values from/to memory in the background, therefore preparing stack cache 155 for incoming virtual machine instructions.
  • As indicated above, all items on stack 400 (regardless of size) are placed into a 32-bit word. This tends to waste space if many small data items are used, but it also keeps things relatively simple and free of lots of tagging or muxing. An entry in stack 400 thus represents a value and not a number of bytes. Long integer and double precision floating-point numbers require two entries. To keep the number of read and write ports low, two cycles to read two long integers or two double precision floating point numbers are required.
  • The mechanism for filling and spilling the operand stack from stack cache 155 out to memory by dribble manager 151 can assume one of several alternative forms. One register at a time can be filled or spilled, or a block of several registers filled or spilled at once. A simple scoreboarded method is appropriate for stack management. In its simplest form, a single bit indicates if the register in stack cache 155 is currently valid. In addition, some embodiments of stack cache 155 use a single bit to indicate whether the data content of the register is saved to stack 400, i.e., whether the register is dirty. In one embodiment, a high-water mark/low-water mark heuristic determines when entries are saved to and restored from stack 400, respectively (FIG. 4A). Alternatively, when the top-of-the-stack becomes close to bottom 401 of stack cache 155 by a fixed, or alternatively, a programmable number of entries, the hardware starts loading registers from stack 400 into stack cache 155. Detailed embodiments of stack management unit 150 and dribble manager unit 151 are described below and in U.S. patent application Ser. No. 08/787,736, entitled “STACK MANAGEMENT UNIT AND METHOD FOR A PROCESSOR HAVING A STACK” naming Marc Tremblay and James Michael O'Connor as inventors, assigned to the assignee of this application, and filed on Jan. 23, 1997 with Attorney Docket No. SP2037, now U.S. Pat. No. 6,038,643, which is incorporated herein by reference in its entirety.
  • In one embodiment, stack management unit 150 also includes an optional local variable look-aside cache 153. Cache 153 is most important in applications where both the local variables and operand stack 423 (FIG. 4A) for a method are not located on stack cache 155. In such instances when cache 153 is not included in hardware processor 100, there is a miss on stack cache 155 when a local variable is accessed, and execution unit 140 accesses data cache unit 160, which in turn slows down execution. In contrast, with cache 153, the local variable is retrieved from cache 153 and there is no delay in execution.
  • One embodiment of local variable, look-aside cache 153 is illustrated in FIG. 4D for method 0 to 2 on stack 400. Local variables zero to M, where M is an integer, for method 0 are stored in plane 421A_0 of cache 153 and plane 421A_0 is accessed when method number 402 is zero. Local variables zero to N, where N is an integer, for method 1 are stored in plane 421A_1 of cache 153 and plane 421A_1 is accessed when method number 402 is one. Local variables zero to P, where P is an integer, for method 1 are stored in plane 421A_2 of cache 153 and plane 421A_2 is accessed when method number 402 is two. Notice that the various planes of cache 153 may be different sizes, but typically each plane of the cache has a fixed size that is empirically determined.
  • When a new method is invoked, e.g, method 2, a new plane 421A_2 in cache 153 is loaded with the local variables for that method, and method number register 402, which in one embodiment is a counter, is changed, e.g., incremented, to point to the plane of cache 153 containing the local variables for the new method. Notice that the local variables are ordered within a plane of cache 153 so that cache 153 is effectively a direct-mapped cache. Thus, when a local variable is needed for the current method, the variable is accessed directly from the most recent plane in cache 153, i.e., the plane identified by method number 402. When the current method returns, e.g., method 2, method number register 402 is changed, e.g., decremented, to point at previous plane 421A_1 of cache 153. Cache 153 can be made as wide and as deep as necessary.
  • Data Cache Unit
  • Data cache unit 160 (DCU) manages all requests for data in data cache 165. Data cache requests can come from dribbling manager 151 or execution unit 140. Data cache controller 161 arbitrates between these requests giving priority to the execution unit requests. In response to a request, data cache controller 161 generates address, data and control signals for the data and tags RAMs in data cache 165. For a data cache hit, data cache controller 161 reorders the data RAM output to provide the right data.
  • Data cache controller 161 also generates requests to I/O bus and memory interface unit 110 in case of data cache misses, and in case of non-cacheable loads and stores. Data cache controller 161 provides the data path and control logic for processing non-cacheable requests, and the data path and data path control functions for handling cache misses.
  • For data cache hits, data cache unit 160 returns data to execution unit 140 in one cycle for loads. Data cache unit 160 also takes one cycle for write hits. In case of a cache miss, data cache unit 160 stalls the pipeline until the requested data is available from the external memory. For both non-cacheable loads and stores, data cache 165 is bypassed and requests are sent to I/O bus and memory interface unit 110. Non-aligned loads and stores to data cache 165 trap in software.
  • Data cache 165 is a two-way set associative, write back, write allocate, 16-byte line cache. The cache size is configurable to 0, 1, 2, 4, 8, 16 Kbyte sizes. The default size is 6 Kbytes. Each line has a cache tag store entry associated with the line. On a cache miss, 16 bytes of data are written into cache 165 from external memory.
  • Each data cache tag contains a 20-bit address tag field, one valid bit, and one dirty bit. Each cache tag is also associated with a least recently used bit that is used for replacement policy. To support multiple cache sizes, the width of the tag fields also can be varied. If a cache enable bit in processor service register is not set, loads and stores are treated like non-cacheable instructions by data cache controller 161.
  • A single sixteen-byte write back buffer is provided for writing back dirty cache lines which need to be replaced. Data cache unit 160 can provide a maximum of four bytes on a read and a maximum of four bytes of data can be written into cache 165 in a single cycle. Diagnostic reads and writes can be done on the caches.
  • Memory Allocation Accelerator
  • In one embodiment, data cache unit 160 includes a memory allocation accelerator 166. Typically, when a new object is created, fields for the object are fetched from external memory, stored in data cache 165 and then the field is cleared to zero. This is a time consuming process that is eliminated by memory allocation accelerator 166. When a new object is created, no fields are retrieved from external memory. Rather, memory allocation accelerator 166 simply stores line of zeros in data cache 165 and marks that line of data cache 165 as dirty. Memory allocation accelerator 166 is particularly advantageous with a write-back cache. Since memory allocation accelerator 166 eliminates the external memory access each time a new object is created, the performance of hardware processor 100 is enhanced.
  • Floating, Point Unit
  • Floating point unit (FPU) 143 includes a microcode sequencer, input/output section with input/output registers, a floating point adder, i.e., an ALU, and a floating point multiply/divide unit. The microcode sequencer controls the microcode flow and microcode branches. The input/output section provides the control for input/output data transactions, and provides the input data loading and output data unloading registers. These registers also provide intermediate result storage.
  • The floating point adder ALU includes the combinatorial logic used to perform the floating point adds, floating point subtracts, and conversion operations. The floating point multiply/divide unit contains the hardware for performing multiply/divide and remainder.
  • Floating point unit 143 is organized as a microcoded engine with a 32-bit data path. This data path is often reused many times during the computation of the result. Double precision operations require approximately two to four times the number of cycles as single precision operations. The floating point ready signal is asserted one-cycle prior to the completion of a given floating point operation. This allows integer unit 142 to read the floating point unit output registers without any wasted interface cycles. Thus, output data is available for reading one cycle after the floating point ready signal is asserted.
  • Execution Unit Accelerators
  • Since the JAVA Virtual Machine Specification of Section I is hardware independent, the virtual machine instructions are not optimized for a particular general type of processor, e.g., a complex instruction-set computer (CISC) processor, or a reduced instruction set computer (RISC) processor. In fact, some virtual machine instructions have a CISC nature and others a RISC nature. This dual nature complicates the operation and optimization of hardware processor 100.
  • For example, the JAVA virtual machine specification defines opcode 171 for an instruction lookupswitch, which is a traditional switch statement. The datastream to instruction cache unit 320 includes an opcode 171, identifying the N-way switch statement, that is followed zero to three bytes of padding. The number of bytes of padding is selected so that first operand byte begins at an address that is a multiple of four. Herein, datastream is used generically to indicate information that is provided to a particular element, block, component, or unit.
  • Following the padding bytes in the datastream are a series of pairs of signed four-byte quantities. The first pair is special. A first operand in the first pair is the default offset for the switch statement that is used when the argument, referred to as an integer key, or alternatively, a current match value, of the switch statement is not equal to any of the values of the matches in the switch statement. The second operand in the first pair defines the number of pairs that follow in the datastream.
  • Each subsequent operand pair in the datastream has a first operand that is a match value, and a second operand that is an offset. If the integer key is equal to one of the match values, the offset in the pair is added to the address of the switch statement to define the address to which execution branches. Conversely, if the integer key is unequal to any of the match values, the default offset in the first pair is added to the address of the switch statement to define the address to which execution branches. Direct execution of this virtual machine instruction requires many cycles.
  • To enhance the performance of hardware processor 100, a look-up switch accelerator 145 is included in hardware processor 100. Look-up switch accelerator 145 includes an associative memory which stores information associated with one or more lookup switch statements. For each lookup switch statement, i.e., each instruction lookupswitch, this information includes a lookup switch identifier value, i.e., the program counter value associated with the lookup switch statement, a plurality of match values and a corresponding plurality of jump offset values.
  • Lookup switch accelerator 145 determines whether a current instruction received by hardware processor 100 corresponds to a lookup switch statement stored in the associative memory. Lookup switch accelerator 145 further determines whether a current match value associated with the current instruction corresponds with one of the match values stored in the associative memory. Lookup switch accelerator 145 accesses a jump offset value from the associative memory when the current instruction corresponds to a lookup switch statement stored in the memory and the current match value corresponds with one of the match values stored in the memory wherein the accessed jump offset value corresponds with the current match value.
  • Lookup switch accelerator 145 further includes circuitry for retrieving match and jump offset values associated with a current lookup switch statement when the associative memory does not already contain the match and jump offset values associated with the current lookup switch statement. Lookup switch accelerator 145 is described in more detail in U.S. patent application Ser. No. 08/788,811, entitled “LOOK-UP SWITCH ACCELERATOR AND METHOD OF OPERATING SAME” naming Marc Tremblay and James Michael O'Connor as inventors, assigned to the assignee of this application, and filed on Jan. 23, 1997 with Attorney Docket No. SP2040, now U.S. Pat. No. 6,076,141, which is incorporated herein by reference in its entirety.
  • In the process of initiating execution of a method of an object, execution unit 140 accesses a method vector to retrieve one of the method pointers in the method vector, i.e., one level of indirection. Execution unit 140 then uses the accessed method pointer to access a corresponding method, i.e., a second level of indirection.
  • To reduce the levels of indirection within execution unit 140, each object is provided with a dedicated copy of each of the methods to be accessed by the object. Execution unit 140 then accesses the methods using a single level of indirection. That is, each method is directly accessed by a pointer which is derived from the object. This eliminates a level of indirection, which was previously introduced by the method pointers. By reducing the levels of indirection, the operation of execution unit 140 can be accelerated. The acceleration of execution unit 140 by reducing the levels of indirection experienced by execution unit 140 is described in more detail in U.S. patent application Ser. No. 08/787,846, entitled “REPLICATING CODE TO ELIMINATE A LEVEL OF INDIRECTION DURING EXECUTION OF AN OBJECT ORIENTED COMPUTER PROGRAM” naming Marc Tremblay and James Michael O'Connor as inventors, assigned to the assignee of this application, and filed on Jan. 23, 1997 with Attorney Docket No. SP2043, now U.S. Pat. No. 5,970,242, which is incorporated herein by reference in its entirety.
  • Getfield-Putfield Accelerator
  • Other specific functional units and various translation lookaside buffer (TLB) types of structures may optionally be included in hardware processor 100 to accelerate accesses to the constant pool. For example, the JAVA virtual machine specification defines an instruction putfield, opcode 181, that upon execution sets a field in an object and an instruction getfield, opcode 180, that upon execution fetches a field from an object. In both of these instructions, the opcode is followed by an index byte one and an index byte two. Operand stack 423 contains a reference to an object followed by a value for instruction putfield, but only a reference to an object for instruction getfield.
  • Index bytes one and two are used to generate an index into the constant pool of the current class. The item in the constant pool at that index is a field reference to a class name and a field name. The item is resolved to a field block pointer which has both the field width, in bytes, and the field offset, in bytes.
  • An optional getfield-putfield accelerator 146 in execution unit 140 stores the field block pointer for instruction getfield or instruction putfield in a cache, for use after the first invocation of the instruction, along with the index used to identify the item in the constant pool that was resolved into the field block pointer as a tag. Subsequently, execution unit 140 uses index bytes one and two to generate the index and supplies the index to getfield-putfield accelerator 146. If the index matches one of the indexes stored as a tag, i.e., there is a hit, the field block pointer associated with that tag is retrieved and used by execution unit 140. Conversely, if a match is not found, execution unit 140 performs the operations described above. Getfield-putfield accelerator 146 is implemented without using self-modifying code that was used in one embodiment of the quick instruction translation described above.
  • In one embodiment, getfield-putfield accelerator 146 includes an associative memory that has a first section that holds the indices that function as tags, and a second section that holds the field block pointers. When an index is applied through an input section to the first section of the associative memory, and there is a match with one of the stored indices, the field block pointer associated with the stored index that matched in input-index is output from the second section of the associative memory.
  • Bounds Check Unit
  • Bounds check unit 147 (FIG. 1) in execution unit 140 is an optional hardware circuit that checks each access to an element of an array to determine whether the access is to a location within the array. When the access is to a location outside the array, bounds check unit 147 issues an active array bound exception signal to execution unit 140. In response to the active array bound exception signal, execution unit 140 initiates execution of an exception handler stored in microcode ROM 141 that in handles the out of bounds array access.
  • In one embodiment, bounds check unit 147 includes an associative memory element in which is stored a array identifier for an array, e.g., a program counter value, and a maximum value and a minimum value for the array. When an array is accessed, i.e., the array identifier for that array is applied to the associative memory element, and assuming the array is represented in the associative memory element, the stored minimum value is a first input signal to a first comparator element, sometimes called a comparison element, and the stored maximum value is a first input signal to a second comparator element, sometimes also called a comparison element. A second input signal to the first and second comparator elements is the value associated with the access of the array's element.
  • If the value associated with the access of the array's element is less than or equal to the stored maximum value and greater than or equal to the stored minimum value, neither comparator element generates an output signal. However, if either of these conditions is false, the appropriate comparator element generates the active array bound exception signal. A more detailed description of one embodiment of bounds check unit 147 is provided in U.S. patent application Ser. No. 08/786,352, entitled “PROCESSOR WITH ACCELERATED ARRAY ACCESS BOUNDS CHECKING” naming Marc Tremblay, James Michael O'Connor, and William N. Joy as inventors, assigned to the assignee of this application, and filed on Jan. 23, 1997 with Attorney Docket No. SP2041, now U.S. Pat. No. 6,014,723, which is incorporated herein by reference in its entirety.
  • The JAVA Virtual Machine Specification defines that certain instructions can cause certain exceptions. The checks for these exception conditions are implemented, and a hardware/software mechanism for dealing with them is provided in hardware processor 100 by information in microcode ROM 141 and program counter and trap control logic 170. The alternatives include having a trap vector style or a single trap target and pushing the trap type on the stack so that the dedicated trap handler routine determines the appropriate action.
  • No external cache is required for the architecture of hardware processor 100. No translation lookaside buffers need be supported.
  • FIG. 5 illustrates several possible add-ons to hardware processor 100 to create a unique system. Circuits supporting any of the eight functions shown, i.e., NTSC encoder 501, MPEG 502, Ethernet controller 503, VIS 504, ISDN 505, I/O controller 506, ATM assembly/reassembly 507, and radio link 508 can be integrated into the same chip as hardware processor 100 of this invention.
  • FIG. 6 is a block diagram of one embodiment of a stack management unit 150. Stack management unit 150 serves as a high speed buffer between stack 400 and hardware processor 100. Hardware processor 100 accesses stack management unit 150 as if stack management unit 150 were stack 400. Stack management unit 150 automatically transfers data between stack management unit 150 and stack 400 as necessary to provide improve the throughput of data between stack 400 and hardware processor 100. In the embodiment of FIG. 1, if hardware processor 100 requires a data word which is not cached in stack management unit 150, data cache unit 160 retrieves the requested data word and places the requested data word at the top of stack cache 155.
  • Stack management unit 150 contains a stack cache, memory circuit 610. Stack cache memory circuit 610 is typically fast memory devices such as a register file or SRAM; however, slower memory devices such as DRAM can also be used. In the embodiment of FIG. 6, access to stack cache memory circuit 610 is controlled by stack control unit 152. A write port 630 allows hardware processor 100 to write data on data lines 635 to stack cache memory circuit 610. Read port 640 and read port 650 allow hardware processor 100 to read data from stack cache memory circuit 610 on data lines 645 and 655, respectively. Two read ports are provided to increase throughput since many operations of stack-based computing systems require two operands from stack 400. Other embodiments of stack cache 155 may provide more or less read and write ports.
  • As explained above, dribble manager unit 151 controls the transfer of data between stack 400 (FIG. 4(a)) and stack cache memory circuit 610. In the embodiment shown in FIG. 1, the transfer of data between stack 400 and stack cache memory circuit 610 goes through data cache unit 160. Dribble manager unit 151 includes a fill control unit 694 and a spill control unit 698. In some embodiments of dribble manager unit 151, fill control unit 694 and spill control unit 698 function independently. Fill control unit 694 determines if a fill condition exists. If the fill condition exists, fill control unit 694 transfers data words from stack 400 to stack cache memory circuit 610 on data lines 675 through a write port 670. Spill control unit 698 determines if a spill condition exists. If the spill condition exists, spill control unit 698 transfers data words from stack cache memory circuit 610 to stack 400 through read port 680 on data lines 685. Write port 670 and read port 680 allows transfers between stack 400 and stack cache memory circuit 610 to occur simultaneously with reads and writes controlled by stack control unit 152. If contention for read and write ports of stack cache memory circuit 610 is not important, dribble manager unit 151 can share read and write ports with stack control unit 152.
  • Although stack management unit 150 is described in the context of buffering stack 400 for hardware processor 100, stack management unit 150 can perform caching for any stack-based computing system. The details of hardware processor 100, are provided only as an example of one possible stack-based computing system for use with the present invention. Thus, one skilled in the art can use the principles described herein to design a stack management unit in accordance to the present invention for any stack-based computing system.
  • FIG. 7 shows a conceptual model of the memory architecture of stack cache memory circuit 610 for one embodiment of stack cache 155. Specifically, in the embodiment of FIG. 7, stack cache memory circuit 610 is a register file organized in a circular buffer memory architecture capable of holding 64 data words. Other embodiments may contain a different number of data words. The circular memory architecture causes data words in excess of the capacity of stack cache memory circuit 610 to be written to previously used registers. If stack cache memory unit 610 uses a different memory device, such as an SRAM, different registers would correspond to different memory locations. One technique to address registers in a circular buffer is to use pointers containing modulo stack cache size (modulo-SCS) addresses to the various registers of stack cache memory circuit 610. As used herein, modulo-N operations have the results of the standard operation mapped to a number between 0 and SCS-1 using a standard MOD N function. Some common modulo operations are defined as follows
      • Modulo-N addition of X and Y=(X+Y) MOD N,
      • Modulo-N subtraction of X and Y=(X−Y) MOD N,
      • Modulo-N increment of X by Y=(X+Y) MOD N,
      • Modulo-N decrement of X by Y=(X−Y) MOD N.
  • One embodiment of the pointer addresses of the registers of stack cache memory circuit 610 are shown in FIG. 7 as numbered 0-63 along the outer edge of stack cache memory circuit 610. Thus for the embodiment of FIG. 7, if 70 data words (numbered 1 to 70) are written to stack cache memory circuit 610 when stack cache memory circuit 610 is empty, data words 1 to 64 are written to registers 0 to 63, respectively and data words 65 to 70 are written subsequently to registers 0 to 5. Prior to writing data words 65 to 70, dribble manager unit 151, as described below, transfers data words 1 to 6 which were in registers 0 to 5 to stack 400. Similarly, as data words 70 to 65 are read out of stack cache memory circuit 610, data words 1 to 6 can be retrieved from stack 400 and placed in memory locations 0 to 5.
  • Since most reads and writes on a stack are from the top of the stack, a pointer OPTOP contains the location of the top of stack 400, i.e. the top memory location. In some embodiments of stack management unit 150, pointer OPTOP is a programmable register in execution unit 140. However other embodiments of stack management unit 150 maintain pointer OPTOP in stack control unit 152. Since pointer OPTOP is often increased by one, decreased by one, or changed by a specific amount, pointer OPTOP, in one embodiment is a programmable up/down counter.
  • Since stack management unit 150 contains the top portion of stack 400, pointer OPTOP indicates the register of stack cache memory circuit 610 containing the most recently written data word in stack cache memory circuit 610, i.e. pointer OPTOP points to the register containing the most recently written data word also called the top register. Some embodiments of stack management unit 150 also contains a pointer OPTOPI (not shown) which points to the register preceding the register pointed to by pointer OPTOP. Pointer OPTOPI can improve the performance of stack management unit 150 since many operations in hardware processor 100 require two data words from stack management unit 150.
  • Pointer OPTOP and pointer OPTOP1 are incremented whenever a new data word is written to stack cache 155. Pointer OPTOP and pointer OPTOP1 are decremented whenever a stacked data word, i.e. a data word already in stack 400, is popped off of stack cache 155. Since some embodiments of hardware processor 100 may add or remove multiple data words simultaneously, pointer OPTOP and OPTOP1 are implemented, in one embodiment as programmable registers so that new values can be written into the registers rather than requiring multiple increment or decrement cycles.
  • If stack cache 155 is organized using sequential addressing, pointer OPTOP1 may also be implemented using a modulo SCS subtractor which modulo-SCS subtracts one from pointer OPTOP. Some embodiments of stack cache 155 may also include pointers OPTOP2 or pointer OPTOP3.
  • Since data words are stored in stack cache memory circuit 610 circularly, the bottom of stack cache memory circuit 610 can fluctuate. Therefore, most embodiments of stack cache memory circuit 610 include a pointer CACHE_BOTTOM to indicate the bottom memory location of stack cache memory circuit 610. Pointer CACHE_BOTTOM is typically maintained by dribble manager unit 151. The process to increment or decrement pointer CACHE_BOTTOM varies with the specific embodiment of stack management unit 150. Pointer CACHE_BOTTOM is typically implemented as a programmable up/down counter.
  • Some embodiments of stack management unit 150 also includes other pointers, such as pointer VARS, which points to a memory location of a data word that is often accessed. For example, if hardware processor 100 is implementing the JAVA Virtual Machine, entire method frames may be placed in stack management unit 150. The method frames often contain local variables that are accessed frequently. Therefore, having pointer. VARS pointed to the first local variable of the active method decreases the access time necessary to read the local variable. Other pointers such as a pointer VARS1 (not shown) and a pointer VARS2 (not shown) may point to other often used memory locations such as the next two local variables of the active method in a JAVA Virtual Machine. In some embodiments of stack management unit 150, these pointers are maintained in stack control unit 152. In embodiments adapted for use with hardware processor 100, pointer VARS is stored in a programmable register in execution unit 140. If stack cache 155 is organized using sequential addressing, pointer VARS1 may also be implemented using a modulo-SCS adder which modulo-SCS adds one to pointer VARS.
  • To determine which data words to transfer between stack cache memory circuit 610 and stack 400, stack management unit 150, typically tags, i.e. tracks, the valid data words and the data words which are stored in both stack cache memory circuit 610 and stack 400. FIG. 8 illustrates one tagging scheme used in some embodiments of stack management unit 150. Specifically, FIG. 8 shows a register 810 from stack cache memory circuit 610. The actual data word is stored in data section 812. A valid bit 814 and a saved bit 816 are used to track the status of register 810. If valid bit 814 is at a valid logic state, typically logic high, data section 812 contains a valid data word. If valid bit 814 is at an invalid logic state, typically logic low, data section 812 does not contain a valid data word. If saved bit 816 is at a saved logic state, typically logic high, the data word contained in data section 812 is also stored in stack 400. However, if saved bit 816 is at an unsaved logic state, typically logic low, the data word contained in data section 812 is not stored in stack 400. Typically, when stack management unit 150 is powered up or reset, valid bit 814 of each register is set to the invalid logic state and saved bit 816 of each register is set to the unsaved logic state.
  • For the embodiment illustrated in FIG. 6 using the tagging method of FIG. 8, when stack control unit 152 writes a data word to a register in stack cache memory circuit 610 through write port 630 the valid bit of that register is set to the valid logic state and the saved bit of that register is set to the unsaved logic state. When dribble manager unit 151 transfer a data word to a register of stack cache memory circuit 610 through write port 670, the valid bit of that register is set to the valid logic state and the saved bit of that register is set to the saved logic state since the data word is currently saved in stack 400.
  • When hardware processor 100 reads a stacked data word using a stack popping operation from a register of stack cache memory circuit 610 through either read port 640 or read port 650 the valid bit of that register is set to the invalid logic state and the saved bit of that location is set to the unsaved logic state. Typically, stack popping operations use the register indicated by pointer OPTOP or pointer OPTOP1.
  • When hardware processor 100 reads a data word with a non-stack popping operation from a register of stack cache memory circuit 610 through either read port 640 or read port 650 the valid bit and saved bit of the register are not changed. For example, if hardware processor 100 is implementing the JAVA Virtual Machine, a local variable stored in stack cache memory circuit 610 in the register indicated by pointer VARS may be used repeatedly and should not be removed from stack cache 155. When dribble manager unit 151 copies a data word from a register of stack cache memory circuit 610 to stack 400 through read port 680, the valid bit of that register remains in the valid logic state since the saved data word is still contained in that register and the saved bit of that register is set to the saved logic state.
  • Since stack cache 155 is generally much smaller than the memory address space of hardware processor 100, the pointers used to access stack cache memory circuit 610 are generally much smaller than general memory addresses. The specific technique used to map stack cache 155 into the memory space of hardware processor 100 can vary. In one embodiment of hardware processor 100 the pointers used to access stack cache memory circuit 610 are only the lower bits of general memory pointers, i.e, the least significant bits. For example, if stack cache memory circuit 610 comprises 64 registers, pointers OPTOP, VARS, and CACHE_BOTTOM need only be six bits long. If hardware processor 100 has a 12 bit address space, pointers OPTOP, VARS, and CACHE_BOTTOM could be the lower-six bits of a general memory pointer. Thus stack cache memory circuit 610 is mapped to a specific segment of the address space having a unique upper six bit combination.
  • Some embodiments of stack cache management unit 150 may be used with purely stacked based computing system so that there is not a memory address space for the system. In this situation, the pointers for accessing stack cache 155 are only internal to stack cache management unit 150.
  • As explained above, hardware processor 100 primarily accesses data near the top of the stack. Therefore, stack management unit 150 can improve data accesses of hardware processor 100 while only caching the top portion of stack 400. When hardware processor 100 pushes more data words to stack management unit 150 than stack cache memory circuit 610 is able to store, the data words near the bottom of stack cache memory circuit 610 are transferred to stack 400. When hardware processor 100 pops data words out of stack cache 155, data words from stack 400 are copied under the bottom of stack cache memory circuit 610, and pointer CACHE_BOTTOM is decremented to point to the new bottom of stack cache memory circuit 610.
  • Determination of when to transfer data words between stack 400 and stack cache memory circuit 610 as well as how many data words to transfer can vary. In general, dribble manager unit 151 should transfer data from stack cache memory circuit 610 to stack 400, i.e. a spill operation, as hardware processor fills stack cache memory circuit 610. Conversely, dribble manager unit 151 should copy data from stack 400 to stack cache memory circuit 610, i.e. a fill operation, as hardware processor empties stack cache memory circuit 610.
  • FIG. 9 shows one embodiment of dribble manager unit 151 in which decisions on transferring data from stack cache memory circuit 610 to stack 400, i.e. spilling data, are based on the number of free registers in stack cache memory circuit 610. Free registers includes registers without valid data as well as registers containing data already stored in stack 400, i.e. registers with saved bit 816 set to the saved logic state. Decisions on transferring data from stack 400 to stack cache memory circuit 610, i.e. filling data, are based on the number of used registers. A used registers contains a valid but unsaved data word in stack cache memory circuit 610.
  • Specifically in the embodiment of FIG. 9, dribble manager unit 151 further includes a stack cache status circuit 910 and a cache bottom register 920, which can be a programmable up/down counter. Stack cache status circuit 910, receives pointer CACHE_BOTTOM from cache bottom register 920 and pointer OPTOP to determine the number of free registers FREE and the number of used registers USED.
  • For a circular buffer using sequential modulo-SCS addressing, as in FIG. 7, the number of free registers FREE is defined as
      • FREE=SCS−(OPTOP-CACHE_BOTTOM+1) MOD SCS, where SCS is the size of stack cache 155. Thus, for the specific pointer values shown in FIG. 7, the number of free registers FREE is 34, as calculated by:
        FREE=64−((27−62+1)MOD 64)=34.
  • Similarly, for a circular buffer using sequential modulo addressing, the number of used registers USED defined as
    USED=(OPTOP−CACHE_BOTTOM+1)MOD SCS.
  • Thus, for the specific pointer values shown in FIG. 7, the number of used registers USED is 30, as calculated by:
    USED=(27−62+1)MOD 64.
  • Thus, stack cache status circuit 910 can be implemented with a modulo SCS adder/subtractor. The number of used registers USED and the number of free registers FREE can also be generated using a programmable up/down counters. For example, a used register can be incremented whenever a data word is added to stack cache 155 and decremented whenever a data word is removed from stack cache 155. Specifically, if pointer OPTOP is modulo-SCS incremented by some amount, the used register is incremented by the same amount. If pointer OPTOP is modulo-SCS decremented by some amount, the used register is decremented by the same amount. However, if pointer CACHE_BOTTOM is modulo-SCS incremented by some amount, the used register is decremented by the same amount. If pointer CACHE_BOTTOM is modulo-SCS. decremented by some amount, the used register is incremented the same amount. The number of free registers FREE can be generated by subtracting the number of used registers USED from the total number of registers.
  • Spill control unit 694 (FIGS. 6 and 9) includes a cache high threshold register 930 and a comparator 940. Comparator 940 compares the value in cache high threshold register 930 to the number of free registers FREE. If the number of free registers FREE is less than the value in cache high threshold register 930, comparator 940 drives a spill signal SPILL to a spill logic level, typically logic high, to indicate that the spill condition exists and one or more data words should be transferred from stack cache memory circuit 610 to stack 400, i.e. a spill operation should be performed. The spill operation is described in more detail below. Typically, cache high threshold register 930 is programmable by hardware processor 100.
  • Fill control unit 698 (FIGS. 6 and 9) includes a cache low threshold register 950 and a comparator 960. Comparator 960 compares the value in cache low threshold register 950 to the number of used registers USED. If the number of used registers is less than the value in cache low threshold register 950, comparator 960 drives a fill signal FILL to a fill logic level, typically logic high, to indicate that the fill condition exists and one or more data words should be transferred from stack 400 to stack cache memory circuit 610, i.e. a fill operation should be performed. The fill operation is described in more detail below. Typically, cache low threshold register 950 is programmable by hardware processor 100.
  • If the value in cache high threshold 930 and cache low threshold 940 is always the same, a single cache threshold register can be used. Fill control unit 698 can be modified to use the number of free registers FREE to drive signal FILL to the fill logic level if then number of free registers is greater than the value in cache low threshold 950, with a proper modification of the value in cache low threshold 950. Alternatively, spill control unit 694 can be modified to use the number of used registers.
  • FIG. 10A shows another embodiment of dribble manager unit 151, which uses a high-water mark/low water mark heuristic to determine when a spill condition or a fill condition exists. Spill control unit 694 includes a high water mark register 1010 implemented as a programmable up/down counter. A comparator 1020 in spill control unit 694 compares the value in high water mark register 1010, i.e. the high water mark, with pointer OPTOP. If pointer OPTOP is greater than the high water mark, comparator 1020 drives spill signal SPILL to the spill logic level to indicate a spill operation should be performed. Since, the high water mark is relative to pointer CACHE_BOTTOM, the high water mark is modulo-SCS incremented and modulo-SCS decremented whenever pointer CACHE_BOTTOM is modulo-SCS incremented or modulo-SCS decremented, respectively.
  • Fill control unit 698 includes a low water mark register 1010 implemented as a programmable up/down counter. A comparator 1030 in fill control unit 698 compares the value in low water mark register 1030, i.e. the low water mark, with pointer OPTOP. If pointer OPTOP is less than the low water mark, comparator 1040 drives fill signal FILL to the fill logic level to indicate a fill operation should be performed. Since the low water mark is relative to pointer CACHE_BOTTOM, the low water mark register is modulo-SCS incremented and modulo-SCS decremented whenever pointer CACHE_BOTTOM is modulo-SCS incremented or modulo-SCS decremented, respectively.
  • FIG. 10B shows an alternative circuit to generate the high water mark and low water mark. Cache high threshold register 930, typically implemented as a programmable register, contains the number of free registers which should be maintained in stack cache memory circuit 610. The high water mark is then calculated by modulo-SCS subtractor 1050 by modulo-SCS subtracting the value in cache high threshold register 930 from pointer CACHE_BOTTOM stored in cache bottom register 920.
  • The low water mark is calculated by doing a modulo-SCS addition. Specifically, cache low threshold register 950 is programmed to contain the minimum number of used data registers desired to be maintained in stack cache memory circuit 610. The low water mark is then calculated by modulo-SCS adder 1060 by modulo-SCS adding the value in cache low threshold register 950 with pointer CACHE_BOTTOM stored in cache bottom register 920.
  • As described above, a spill operation is the transfer of one or more data words from stack cache memory circuit 610 to stack 400. In the embodiment of FIG. 1, the transfers occurs though data cache unit 160. The specific interface between stack management unit 150 and data cache unit 160 can vary. Typically, stack management unit 150, and more specifically dribble manager unit 151, sends the data word located at the bottom of stack cache 155, as indicated by pointer CACHE_BOTTOM from read port 680 to data cache unit 160. The value of pointer CACHE_BOTTOM is also provided to data cache unit 160 so that data cache unit 160 can address the data word appropriately. The saved bit of the register indicated by pointer CACHE_BOTTOM is set to the saved logic level. In addition, pointer CACHE_BOTTOM is modulo-SCS incremented by one. Other registers as described above may also be modulo-SCS incremented by one. For example, high water mark register 1010 (FIG. 10A) and low water mark 1030 would be modulo-SCS incremented by one. Some embodiments of dribble manager unit 157. transfer multiple words for each spill operation. For these embodiments, pointer CACHE_BOTTOM is modulo-SCS incremented by the number words transferred to stack 400.
  • In embodiments using a saved bit and valid bit, as shown in FIG. 8, some optimization is possible. Specifically, if the saved bit of the data register pointed to by pointer CACHE_BOTTOM is at the saved logic level, the data word in that data register is already stored in stack 400. Therefore, the data word in that data register does not need to be copied to stack 400. However, pointer CACHE_BOTTOM is still modulo-SCS incremented by one.
  • A fill operation transfers data words from stack 400 to stack cache memory circuit 610. In the embodiment of FIG. 1, the transfers occurs though data cache unit 160. The specific interface between stack management unit 150 and data cache unit 160 can vary. Typically, stack management unit 150, and more specifically dribble manager unit 151, determines whether the data register preceding the data register pointed by CACHE_BOTTOM is free, i.e. either the saved bit is in the saved logic state or the valid bit is in the invalid logic state. If the data register preceding the data register pointed to by pointer CACHE_BOTTOM is free, dribble manager unit 151 requests a data word from stack 400 by sending a request with the value of pointer CACHE_BOTTOM modulo-SCS minus one. When the data word is received from data cache unit 160, pointer CACHE_BOTTOM is modulo-SCS decremented by one and the received data word is written to the data register pointed to by pointer CACHE_BOTTOM through write port 670. Other registers as described above may also be modulo-SCS decremented. The saved bit and valid bit of the register pointed to by pointer CACHE_BOTTOM are set to the saved logic state and valid logic state, respectively. Some embodiments of dribble manager unit 151 transfer multiple words for each spill operation. For these embodiments, pointer CACHE_BOTTOM is modulo-SCS decremented by the number words transferred to stack 400.
  • In embodiments using a saved bit and valid bit, as shown in FIG. 8, some optimization is possible. Specifically, if the saved bit and valid bit of the data register preceding the data register pointed to by pointer CACHE_BOTTOM is at the saved logic level and the valid logic level, respectively, then the data word in that data register was never overwritten. Therefore, the data word in that data register does not need to be copied from stack 400. However, pointer CACHE_BOTTOM is still modulo-SCS decremented by one.
  • As stated above, in one embodiment of stack cache 155, hardware processor 100 accesses stack cache memory circuit 610 (FIG. 6) through write port 630, read port 640 and read port 650. Stack control unit 152 generates pointers for write port 630, read port 640, and read port 650 based on the requests of hardware processor 100. FIG. 11 shows a circuit to generate pointers for a typical operation which reads two data words from stack cache 155 and writes one data word to stack cache 155. The most common stack manipulation for a stack-based computing system is to pop the top two data words off of the stack and to push a data word onto the top of the stack. Therefore, the circuit of FIG. 11 is configured to be able to provide read pointers to the value of pointer OPTOP and the value of pointer OPTOP modulo-SCS minus one, and a write pointer to the current value of OPTOP modulo-SCS minus one.
  • Multiplexer (MUX) 1110 drives a read pointer RP1 for read port 640. A select line RS1 controlled by hardware processor 100 determines whether multiplexer 1110 drives the same value as pointer OPTOP or a read address R_ADDR1 as provided by hardware processor 100.
  • Multiplexer 1120 provides a read pointer RP2 for read port 650. Modulo adder 1140 modulo-SCS adds negative one to the value of pointer OPTOP and drives the resulting sum to multiplexer 1120. A select line RS2 controlled by hardware processor 100 determines whether multiplexer 1120 drives the value from modulo adder 1140 or a read address R_ADDR2 as provided by hardware processor 100.
  • Multiplexer 1130 provides a write pointer WP for write port 630. A modulo adder 1150 modulo-SCS adds one to the value of pointer OPTOP and drives the resulting sum to multiplexer 1130. Select lines WS controlled by hardware processor 100 determines whether multiplexer 1130 drives the value from modulo-SCS adder 1140, the value from modulo-SCS adder 1150, or a write address W_ADDR as provided by hardware processor 100.
  • FIG. 12 shows a circuit that generates a read pointer R for read port 640 or read port 650 in embodiments allowing accessing stack cache memory circuit using pointer VARS. Multiplexer 1260 drives read pointer R to one of several input values received on input ports 1261-1267 as determined by selection signals RS. Selection signals RS are controlled by hardware processor 100. The value of pointer OPTOP is driven to input port 1261. Modulo-SCS adder 1210 drives the modulo-SCS sum of the value of pointer OPTOP with negative one to input port 1262. Modulo-SCS adder 1210 drives the modulo-SCS sum of the value of pointer OPTOP with negative two to input port 1263. The value of pointer VARS is driven to input port 1264. Modulo-SCS adder 1230 drives the modulo-SCS sum of the value of pointer VARS with one to input port 1265. Modulo-SCS adder 1240 drives the modulo-SCS sum of the value of pointer VARS with two to input port 1266. Modulo adder-SCS 1250 drives the modulo-SCS sum of the value of pointer VARS with three to input port 1263. Other embodiments may provide other values to the input ports of multiplexer 1260.
  • Thus by using the stack cache according to the principles of the invention, a dribbling management unit can efficiently control transfers between the stack cache and the stack. Specifically, the dribbling management unit is able to transfer data out of the stack cache to make room for additional data as necessary and transfer data into the stack cache as room becomes available transparently to the stack-based computing system using the stack management unit.
  • The various embodiments of the structure and method of this invention that are described above are illustrative only of the principles of this invention and are not intended to limit the scope of the invention to the particular embodiments described. In view of this disclosure, those skilled-in-the-art can define other memory circuits, registers, counters, stack-based computing systems, dribble management units, fill control units, spill control units, read ports, write ports, and use these alternative features to create a method or system of stack caching according to the principles of this invention.
  • SECTION I
  • The JAVA Virtual Machine Specification
    • ©1993, 1994, 1995 Sun Microsystems, Inc.
    • 2550 Garcia Avenue, Mountain View, Calif.
    • 94043-1100 U.S.A.
  • All rights reserved. This BETA quality release and related documentation are protected by copyright and distributed under licenses restricting its use, copying, distribution, and decompilation. No part of this release or related documentation may be reproduced in any form by any means without prior written authorization of Sun and its licensors, if any.
  • Portions of this product may be derived from the UNIX® and Berkeley 4.3 BSD systems, licensed from UNIX System Laboratories, Inc. and the University of California, respectively. Third-party font software in this release is protected by copyright and licensed from Sun's Font Suppliers.
  • RESTRICTED RIGHTS LEGEND: Use, duplication, or disclosure by the United States Government is subject to the restrictions set forth in DFARS 252.227-7013 (c) (1) (ii) and FAR 52.227-19.
  • The release described in this manual may be protected by one or more U.S. patents, foreign patents, or pending applications.
  • Trademarks
  • Sun, Sun Microsystems, Sun Microsystems Computer Corporation, the Sun logo, the Sun Microsystems Computer Corporation logo, WebRunner, JAVA, FirstPerson and the FirstPerson logo and agent are trademarks or registered trademarks of Sun Microsystems, Inc. The “Duke” character is a trademark of Sun Microsystems, Inc. and Copyright (c) 1992-1995 Sun Microsystems, Inc. All Rights Reserved. UNIX® is a registered trademark in the United States and other countries, exclusively licensed through X/Open Company, Ltd. OPEN LOOK is a registered trademark of Novell, Inc. All other product names mentioned herein are the trademarks of their respective owners.
  • All SPARC trademarks, including the SCD Compliant Logo, are trademarks or registered trademarks of SPARC International, Inc. SPARCstation, SPARCserver, SPARCengine, SPARCworks, and SPARCompiler are licensed exclusively to Sun Microsystems, Inc; Products bearing SPARC trademarks are based upon an architecture developed by Sun Microsystems, Inc.
  • The OPEN LOOK® and Sun™ Graphical User Interfaces were developed by Sun Microsystems, Inc. for its users and licensees. Sun acknowledges the pioneering efforts of Xerox in researching and developing the concept of visual or graphical user interfaces for the computer industry. Sun holds a non-exclusive license from Xerox to the Xerox Graphical User Interface, which license also covers Sun's licensees who implement OPEN LOOK GUIs and otherwise comply with Sun's written license agreements.
  • X Window System is a trademark and product of the Massachusetts Institute of Technology.
  • THIS PUBLICATION IS PROVIDED “AS IS” WITHOUT WARRANTY OF ANY KIND, EITHER EXPRESS OR IMPLIED, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE, OR NON-INFRINGEMENT.
  • THIS PUBLICATION COULD INCLUDE TECHNICAL INACCURACIES OR TYPOGRAPHICAL ERRORS. CHANGES ARE PERIODICALLY ADDED TO THE INFORMATION HEREIN; THESE CHANGES WILL BE INCORPORATED IN NEW EDITIONS OF THE PUBLICATION. SUN MICROSYSTEMS, INC. MAY MAKE IMPROVEMENTS AND/OR CHANGES IN THE PRODUCT(S) AND/OR THE PROGRAM(S) DESCRIBED IN THIS PUBLICATION AT ANY TIME.
  • Preface
  • This document describes version 1.0 of the JAVA Virtual Machine and its instruction set. We have written this document to act as a specification for both compiler writers, who wish to target the machine, and as a specification for others who may wish to implement a compliant JAVA Virtual Machine.
  • The JAVA Virtual Machine is an imaginary machine that is implemented by emulating it in software on a real machine. Code for the JAVA Virtual Machine is stored in class files, each of which contains the code for at most one public class.
  • Simple and efficient emulations of the JAVA Virtual Machine are possible because the machine's format is compact and efficient bytecodes. Implementations whose native code speed approximates that of compiled C are also possible, by translating the bytecodes to machine code, although Sun has not released such implementations at this time.
  • The rest of this document is structured as follows:
      • Chapter 1 describes the architecture of the JAVA Virtual Machine;
      • Chapter 2 describes the class file format;
      • Chapter 3 describes the bytecodes; and
      • Appendix A contains some instructions generatedinternally by Sun's implementation of the JAVA Virtual Machine. While not strictly part of the specification we describe these here so that this specification can serve as a reference for our implementation. As more implementations of the JAVA Virtual Machine become available, we may remove Appendix A from future releases.
      • Sun will license the JAVA Virtual Machine trademark and logo for use with compliant implementations of this specification. If you are considering constructing your own implementation of the JAVA Virtual Machine please contact us, at the email address below, 60 that we can work together to insure 100% compatibility of your implementation.
  • Send comments on this specification or questions about implementing the JAVA Virtual Machine to our electronic mail address:JAVA@JAVA.sun.com.
  • 1. JAVA Virtual Machine Architecture
  • 1.1 Supported Data Types
  • The virtual machine data types include the basic data types of the JAVA language:
  • byte //1-byte signed 2's complement integer
  • short //2-byte signed 2's complement integer
  • int //4-byte signed 2's complement integer
  • long //8-byte signed 2's complement integer
  • float //4-byte IEEE 754 single-precision float
  • double //8-byte IEEE 754 double-precision float
  • char //2-byte unsigned Unicode character
  • Nearly all JAVA type checking is done at compile time. Data of the primitive types shown above need not be tagged by the hardware to allow execution of JAVA. Instead, the bytecodes that operate on primitive values indicate the types of the operands so that, for example, the iadd, ladd, fadd, and dadd instructions each add two numbers, whose types are int, long, float, and double, respectively
  • The virtual machine doesn't have separate instructions for boolean types. Instead, integer instructions, including integer returns, are used to operate on boolean values; byte arrays are used for arrays of boolean.
  • The virtual machine specifies that floating point be done in IEEE 754 format, with support for gradual underflow. Older computer architectures that do not have support for IEEE format may run JAVA numeric programs very slowly.
  • Other virtual machine data types include:
      • object //4-byte reference to a JAVA object
      • returnAddress //4 bytes, used with jsr/ret/jsr_w/ret_w instructions
        Note: JAVA arrays are treated as objects.
  • This specification does not require any particular internal structure for objects. In our implementation an object reference is to a handle, which is a pair of pointers: one to a method table for the object, and the other to the data allocated for the object. Other implementations may use inline caching, rather than method table dispatch such methods are likely to be faster on hardware that is emerging between now and the year 2000.
  • Programs represented by JAVA Virtual Machine bytecodes are expected to maintain proper type discipline and an implementation may refuse to execute a bytecode program that appears to violate such type discipline.
  • While the JAVA Virtual Machines would appear to be limited by the bytecode deònition to running on a 32-bit address space machine, it is possible to build a version of the JAVA Virtual Machine that automatically translates the bytecodes into a 64-bit form. A description of this transformation is beyond the scope of the JAVA Virtual Machine Specification.
  • 1.2 Registers
  • At any point the virtual machine is executing the code of a single method, and the pc register contains the address of the next bytecode to be executed.
  • Each method has memory space allocated for it to hold:
      • a set of local variables, referenced by a vars register;
      • an operand stack, referenced by an optop register; and
      • a execution environment structure, referenced by a frame register.
  • All of this space can be allocated at once, since the size of the local variables and operand stack are known at compile time, and the size of the execution environment structure is well-known to the interpreter.
  • All of these registers are 32 bits wide.
  • 1.3 Local Variables
  • Each JAVA method uses a fixed-sized set of local variables. They are addressed as word offsets from the vary register. Local variables are all 32 bits wide.
  • Long integers and double precision floats are considered to take up two local variables but are addressed by the index of the first local variable. (For example, a local variable with index containing a double precision float actually occupies storage at indices n and n+1.) The virtual machine specifcation does not require 64-bit values in local variables to be 64-bit aligned. Implementor are free to decide the appropriate way to divide long integers and double precision floats into two words.
  • Instructions are provided to load the values of local variables onto the operand stack and store values from the operand stack into local variables.
  • 1.4 The Operand Stack
  • The machine instructions all take operands from an operand stack, operate on them, and return results to the stack. We chose a stack organization so that it would be easy to emulate the machine efficiently on machines with few or irregular registers such as the Intel 486 microprocessor.
  • The operand stack is 32 bits wide. It is used to pass parameters to methods and receive method results, as well as to supply parameters for operations and save operation results.
  • For example, execution of instruction iadd adds two integers together. It expects that the two integers are the top two words on the operand stack, and were pushed there by previous instructions. Both integers are popped from the stack, added, and their sum pushed back onto the operand stack. Subcomputations may be nested on the operand stack, and result in a single operand that can be used by the is nesting computation.
  • Each primitive data type has specialized instructions that know how to operate on operands of that type. Each operand requires a single location on the stack, except for long and double operands, which require two locations.
  • Operands must be operated on by operators appropriate to their type. It is illegal, for example, to push two integers and then treat them as a long. This restriction is enforced, in the Sun implementation, by the bytecode verifier. However, a small number of operations (the dup opcodes and swap) operate on runtime data areas as raw values of a given width without regard to type.
  • In our description of the virtual machine instructions below, the effect of an instruction's execution on the operand stack is represented textually, with the stack growing from left to right, and each 32-bit word separately represented. Thus:
      • Stack: . . . , value1, value2
        Figure US20070277021A1-20071129-P00001
        . . . , value3
        shows an operation that begins by having value2 on top of the stack with value1 just beneath it. As a result of the execution of the instruction, value1 and value2 are popped from the stack and replaced by value3, which has been calculated by the instruction. The remainder of the stack, represented by an ellipsis, is unaffected by the instruction's execution.
  • The types long and double take two 32-bit words on the operand stack:
      • Stack: . . .
        Figure US20070277021A1-20071129-P00001
        . . . , value-word1, value-word2
  • This specification does not say how the two words are selected from the 64-bit long or double value; it is only necessary that a particular implementation be internally consistent.
  • 1.5 Execution Environment
  • The information contained in the execution environment is used to do dynamic linking, normal method returns, and exception propagation.
  • 1.5.1 Dynamic Linking
  • The execution environment contains references to the interpreter symbol table for the current method and current class, in support of dynamic linking of the method code. The class file code for a method refers to methods to-be called and variables to be accessed symbolically. Dynamic linking translates these symbolic method calls into actual method calls, loading classes as necessary to resolve as-yet-undefined symbols, and translates variable accesses into appropriate offsets in storage structures associated with the runtime location of these variables.
  • This late binding of the methods and variables makes changes in other classes that a method uses less likely to break this code.
  • 1.5.2 Normal Method Returns
  • If execution of the current method completes normally, then a value is returned to the calling method. This occurs when the calling method executes a return instruction appropriate to the return type.
  • The execution environment is used in this case to restore the registers of the caller, with the program counter of the caller appropriately incremented to skip the method call instruction. Execution then continues in the calling method's execution environment.
  • 1.5.3 Exception and Error Propagation
  • An exceptional condition, known in JAVA as an Error or Exception, which are subclasses of Throwable, may arise in a program because of:
      • a dynamic linkage failure, such as a failure to find a needed class file;
      • a run-time error, such as a reference through a null pointer;
      • an asynchronous event, such as is thrown by Thread.stop, from another thread; and
      • the program using a throw statement.
        When an exception occurs:
      • A list of catch clauses associated with the current method is examined. Each catch clause describes the instruction range for which it is active, describes the type of exception that it is to handle, and has the address of the code to handle it.
      • An exception matches a catch clause if the instruction that caused the exception is in the appropriate instruction range, and the exception type is a subtype of the type of exception that the catch clause handles. If a matching catch clause is found, the system branches to the specified handler. If no handler is found, the process is repeated until all the nested catch clauses of the current method have been exhausted.
      • The order of the catch clauses in the list is important. The virtual machine execution continues at the first matching catch clause. Because JAVA code is structured, it is always possible to sort all the exception handlers for one method into a single list that, for any possible program counter value, can be searched in linear order to find the proper (innermost containing applicable) exception handler for an exception occurring at that program counter value.
      • If there is no matching catch clause then the current method is said to have as its outcome the uncaught exception. The execution state of the method that called this method is restored from the execution environment, and the propagation of the exception continues, as though the exception had just occurred in this caller.
        1.5.4 Additional Information
  • The execution environment may be extended with additional implementation-specified information, such as debugging information.
  • 1.6 Garbage Collected Heap
  • The JAVA heap is the runtime data area from which class instances (objects) are allocated. The JAVA language is designed to be garbage collected—it does not, give the programmer the ability to deallocate objects explicitly. The JAVA language does not presuppose any particular kind of garbage collection; various algorithms may be used depending on system requirements.
  • 1.7 Method Area
  • The method area is analogous to the store for compiled code in conventional languages or the text segment in a UNIX process. It stores method code (compiled JAVA code) and symbol tables. In the current JAVA implementation, method code is not part of the garbage-collected heap, although this is planned for a future release.
  • 1.8 The JAVA Instruction Set
  • An instruction in the JAVA instruction set consists of a one-byte opcode specifying the operation to be performed, and zero or more operands supplying parameters or data that will be used by the operation. Many instructions have no operands and consist only of an opcode.
  • The inner loop of the virtual machine execution is effectively:
    do {
    fetch an opcode byte
    execute an action depending on the value of
    the opcode
    } while (there is more to do);
  • The number and size of the additional operands is determined by the opcode. If an additional operand is more than one byte in size, then it is stored in big-endian order—high order byte first. For example, a 16-bit parameter is stored as two bytes whose value is:
    first_byte*256+second_byte
  • The bytecode instruction stream is only byte-aligned, with the exception being the tableswitch and lookupswitch instructions, which force alignment to a 4-byte boundary within their instructions.
  • These decisions keep the virtual machine code for a compiled JAVA program compact and reflect a conscious bias in favor of compactness at some possible cost in performance.
  • 1.9 Limitations
  • The per-class constant pool has a maximum of 65535 entries. This acts as an internal limit on the total complexity of a single class.
  • The amount of code per method is limited to 65535 bytes by the sizes of the indices in the code in the exception table, the line number table, and the local variable table.
  • Besides this limit, the only other limitation of note is that the number of words of arguments in a method call is limited to 255.
  • 2. Class File Format
  • This chapter documents the JAVA class (.class) file format.
  • Each class file contains the compiled version of either a JAVA class or a JAVA interface. Compliant JAVA interpreters must be capable of dealing with all class files that conform to the following specification.
  • A JAVA class file consists of a stream of 8-bit bytes. All 16-bit and 32-bit quantities are constructed by reading in two or four 8-bit bytes, respectively. The bytes are joined together in network (big-endian) order, where the high bytes come first. This format is supported by the JAVA JAVA.io.DataInput and JAVA.io.DataOutput interfaces, and classes such as JAVA.io.DataInputStream and JAVA.io.DataOutputStream.
  • The class file format is described here using a structure notation. Successive fields in the structure appear in the external representation without padding or alignment. Variable size arrays, often of variable sized elements, are called tables and are commonplace in these structures.
  • The types u1, u2, and u4 mean an unsigned one-, two-, or four-byte quantity, respectively, which are read by method such as readUnsignedByte, readUnsignedShort and readInt of the JAVA.io.DataInput interface.
  • 2.1 Format
  • The following pseudo-structure gives a top-level description of the format of a class file:
    ClassFile {
    u4 magic;
    u2 minor_version;
    u2 major_version;
    u2 constant_pool_count;
    cp_info constant_pool[constant_pool_count − 1];
    u2 access_flags;
    u2 this_class;
    u2 super_class;
    u2 interfaces_count;
    u2 interfaces[interfaces_count];
    u2 fields_count;
    field_info fields[fields_count];
    u2 methods_count;
    method_info methods[methods_count];
    u2 attributes_count;
    attribute_info attributes[attribute_count];
    }

    magic
  • This field must have the value 0×CAFEBABE.
  • minor_version, major_version
  • These fields contain the version number of the. JAVA compiler that produced this class file. An implementation of the virtual machine will normally support some range of minor version numbers 0-n of a particular major version number. If the minor version number is incremented the new code won't run on the old virtual machines, but it is possible to make a new virtual machine which can run versions up to n+1.
  • A change of the major version number indicates a major incompatible change, one that requires a different virtual machine that may not support the old major version in any way.
  • The current major version number is 45; the current minor version number is 3.
  • constant_Pool_count
  • This field indicates the number of entries in the constant pool in the class file.
  • constant_pool
  • The constant pool is a table of values. These values are the various string constants, class names, field names, and others that are referred to by the class structure or by the code.
  • constant_pool[0] is always unused by the compiler, and may be used by an implementation for any purpose.
  • Each of the constant_pool entries 1 through constant_pool_count−1 is a variable-length entry, whose format is given by the first “tag” byte, as described in section 2.3.
  • access_flags
  • This field contains a mask of up to sixteen modifiers used with class, method, and field declarations. The same encoding is used on similar fields in field_info and method_info as described below. Here is the encoding:
    Flag Name Value Meaning Used By
    ACC_PUBLIC 0x0001 Visible to everyone Class,
    Method,
    Variable
    ACC_PRIVATE 0x0002 Visible only to the Method,
    defining class Variable
    ACC_PROTECTED 0x0004 Visible to Method,
    subclasses Variable
    ACC_STATIC 0x0008 Variable or method Method,
    is static Variable
    ACC_FINAL 0x0010 No further sub- Class,
    classing, over- Method,
    riding, or assign- Variable
    ment after ini-
    tialization
    ACC_SYNCHRONIZED 0x0020 Wrap use in monitor Method
    lock
    ACC_VOLATILE 0x0040 Can't cache Variable
    ACC_TRANSIENT 0x0080 Not to be written Variable
    or read by a per-
    sistent object
    manager
    ACC_NATIVE 0x0100 Implemented in a Method
    language other
    than JAVA
    ACC_INTERFACE 0x0200 Is an interface Class
    ACC_ABSTRACT 0x0400 No body provided Class,
    Method

    this_class
  • This field is an index into the constant pool; constant_pool [this_class] must be a CONSTANT_class.
  • super_class
  • This field is an index into the constant pool. If the value of super_class is nonzero, then constant_pool [super_class] must be a class, and gives the index of this class's superclass in the constant pool.
  • If the value of super_class is zero, then the class being defined must be JAVA.lang.Object, and it has no superclass.
  • interfaces_count
  • This field gives the number of interfaces that this class implements.
  • interfaces
  • Each value in this table is an index into the constant pool. If a table value is nonzero (interfaces[i]|=0, where 0<=i <interfaces_count), then constant_pool [interfaces[i]] must be an interface that this class implements.
  • fields_count
  • This field gives the number of instance variables, both static and dynamic, defined by this class. The fields table includes only those variables that are defined explicitly by this class. It does not include those instance variables that are accessible from this class but are inherited from superclasses.
  • fields
  • Each value in this table is a more complete description of a field in the class. See section 2.4 for more information on the field_info structures.
  • methods_count
  • This field indicates the number of methods, both static and dynamic, defined by this class. This table only includes those methods that are explicitly defined by this class. It does not include inherited methods.
  • methods
  • Each value in this table is a more complete description of a method in the class. See section 2.5 for more information on the method_info structure.
  • attributes_count
  • This field indicates the number of additional attributes about this class.
  • attributes
  • A class can have any number of optional attributes associated with it. Currently, the only class attribute recognized is the “SourceFile” attribute, which indicates the name of the source file from which this class file was compiled. See section 2.6 for more information on the attribute_info structure.
  • 2.2 Signatures
  • A signature is a string representing a type of a method, field or array.
  • The field signature represents the value of an argument to a function or the value of a variable. It is a series of bytes generated by the following grammar:
    <field_signature> ::= <field_type>
    <field_type> ::=<base_type>|<object_type>|
    <array_type>
    <base_type> ::= B|C|D|F|I|J|S|Z
    <object_type> ::= L<fullclassname>;
    <array_type> ::=[<optional_size><field_type>
    <optional_size> ::= [0-9]
  • The meaning of the base types is as follows:
    B byte signed byte
    C char character
    D double double
    precision IEEE
    float
    F float single
    precision IEEE
    float
    I int integer
    J long long integer
    L<fullclassname>; . . . an object of
    the given
    class
    S short signed short
    Z boolean true or false
    [<field sig> . . . array
  • A return-type signature represents the return value from a method. It is a series of bytes in the following grammar:
    <return_signature>::=<field_type>|V
  • The character V indicates that the method returns no value. Otherwise, the signature indicates the type of the return value.
  • An argument signature represents an argument passed to a method:
    <argument_signature>::=<field_type>
  • A method signature represents the arguments that the method expects, and the value that it returns.
    <method_signature> ::= (<arguments_signature>)
    <return_signature>
    <arguments_signature> ::= <argument_signature>*

    2.3 Constant Pool
  • Each item in the constant pool begins with a 1-byte tag:. The table below lists the valid tags and their values.
    Constant Type Value
    CONSTANT_Class
    7
    CONSTANT_Fieldref 9
    CONSTANT_Methodref 10
    CONSTANT_InterfaceMethodref 11
    CONSTANT_String 8
    CONSTANT_Integer 3
    CONSTANT_Float 4
    CONSTANT_Long 5
    CONSTANT_Double 6
    CONSTANT_NameAndType 12
    CONSTANT_Utf8 1
    CONSTANT_Unicode 2
  • Each tag byte is then followed by one or more bytes giving more information about the specific constant.
  • 2.3.1 CONSTANT_Class
  • CONSTANT_Class is used to represent a class or an interface.
    CONSTANT_Class_info {
    u1 tag;
    u2 name_index;
    }

    tag
      • The tag will have the value CONSTANT_Class name_index
        • constant_pool[name_index] is a CONSTANT_Utf8 giving the string name of the class.
  • Because arrays are objects, the opcodes anewarray and multianewarray can reference array “classes” via CONSTANT_Class items in the constant pool. In this case, the name of the class is its signature. For example, the class name for
      • int [ ] [ ]
        is
      • [[I
        The class name for
      • Thread[ ]
        is
      • “[LJAVA.lang.Thread;”
        2.3.2 CONSTANT_{Fieldref,Methodref, InterfaceMethodref}
  • Fields, methods, and interface methods are represented by similar structures.
    CONSTANT_Fieldref_info {
    u1 tag;
    u2 class_index;
    u2 name_and_type_index;
    }
    CONSTANT_Methodref_info {
    u1 tag;
    u2 class_index;
    u2 name_and_type_index;
    }
    CONSTANT_InterfaceMethodref_info {
    u1 tag;
    u2 class_index;
    u2 name_and_type_index;
    }

    tag
  • The tag will have the value CONSTANT_Fieldref, CONSTANT_Methodref, or CONSTANT_InterfaceMethodref.
  • class_index
  • constant_pool [class_index] will be an entry of type CONSTANT_Class giving the name of the class or interface containing the field or method.
  • For CONSTANT_Fieldref and CONSTANT_Methodref, the CONSTANT_Class item must be an actual class. For CONSTANT_InterfaceMethodref, the item must be an interface which purports to implement the given method.
  • name_and_type_index
  • constant_pool [name_and_type_index] will be an entry of type CONSTANT_NameAndType. This constant pool entry indicates the name and signature of the field or method.
  • 2.3.3 CONSTANT_String
  • CONSTANT_String is used to represent constant objects of the built-in type String.
    CONSTANT_String_info {
    u1 tag;
    u2 string_index;
    }

    tag
  • The tag will have the value CONSTANT_String
  • string_index
      • constant_pool [string_index] is a CONSTANT_Utf8 string giving the value to which the String object is initialized.
        2.3.4 CONSTANT_Integer andCONSTANT_Float
  • CONSTANT_Integer andCONSTANT_Float represent four-byte constants.
    CONSTANT_Integer_info {
    u1 tag;
    u4 bytes;
    }
    CONSTANT_Float_info {
    u1 tag;
    u4 bytes;
    }

    tag
      • The tag will have the value CONSTANT_Integer or CONSTANT_Float
        bytes
  • For integers, the four bytes are the integer value. For floats, they are the IEEE 754 standard representation of the floating point value. These bytes are in network (high byte first) order.
  • 2.3.5 CONSTANT_Long and CONSTANT_Double
  • CONSTANT_Long andCONSTANT_Double represent eight-byte constants.
    CONSTANT_Long_info {
    u1 tag;
    u4 high_bytes;
    u4 low_bytes;
    }
    CONSTANT_Double_info {
    u1 tag;
    u4 high_bytes;
    u4 low_bytes;
    }
  • All eight-byte constants take up two spots in the constant pool. If this is the nth item in the constant pool, then the next item will be numbered n+2.
  • tag
  • The tag will have the value CONSTANT_Long or CONSTANT_Double.
  • high_bytes, low_bytes
  • For CONSTANT_Long, the 64-bit value is (high_bytes<<32)+low_bytes.
  • For CONSTANT_Double, the 64-bit value;high_bytes and low_bytes together represent the standard IEEE 754 representation of the double-precision floating point number.
  • 2.3.6 CONSTANT_NameAndType
  • CONSTANT_NameAndType is used to represent a field or method, without indicating which class it belongs to.
    CONSTANT_NameAndType_info {
    u1 tag;
    u2 name_index;
    u2 signature_index;
    }

    tag
  • The tag will have the valueCONSTANT_NameAndType.
  • name_index
  • constant_pool [name_index] is a CONSTANT_Utf8 string giving the name of the field or method.
  • signature_index
  • constant_pool [signature_index] is a CONSTANT_Utf8 string giving the signature of the field or method.
  • 2.3.7 CONSTANT_Utf8 and CONSTANT_Unicode
  • CONSTANT_Utf8 andCONSTANT_Unicode are used to represent constant string values.
  • CONSTANT_Utf8 strings are “encoded” so that strings containing only non-null ASCII characters, can be represented using only one byte per character, but characters of up to 16 bits can be represented:
  • All characters in the range 0×0001 to 0×007F are represented by a single byte:
    Figure US20070277021A1-20071129-C00001
  • The null character (0×0000) and characters in the range 0×0080 to 0×07FF are represented by a pair of two bytes:
    Figure US20070277021A1-20071129-C00002
  • Characters in the range 0×0800 to 0×FFFF are represented by three bytes:
    Figure US20070277021A1-20071129-C00003
  • There are two differences between this format and the “standard” UTF-8 format. First, the null byte (0×00) is encoded in two-byte format rather than one-byte, so that our strings never have embedded nulls. Second, only the one-byte, two-byte, and three-byte formats are used. We do not recognize the longer formats.
    CONSTANT_Utf8_info {
    u1 tag;
    u2 length;
    u1 bytes[length];
    }
    CONSTANT_Unicode_info {
    u1 tag;
    u2 length;
    u2 bytes [length];
    }

    tag
      • The tag will have the value CONSTANT_Utf8 or CONSTANT_Unicode.
        length
      • The number of bytes in the string. These strings are not null terminated.
        bytes
      • The actual bytes of the string.
  • 2.4 Fields
  • The information for each field immediately follows the field_count field in the class file. Each field is described by a variable length field_info structure. The format of this structure is as follows:
    field_info {
    u2 access_flags;
    u2 name_index;
    u2 signature_index;
    u2 attributes_count;
    attribute_info attributes[attribute_count];
    }
    access_flags
  • This is a set of sixteen flags used by classes, methods, and fields to describe various properties and how they many be accessed by methods in other classes. See the table “Access Flags” which indicates the meaning of the bits in this field.
  • The possible fields that can be set for a field are ACC_PUBLIC, ACC_PRIVATE, ACC_PROTECTED, ACC_STATIC, ACC_FINAL, ACC_VOLATILE, and ACC_TRANSIENT.
  • At most one of ACC_PUBLIC, ACC_PROTECTED, and ACC_PRIVATE can be set for any method.
  • name_index
  • constant_pool [name_index] is a CONSTANT_Utf8 string which is the name of the field.
  • signature_index
  • constant_pool [signature_index] is a CONSTANT_Utf8 string which is the signature of the field. See the section “Signatures” for more information on signatures.
  • attributes_count
  • This value indicates the number of additional attributes about this field.
  • attributes
  • A field can have any number of optional attributes associated with it. Currently, the only field attribute recognized is the “ConstantValue” attribute, which indicates that this field is a static numeric constant, and indicates the constant value of that field.
  • Any other attributes are skipped.
  • 2.5 Methods
  • The information for each method immediately follows the method count_field in the class file. Each method is described by a variable length method_info structure. The structure has the following format:
    method_info {
    u2 access_flags;
    u2 name_index;
    u2 signature_index;
    u2 attributes_count;
    attribute_info attributes [attribute_count];
    }
    access_flags
  • This is a set of sixteen flags used by classes, methods, and fields to describe various properties and how they many be accessed by methods in other classes. See the table “Access Flags” which gives the various bits in this field.
  • The possible fields that can be set for a method are ACC_PUBLIC, ACC_PRIVATE, ACC_PROTECTED, ACC_STATIC, ACC_FINAL, ACC_SYNCHRONIZED, ACC_NATIVE, and ACC_ABSTRACT.
  • At most one of ACC_PUBLIC, ACC_PROTECTED, and ACC_PRIVATE can be set for any method.
  • name_index
  • constant_pool[name_index] is a CONSTANT_Utf8 string giving the name of the method.
  • signature_index
  • constant_pool [signature_index] is a CONSTANT_Utf8 string giving the signature of the field. See the section “Signatures” for more information on signatures.
  • attributes_count
  • This value indicates the number of additional attributes about this field.
  • attributes
  • A field can have any number of optional attributes associated with it. Each attribute has a name, and other additional information. Currently, the only field attributes recognized are the “Code” and “Exceptions” attributes, which describe the bytecodes. that are executed to perform this method, and the JAVA Exceptions which are declared to result from the execution of the method, respectively.
  • Any other attributes are skipped.
  • 2.6 Attributes
  • Attributes are used at several different places in the class format. All attributes have the following format:
    GenericAttribute_info {
    u2 attribute_name;
    u4 attribute_length;
    u1 info[attribute_length];
    }
  • The attribute_name is a 16-bit index into the class's constant pool; the value of constant_pool [attribute_name] is a CONSTANT_Utf8 string giving the name of the attribute. The field attribute length indicates the length of the subsequent information in bytes. This length does not include the six bytes of the attribute_name and attribute_length.
  • In the following text, whenever we allow attributes, we give the name of the attributes that are currently understood. In the future, more attributes will be added. Class file readers are expected to skip over and ignore the information in any attribute they do not understand.
  • 2.6.1 SourceFile
  • The “SourceFile” attribute has the following format:
    SourceFile_attribute {
    u2 attribute_name_index;
    u4 attribute_length;
    u2 sourcefile_index;
    }

    attribute_name_index
  • constant_pool [attribute_name_index] is the CONSTANT_Utf8 string “SourceFile”.
  • attribute_length
  • The length of a SourceFile_attribute must be 2.
  • sourcefile_index
  • constant_pool [sourcefile_index] is a CONSTANT_Utf8 string giving the source file from which this class file was compiled.
  • 2.6.2 ConstantValue
  • The “ConstantValue” attribute has the following format:
    ConstantValue_attribute {
    u2 attribute_name_index;
    u4 attribute_length;
    u2 constantvalue_index;
    }

    attribute_name_index
  • constant_pool [attribute_name_index] is the CONSTANT_Utf8 string “ConstantValue”.
  • attribute_length
  • The length of a ConstantValue_attribute must be 2.
  • constantvalue_index
  • constant_pool [constantvalue_index] gives the constant value for this field.
  • The constant pool entry must be of a type appropriate to the field, as shown by the following table:
    long CONSTANT_Long
    float CONSTANT_Float
    double CONSTANT_Double
    int, short, char, byte, boolean CONSTANT_Integer

    2.6.3 Code
  • The “Code” attribute has the following format:
    Code_attribute {
    u2 attribute_name_index;
    u4 attribute_length;
    u2 max_stack;
    u2 max_locals;
    u4 code_length;
    u1 code[code_length];
    u2 exception_table_length;
    { u2 start_pc;
    u2 end_pc;
    u2 handler_pc;
    u2 catch_type;
    } exception_table[exception_table_length];
    u2 attributes_count;
    attribute_info attributes [attribute_count];
    }

    attribute_name_index
  • constant_pool [attribute_name_index] is the CONSTANT_Utf8 string “Code”.
  • attribute_length
  • This field indicates the total length of the “Code” attribute, excluding the initial six bytes.
  • max_stack
  • Maximum number of entries on the operand stack that will be used during execution of this method. See the other chapters in this spec for more information on the operand stack.
  • max_locals
  • Number of local variable slots used by this method. See the other chapters in this spec for more information on the local variables.
  • code_length
  • The number of bytes in the virtual machine code for this method.
  • code
  • These are the actual bytes of the virtual machine code that implement the method. When read into memory, if the first byte of code is aligned onto a multiple-of-four boundary the tableswitch and tablelookup opcode entries will be aligned; see their description for more information on alignment requirements.
  • exception_table_length
  • The number of entries in the following exception table.
  • exception_table
  • Each entry in the exception table describes one exception handler in the code.
  • start_pc, end_pc
  • The two fieldsstart_pc and end_pc indicate the ranges in the code at which the exception handler is active. The values of both fields are offsets from the start of the code.start_pc is inclusive.end_pc is exclusive.
  • handler_pc
  • This field indicates the starting address of the exception handler. The value of the field is an offset from the start of the code.
  • catch_type
  • If catch_type is nonzero, then constant_pool [catch_type] will be the class of exceptions that this exception handler is designated to catch. This exception handler should only be called if the thrown exception is an instance of the given class.
  • If catch_type is zero, this exception handler should be called for all exceptions.
  • attributes_count
  • This field indicates the number of additional attributes about code. The “Code” attribute can itself have attributes.
  • attributes
  • A “Code” attribute can have any number of optional attributes associated with it. Each attribute has a name, and other additional information. Currently, the only code attributes defined are the “LineNumberTable” and “LocalVariableTable,” both of which contain debugging information.
  • 2.6.4 Exceptions Table
  • This table is used by compilers which indicate which Exceptions a method is declared to throw:
    Exceptions_attribute {
    u2 attribute_name_index;
    u4 attribute_length;
    u2 number_of_exceptions;
    u2 exception_index_table [number_of_ex-
    ceptions];
    }

    attribute_name_index
  • constant_pool [attribute_name_index] will be the CONSTANT_Utf8 string “Exceptions”.
  • attribute_length
  • This field indicates the total length of the Exceptions_attribute, excluding the initial six bytes.
  • number_of_exceptions
  • This field indicates the number of entries in the following exception index table.
  • exception_index_table
  • Each value in this table is an index into the constant pool. For each table element (exception_index_table [i]|=0, where 0<=i <number_of_exceptions), then constant_pool [exception_index+table [i]] is a Exception that this class is declared to throw.
  • 2.6.5 LineNumberTable
  • This attribute is used by debuggers and the exception handler to determine which part of the virtual machine code corresponds to a given location in the source. The LineNumberTable_attribute has the following format:
    LineNumberTable_attribute {
    u2 attribute_name_index;
    u4 attribute_length;
    u2 line_number_table_length;
    { u2 start_pc;
    u2 line_number;
    } line_number_table[line
    number_table_length];
    }

    attribute_name_index
  • constant_pool [attribute_name_index] will be the CONSTANT_Utf8 string “LineNumberTable”.
  • attribute_length
  • This field indicates the total length of the LineNumberTable_attribute, excluding the initial six bytes.
  • line_number_table_length
  • This field indicates the number of entries in the following line number table.
  • line_number_table
  • Each entry in the line number table indicates that the line number in the source file changes at a given point in the code.
  • start_pc
  • This field indicates the place in the code at which the code for a new line in the source begins. source_pc <<SHOULD THAT BEstart_pc?>> is an offset from the beginning of the code.
  • line_number
  • The line number that begins at the given location in the file.
  • 2.6.6 LocalVariableTable
  • This attribute is used by debuggers to determine the value of a given local variable during the dynamic execution of a method. The format of the LocalVariableTable_attribute is as follows:
    LocalVariableTable_attribute {
    u2 attribute_name_index;
    u4 attribute_length;
    u2 local_variable_table_length;
    { u2 start_pc;
    u2 length;
    u2 name_index;
    u2 signature_index;
    u2 slot;
    } local_variable_table[local
    variable_table_length];
    }

    attribute_name_index
  • constant_pool [attribute_name_index] will be the CONSTANT_Utf8 string “LocalVariableTable”.
  • attribute_length
  • This field indicates the total length of the LineNumberTable_attribute, excluding the initial six bytes.
  • local_variable_table_length
  • This field indicates the number of entries in the following local variable table.
  • local_variable_table
  • Each entry in the local variable table indicates a code range during which a local variable has a value. It also indicates where on the stack the value of that variable can be found.
  • start_pc, length
  • The given local variable will have a value at the code between start_pc andstart_pc+length. The two values are both offsets from the beginning of the code.
  • name_index, signature_index
  • constant_pool [name_index] and constant_pool [signature_index] are CONSTANT_Utf8 strings giving the name and signature of the local variable.
  • slot
  • The given variable will be the slotth local variable in the method's frame.
  • 3. The Virtual Machine Instruction Set
  • 3.1 Format for the Instructions
  • JAVA Virtual Machine instructions are represented in this document by an entry of the following form.
  • instruction name
  • Short description of the instruction
  • Syntax:
    opcode=number
    operand1
    operand2
    ...
  • Stack: . . . , value1, value2
    Figure US20070277021A1-20071129-P00001
    . . . , value3
      • A longer description that explains the functions of the instruction and indicates any exceptions that might be thrown during execution.
  • Each line in the syntax table represents a single 8-bit byte.
  • Operations of the JAVA Virtual Machine most often take their operands from the stack and put their results back on the stack. As a convention, the descriptions do not usually mention when the stack is the source or destination of an operation, but will always mention when it is not. For instance, instruction iload has the short description “Load integer from local variable.” Implicitly, the integer is loaded onto the stack. Instruction iadd is described as “Integer add”; both its source and destination are the stack.
  • Instructions that do not affect the control flow of a computation may be assumed to always advance the virtual machine program counter to the opcode of the following instruction. Only instructions that do affect control flow will explicitly mention the effect they have on the program counter.
  • 3.2 Pushing Constants onto the Stack
  • bipush
  • Push one-byte signed integer
  • Syntax:
    bipush=16
    byte1
  • Stack: . . . => . . . , value
  • byte1 is interpreted as a signed 8-bitvalue. This value is expanded to an integer and pushed onto the operand stack.
  • sipush
      • Push two-byte signed integer
  • Syntax:
    sipush=17
    byte1
    byte2
      • Stack: . . . => . . . , item
  • byte1 and byte2 are assembled into a signed 16-bit value. This value is expanded to an integer and pushed onto the operand stack.
  • ldc1
  • Push item from constant pool
  • Syntax:
    ldc1=18
    indexbyte1
  • Stack: . . . => . . . , item
  • indexbyte1 is used as an unsigned 8-bit index into the constant pool of the current class. The item at that index is resolved and pushed onto the stack. If a String is being pushed and there isn't enough memory to allocate space for it then an OutOfMemoryError is thrown.
  • Note: A String push results in a reference to an object.
  • ldc2
  • Push item from constant pool
  • Syntax:
    ldc2=19
    indexbyte1
    indexbyte2
  • Stack: . . . => . . . , item
  • indexbyte1 and indexbyte2 are used to construct an unsigned 16-bit index into the constant pool of the current class. The item at that index is resolved and pushed onto the stack. If a String is being pushed and there isn't enough memory to allocate space for it then an OutOfMemoryError is thrown.
  • Notes A String push results in a reference to an object.
  • ldc2w
  • Push long or double from constant pool
  • Syntax:
    ldc2w=20
    indexbyte1
    indexbyte2
  • Stack: . . . => . . . , constant-word1, constant-word2 indexbyte1 and indexbyte2 are used to construct an unsigned 16-bit index into the constant pool of the current class. The two-word constant that index is resolved and pushed onto the stack.
  • aconst_null
  • Push null object reference
  • Syntax:.
    aconst_null=1
  • Stack: . . . => . . . ,null
  • Push the null object reference onto the stack.
  • iconst_m1
  • Push integer constant −1
  • Syntax:
    iconst_ml=2
  • Stack: . . . => . . . , 1
  • Push the integer −1 onto the stack.
  • iconst_<n>
  • Push integer constant
  • Syntax:
    iconst_<n>
  • Stack: . . . => . . ., <n>
  • Forms: iconst0=3, iconst 1=4, iconst 2=5, iconst 3=6, iconst4=7, iconst5=8
  • Push the integer <n> onto the stack.
  • lconst_<l>
  • Push long integer constant
  • Syntax:
    lconst_<l>
  • Stack: . . . => . . . <l>-word1, <l>-word2
  • Forms: lconst0=9, iconst 1=10
  • Push the long integer <I>onto the stack.
  • fconst_<f>
  • Push single float
  • Syntax:
    fconst_<f>
  • Stack: . . . => . . . <f>
  • Forms: fconst0=11, fconst 1=12, fconst 2=13
  • Push the single-precision floating point number <f> onto the stack.
  • dconst_<d>
  • Push double float
  • Syntax:
    dconst_<d>
  • Stack: . . . => . . . , <d>-word1, <d>-word2
  • Forms: dconst0=14, dconst 1=15
  • Push the double-precision floating point number <d> onto the stack.
  • 3.3 Loading Local Variables Onto the Stack
  • lload
  • Load integer from local variable
  • Syntax:
    iload=21
    vindex
  • Stack: . . . => . . . , value
  • The value of the local variable at vindex in the current JAVA frame is pushed onto the operand stack.
  • iload_<n>
  • Load integer from local variable
  • Syntax:
    iload_<n>
  • Stack: . . . => . . . , value
  • Forms: iload0=26, iload 1=27,iload 2=28, iload 3=29
  • The value of the local variable at <n> in the current JAVA frame is pushed onto the operand stack.
  • This instruction is the same as iload with a vindex of <n>, except that the operand <n> is implicit.
  • iload
  • Load long integer from local variable
  • Syntax:
    iload = 22
    vindex
  • Stack: . . . =>. . . , value-word1, value-work2
  • The value of the local variables at vindex and vindex+1 in the current JAVA frame is pushed onto the operand stack.
  • lload_<n>
  • Load long integer from local variable
  • Syntax:
    iload_<n>
  • Stack: . . . => . . . , value-word1, value-word2
  • Forms: lload0=30, lload 1=31, lload 2=32, lload 3=33
  • The value of the local variables at <n> and <n>+1 in the current JAVA frame is pushed onto the operand stack.
  • This instruction is the same as lload with a vindex of <n>, except that the operand <n> is implicit.
  • fload
  • Load single float from local variable
  • Syntax:
    fload = 23
    vindex
  • Stack: . . . => . . . , value
  • The value of the local variable at vindex in the current JAVA frame is pushed onto the opera and stack.
  • fload_<n>
  • Load single float from local variable
  • Syntax:
    fload_<n>
  • Stack: . . . => . . . ,value
  • Forms: fload0=34, fload 1=35, fload 2=36, fload 3=37
  • The value of the local variable at <n> in the current JAVA frame is pushed onto the operand stack.
  • This instruction is the same as fload with a vindex of <n>, except that the operand <n> is implicit.
  • dload
  • Load double float from local variable
  • Syntax:
    dload = 24
    vindex
  • Stack: . . . => . . . , value-word1, value-word2
  • The value of the local variables at vindex and vindex+1 in the current JAVA frame is pushed onto the operand stack.
  • dload_<n>
  • Load double float from local variable
  • Syntax:
    dload_<n>
  • Stack: . . . => . . . value-word1, value-word2
  • Forms: dload0=38, dload 1=39, dload 2=40, dload 3=41
  • The value of the local variables at <n> and <n>+1 in the current JAVA frame is pushed onto the operand stack.
  • This instruction is the same as dload with a vindex of <n>, except that the operand <on> is implicit.
  • aload
  • Load object reference from local variable
  • Syntax:
    aload = 25
    vindex
  • Stack: . . . => . . . , value
  • The value of the local variable at vindex in the current JAVA frame is pushed onto the operand stack.
  • aload_<n>
  • Load object reference from local variable
  • Syntax:
    aload_<n>
  • Stack: . . . => . . . , value
  • Forms: aload0=42, aload 1=43, aload 2=44, aload 3=45
  • The value of the local variable at <n> in the current JAVA frame is pushed onto the operand stack.
  • This instruction is the same as aload with a vindex of <n>, except that the operand <n> is implicit.
  • 3.4 Storing Stack Values into Local Variables
  • istore
  • Store integer into local variable
  • Syntax:
    istore = 54
    vindex
  • Stack: . . . , value => . . .
      • value must be an integer. Local variable vindex in the current JAVA frame is set to value.
        istore_<n>
  • Store integer into local variable
  • Syntax:
    istore_<n>
  • Stack: . . . , value => . . .
  • Forms: istore0=59, istore 1=60, istore 2=61, istore 3=62
  • value must be an integer. Local variable <n> in the current JAVA frame is set to value.
  • This instruction is the same as istore with a vindex of <n>, except that the operand <n> is implicit.
  • lstore
  • Store long integer into local variable
  • Syntax:
    lstore = 55
    vindex
  • Stack: . . . , value-word1, value-word2=> . . .
  • value must be a long integer. Local variables vindex+1 in the current JAVA frame are set to value.
  • lstore_<n>
  • Store long integer into local variable
  • Syntax:
    lstore_<n>
  • Stack: . . . value-word1, value-word2=>
  • Forms: lstore0=63, lstore 1=64, lstore 2=65, lstore 3=66
  • value must be a long integer. Local variables <n> and <n>+1 in the current JAVA frame are set to value.
  • This instruction is the same as lstore with a vindex of <n>, except that the operand <n>is implicit. fstore
  • Store single float into local variable
  • Syntax:
    fstore =56
    vindex
  • Stack: . . . , value=>. . .
  • value must be a single-precision floating point number. Local variable vindex in the current JAVA frame is set to value.
  • fstore_<n>
  • Store single float into local variable
  • Syntax:
    fstore_<n>
  • Stack: . . . value=> . . .
  • Forms: fstore0=67, fstore 1=68, fstore 2=69, fstore 3=70
  • value must be a single-precision floating point number. Local variable <n> in the current JAVA frame is set to value.
  • This instruction is the same as fstore with a vindex of <n>, except that the operand <n> is implicit.
  • dstore
  • Store double float into local variable
  • Syntax:
    dstore = 57
    vindex
  • Stack: . . . , value-word1, value-word2=> . . .
  • value must be a double-precision floating point number. Local variables vindex and vindex+1 in the current JAVA frame are set to value.
  • dstore_<n>
  • Store double float into local variable
  • Syntax:
    dstore_<n>
  • Stack: . . . , value-word1, value-word2=> . . .
  • Forms: dstore0=71, dstore 1=72, dstore 2=73, dstore 3=74
  • value must be a double-precision floating point number. Local variables <n> and <n>+1 in the current JAVA frame are set to value.
  • This instruction is the same as dstore with a vindex of <n>, except that the operand <n> is implicit.
  • astore
  • Store object reference into local variable
  • Syntax:
    astore=58
    vindex
  • Stack: . . . , value=> . . .
  • value must be a return address or a reference to an object. Local variable vindex in the current JAVA frame is set to value.
  • astore_<n>
  • Store object reference into local variable
  • Syntax:
    astore_<n>
  • Stack: . . . , value=> . . .
  • Forms: astore0=75, astore 1=76, astore 2=77, astore 3=78
  • value must be a return address or a reference to an object. Local variable <n> in the current JAVA frame is set to value.
  • This instruction is the same as astore with a vindex of <n>, except that the operand <n> is implicit.
  • iinc
  • Increment local variable by constant
  • Syntax:
    iinc = 132
    vindex
    const
  • Stack: no change
  • Local variable vindex in the current JAVA frame must contain an integer. Its value is incremented by the value const, where const is treated as a signed 8-bit quantity.
  • 3.5 Wider Index for Loading, Storing and Incrementing
  • Wide
  • Wider index for accessing local variables in load, store and increment.
  • Syntax:
    wide = 196
    vindex2
  • Stack: no change
  • This bytecode must precede one of the following bytecodes: iload, lload, fload, dload, aload, istore, lstore, fstore, dstore, astore, iinc. The vindex of the following bytecode and vindex2 from this bytecode are assembled into an unsigned 16-bit index to a local variable in the current JAVA frame. The following bytecode operates as normal except for the use of this wider index.
  • 3.6 Managing Arrays
  • newarray
  • Allocate new array
  • Syntax:
    newarray = 188
    atype
  • Stack: . . . , size=>result
  • size must be an integer. It represents the number of elements in the new array.
  • atype is an internal code that indicates the type of array to allocate. Possible values for atype are as follows:
    T_BOOLEAN 4
    T_CHAR 5
    T_FLOAT 6
    T_DOUBLE 7
    T_BYTE 8
    T_SHORT 9
    T_INT 10
    T_LONG 11
  • A new array of atype, capable of holding size elements, is allocated, and result is a reference to this new object. Allocation of an array large enough to contain size items of atype is attempted. All elements of the array are initialized to zero.
  • If size is less than zero, a NegativeArraySizeException is thrown. If there is not enough memory to allocate the array, anOutOfMemoryError is thrown.
  • anewarray
  • Allocate new array of references to objects
  • Syntax:
    anewarray = 189
    indexbyte1
    indexbyte2
  • Stack: . . . , size=>result
  • size must be an integer. It represents the number of elements in the new array. indexbyte1 and indexbyte2 are used to construct an index into the constant pool of the current class. The item at that index is resolved. The resulting entry must be a class.
  • A new array of the indicated class type and capable of holding size elements is allocated, and result is a reference to this new object. Allocation of an array large enough to contain size items of the given class type is attempted. All elements of the array are initialized to null.
  • If size is less than zero, a NegativeArraySizeException is thrown. If there is not enough memory to allocate the array, an OutOfMemoryError is thrown.
  • anewarray is used to create a single dimension of an array of object references. For example, to create
      • new Thread[7]
  • the following code is used:
      • bipush 7
      • anewarray <Class “JAVA.lang.Thread”>
  • anewarray can also be used to create the first dimension of a multi-dimensional array. For example, the following array declaration:
      • new int[6] [ ]
  • is created with the following code:
      • bipush 6
      • anewarray <Class “[I”>
  • See CONSTANT_Class in the “Class File Format” chapter for information on array class names.
  • multianewarray
  • Allocate new multi-dimensional array
  • Syntax:
    multianwearray=197
    indexbyte1
    indexbyte2
    dimensions
  • Stack: . . . , size1 size2 . . . sizen=>result
  • Each size must be an integer. Each represents the number of elements in a dimension of the array.
  • indexbyte1 and indexbyte2 are used to construct an index into the constant pool of the current class. The item at that index is resolved. The resulting entry must be an array class of one or more dimensions.
  • dimensions has the following aspects:
      • It must be an integer ≧1.
      • It represents the number of dimensions being created. It must be s the number of dimensions of the array class.
      • It represents the number of elements that are popped off the stack. All must be integers greater than or equal to zero. These are used as the sizes of the dimension. For example, to create
        • new int[6] [3] [ ]
      • the following code is used:
        • bipush 6
        • bipush 3
        • multianewarray <Class “[[[I”>2
  • If any of the size arguments on the stack is less than zero, a NegativeArraySizeException is thrown. If there is not enough memory to allocate the array, an OutOfMemoryError is thrown.
  • The result is a reference to the new array object.
  • Note: It is more efficient to use newarray or anewarray when creating a single dimension.
  • See CONSTANT_Class in the “Class File Format” chapter for information on array class names.
  • arraylength
  • Get length of array
  • Syntax;
    arraylength=190
  • Stack: . . . , objectref=> . . . , length
  • objectref must be a reference to an array object. The length of the array is determined and replaces objectref on the top of the stack.
  • If the objectref is null, a NullPointerException is thrown.
  • iaload
  • Load integer from array
  • Syntax:
    iaload=46
  • Stack: . . . , arrayref, index=> . . . , value
  • arrayref must be a reference to an array of integers.index must be an integer. The integer value at position number index in the array is retrieved and pushed onto the top of the stack.
  • If arrayref is null a NullPointerException is thrown. If index is not within the bounds of the array an ArrayIndexOutOfBoundsException is thrown.
  • laload
  • Load long integer from array
  • Syntax:
    laoad=47
  • Stack: . . . , arrayref, index=> . . . , value-word1, value-word2
  • arrayref must be a reference to an array of long integers, index must be an integer. The long integer value at position number index in the array is retrieved and pushed onto the top of the stack.
  • If arrayref is null a NullPointerException is thrown. If index is not within the bounds of the array an ArrayIndexOutOfBoundsException is thrown.
  • faload
  • Load single float from array
  • Syntax:
    faload=48
  • Stack: . . . , arrayref, index=> . . . , value
  • arrayref must be a reference to an array of single-precision floating point numbers, index must be an integer. The single-precision floating point number value at position number index in the array is retrieved and pushed onto the top of the stack.
  • If arrayref is null a NullPointerException is thrown. If index is not within the bounds of the array an ArrayIndexOutOfBoundsException is thrown.
  • daload
  • Load double float from array
  • Syntax:
    daload=49
  • Stack: . . . , arrayref, index=> . . . , value-word1, value-word2
  • arrayref must be a reference to an array of double-precision floating point numbers. index must be an integer. The double-precision floating point number value at position number index in the array is retrieved and pushed onto the top of the stack.
  • If arrayref is null a NullPointerException is thrown. If index is not within the bounds of the array an ArrayIndexOutOfBoundsException is thrown.
  • aaload
  • Load object reference from array
  • Syntax:
    aaload=50
  • Stack: . . . , arrayref, index=> . . . , value
  • arrayref must be a reference to an array of references to objects. index must be an integer. The object reference at position number index in the array is retrieved and pushed onto the top of the stack.
  • If arrayref is null a NullPointerException is thrown. If index is not within the bounds of the array an ArrayIndexOutOfBoundsException is thrown.
  • baload
  • Load signed byte from array.
  • Syntax:
    baload=51
  • Stack: . . . , arrayref, index=> . . . value
  • arrayref must be a reference to an array of signed bytes. index must be an integer. The signed byte value at position number index in the array is retrieved, expanded to an integer, and pushed onto the top of the stack.
  • If arrayref is null a NullPointerException is thrown. If index is not within the bounds of the array an ArrayIndexOutOfBoundsException is thrown.
  • caload
  • Load character from array
  • Syntax:
    caload=52
  • Stack: . . . , arrayref, index=> . . . ,value
  • arrayref must be a reference to an array of characters. index must be an integer. The character value at position number index in the array is retrieved, zero-extended to an integer, and pushed onto the top of the stack.
  • If arrayref is null a NullPointerException is thrown. If index is not within the bounds of the array an ArrayIndexOutOfBoundsException is thrown.
  • saload
  • Load short from array
  • Syntax:
    saload=53
  • Stack: . . . , arrayref, index=> . . . , value
  • arrayref must be a reference to an array of short integers. index must be an integer. The ;signed short integer value at position number index in the array is retrieved, expanded to an integer, and pushed onto the top of the stack.
  • If arrayref is null, a NullPointerException is thrown. If index is not within the bounds of the array an ArrayIndexOutOfBoundsException is thrown.
  • iastore
  • Store into integer array
  • Syntax:
    iastore=79
  • Stack: . . . , arrayref, index, value=> . . .
  • arrayref must be a reference to an array of integers, index must be an integer, and value an integer. The integer value is stored at position index in the array.
  • If arrayref is null, a NullPointerException is thrown. If index is not within the bounds of the array an ArrayIndexOutOfBoundsException is thrown.
  • lastore
  • Store into long integer array
  • Syntax:
    lastore=80
  • Stack: . . . , arrayref, index, value-word1, value-word2=> . . .
  • arrayref must be a reference to an array of long integers, index must be an integer, and value a long integer. The long integer value is stored at position index in the array.
  • If arrayref is null, a NullPointerException is thrown. If index is not within the bounds of the array, an ArrayIndexOutOfBoundsException is thrown.
  • fastore
  • Store into single float array
  • Syntax:
    fastore=81
  • Stack: . . . , arrayref, index, value=> . . .
  • arrayref must be an array of single-precision floating point numbers, index must be an integer, and value a single-precision floating point number. The single float value is stored at position index in the array.
  • If arrayref is null, a NullPointerException is thrown. If index is not within the bounds of the array an ArrayIndexOutOfBoundsException is thrown.
  • dastore
  • Store into double float array
  • Syntax:
    dastore=82
  • Stack: . . . , arrayref, index, value-word1, value-word2=> . . .
  • arrayref must be a reference to an array of double-precision floating point numbers, index must be an integer, and value a double-precision floating point number. The double float value is stored at position index in the array.
  • If arrayref is null, a NullPointerException is thrown. If index is not within the bounds of the array an ArrayIndexOutOfBoundsException is thrown.
  • aastore
  • Store into object reference array
  • Syntax:
    aastore=83
  • Stack: . . . , arrayref, index, value=> . . .
  • arrayref must be a reference to an array of references to objects, index must be an integer, and value a reference to an object. The object reference value is stored at position index in the array.
  • If arrayref is null, a NullPointerException is thrown. If index is not within the bounds of the array, an ArrayIndexOutOfBoundsException is thrown.
  • The actual type of value must be conformable with the actual type of the elements of the array. For example, it is legal to store an instance of class Thread in an array of class Object, but not vice versa. An ArrayStoreException is thrown if an attempt is made to store an incompatible object reference.
  • bastore
  • Store into signed byte array
  • Syntax:
    bastore=84
  • Stack: . . . , arrayref, index, value=> . . .
  • arrayref must be a reference to an array of signed bytes, index must be an integer, and value an integer. The integer value is stored at position index in the array. If value is too large to be a signed byte, it is truncated.
  • If arrayref is null, a NullPointerException is thrown. If index is not within the bounds of the array an ArrayIndexOutOfBoundsException is thrown.
  • castore
  • Store into character array
  • Syntax:
    castore=85
  • Stack: . . . , arrayref, index, value=> . . .
  • arrayref must be an array of characters, index must be an integer, and value an integer. The integer value is stored at position index in the array. If value is too large to be a character, it is truncated.
  • If arrayref is null, a NullPointerException is thrown. If index is not within the bounds of [the array an ArrayIndexOutOfBoundsException is thrown.
  • sastore
  • Store into short array
  • Syntax:
    sastore=86
  • Stack: . . . , array, index, value=> . . .
  • arrayref must be an array of shorts, index must be an integer, and value an integer. The integer value is stored at position index in the array. If value is too large to be an short, it is truncated.
  • If arrayref is null, a NullPointerException is thrown. If index is not within the bounds of the array an ArrayIndexOutOfBoundsException is thrown.
  • 3.7 Stack Instructions
  • nop
  • Do nothing
  • Syntax:
    nop=0
  • Stack: no change
  • Do nothing.
  • pop
  • Pop top stack word
  • Syntax:
    pop=87
  • Stack: . . . , any=> . . .
  • Pop the top word from the stack.
  • pop2
  • Pop top two stack words
  • Syntax:
    pop2=89
  • Stack: . . . , any2, any1=> . . .
  • Pop the top two words from the stack.
  • dup
  • Duplicate top stack word
  • Syntax:
    dup=89
  • Stack: . . . , any=> . . . , any,any
  • Duplicate the top word on the stack.
  • dup2
  • Duplicate top two stack words
  • Syntax:
    dup2=92
  • Stack: . . . , any2,any1=> . . . , any2, any1,any2, any1
  • Duplicate the top two words on the stack.
  • dup_x1
  • Duplicate top stack word and put two down
  • Syntax:
    dup_x1=90
  • Stack: . . . , any2, any1=> . . . , any1, any2, any1
  • Duplicate the top word on the stack and insert the copy two words down in the stack.
  • dup2_x1
  • Duplicate top two stack words and put two down
  • Syntax:
    dup2_x1=93
  • Stack: . . . , any3, any2, any1=> . . . , any2, any1, any3, any2, any1
  • Duplicate the top two words on the stack and insert the copies two words down in the stack.
  • dup_x2
  • Duplicate top stack word and put three down
  • Syntax:
    dup_x2=91
  • Stack: . . . , any3, any2, any1=> . . . , any1, any3, any2, any1
  • Duplicate the top word on the stack and insert the copy three words down in the stack.
  • dup2_x2
  • Duplicate top two stack words and put three down
  • Syntax:
    dup2_x2=94
  • Stack: . . . , any4, any3, any2, any1=> . . . , any2, any1, any4, any3, any2, any1
  • Duplicate the top two words on the stack and insert the copies three words down in the stack.
  • swap
  • Swap top two stack words
  • Syntax:
    swap=95
  • Stack: . . . , any2, any1=> . . . , any2, any1
  • Swap the top two elements on the stack.
  • 3.8 Arithmetic Instructions
  • iadd
  • Integer add
  • Syntax:
    iadd=96
  • Stack: . . . , value1, value2=> . . . , result
  • value1 and value 2 must be integers. The values are added and are replaced on the stack by their integer sum.
  • ladd
  • Long integer add
  • Syntax:
    ladd=97
  • Stack: . . . , value1-word1, value1-word2, value2-word1, value2-word2=> . . . , result-word1, result-word2
  • value1 and value 2 must be long integers. The values are added and are replaced on the stack by their long integer sum.
  • fadd
  • Single floats add
  • Syntax:
    fadd=98
  • Stack: . . . , value1, value2=> . . . , result
  • value1 and value 2 must be single-precision floating point numbers. The values are added and are replaced on the stack by their single-precision floating point sum.
  • dadd
  • Double floats add
  • Syntax:
    dadd=99
  • Stack: . . . , value1-word1, value1-word2, value2-word1, value2-word2=> . . . , result-word1, result-word2
  • value1 and value 2 must be double-precision floating point numbers. The values are added and are replaced on the stack by their double-precision floating point sum.
  • isub
  • Integer subtract
  • Syntax:
    isub=100
  • Stack: . . . , value1, value2=> . . . , result
  • value1 and value 2 must be integers. value2 is subtracted from value1, and both values are replaced on the stack by their integer difference.
  • lsub
  • Long integer subtract
  • Syntax:
    lsub=101
  • Stack: . . . , value1-word1, value1-word2, value2-word1, value2-word2=> . . . , result-word1, result-word2
  • value1 and value 2 must be long integers. valuo2 is subtracted from value1 , and both values are replaced on the stack by their long integer difference.
  • fsub
  • Single float subtract
  • Syntax:
    fsub=102
  • Stack: . . . , value1, value2=> . . . , result
  • value1 and value 2 must be single-precision floating point numbers. value2 is subtracted from value1, and both values are replaced on the stack by their single-precision floating point difference.
  • dsub
  • Double float subtract
  • Syntax:
    dsub=103
  • Stack: . . . , value1-word1, value1-word2, value2-word1, value2-word2=> . . . , result-word1, result-word2
  • value1 and value 2 must be double-precision floating point numbers. value2 is subtracted from value1, and both values are replaced on the stack by their double-precision floating point difference.
  • imul
  • Integer multiply
  • Syntax:
    imul=104
  • Stack: . . . , value1, value2=> . . . , result
  • value1 and value 2 must be integers. Both values are replaced on the stack by their integer product.
  • lmul
  • Long integer multiply
  • Syntax:
    lmul=105
  • Stack: . . . , value1-word1, value1-word2, value2-word1, value2-word2=> . . . , result-word1, result-word2
  • value1 and value 2 must be long integers. Both values are replaced on the stack by their long integer product.
  • fmul
  • Single float multiply
  • Syntax:
    fmul=106
  • Stack: . . . , value1, value2=> . . . , result
  • value1 and value 2 must be single-precision floating point numbers. Both values are replaced on the stack by their single-precision floating point product.
  • dmul
  • Double float multiply
  • Syntax:
    dmul=107
  • Stack: . . . , value1-word1, value1-word2, value2-word1, value2-word2=> . . . result-word1, result-word2
  • value1 and value 2 must be double-precision floating point numbers. Both values are replaced on the stack by their double-precision floating point product.
  • idiv
  • Integer divide
  • Syntax:
    idiv=108
  • Stack: . . . , value1, value2=> . . . , result
  • value1 and value 2 must be integers. value1 is divided by value2, and both values are replaced on the stack by their integer quotient.
  • The result is truncated to the nearest integer that is between it and 0. An attempt to divide by zero results in a “/ by zero” ArithmeticException being thrown.
  • ldiv
  • Long integer divide
  • Syntax:
    ldiv=109
  • Stack: . . . , value1-word1, value1-word2, value2-word1, value2-word2=> . . . , result-word1, result-word2
  • value1 and value 2 must be long integers. value1 is divided by value2, and both values are replaced on the stack by their long integer quotient.
  • The result is truncated to the nearest integer that is between it and 0. An attempt to divide by zero results in a “/ by zero” ArithmeticException being thrown.
  • fdiv
  • Single float divide
  • Syntax:
    fdiv=110
  • Stack: . . . , value1, value2=> . . . , result
  • value1 and value 2 must be single-precision floating point numbers. value1 is divided by value2, and both values are replaced on the stack by their single-precision floating point quotient.
  • Divide by zero results in the quotient being NaN.
  • ddiv
  • Double float divide
  • Syntax:
    ddiv=111
  • Stack: . . . , value1-word1, value1-word2, value2-word1, value2-word2=> . . . result-word1, result-word2
  • value1 and value 2 must be double-precision floating point numbers. value1 is divided by value2, and both values are replaced on the stack by their double-precision floating point quotient.
  • Divide by zero results in the quotient being NaN.
  • irem
  • Integer remainder
  • Syntax:
    irem=112
  • Stack: . . . , value1, value2=> . . . , result
  • value1 and value 2 must both be integers. value1 is divided by value2, and both values are replaced on the stack by their integer remainder.
  • An attempt to divide by zero results in a “/ by zero” ArithmeticException being thrown.
  • lrem
  • Long integer remainder
  • Syntax:
    lrem=113
  • Stack: . . . value1-word1, value1-word2, value2-word1, value2-word2=> . . . , result-word1, result-word2
  • value1 and value 2 must both be long integers. value1 is divided by value2, and both values are replaced on the stack by their long integer remainder.
  • An attempt to divide by zero results in a “/ by zero” ArithmeticException being thrown.
  • frem
  • Single float remainder
  • Syntax:
    frem=114
  • Stacks . . . , value1, value2=> . . . , result
  • value1and value 2 must both be single-precision floating point numbers. value1 is divided by value2, and the quotient is truncated to an integer, and then multiplied by value2. The product is subtracted from value1. The result, as a single-precision floating point number, replaces both values on the stack. result=value1−(integral_part(value1/value2) *value2), where integral_part( ) rounds to the nearest integer, with a tie going to the even number.
  • An attempt to divide by zero results in NaN.
  • drem
  • Double float remainder
  • Syntax:
    drem=115
  • Stack: . . . , value1-word1, value1-word2, value2-word1, value2-word2=> . . . , result-word1, result-word2
  • value1 and value 2 must both be double-precision floating point numbers. value1 is divided by value2, and the quotient is truncated to an integer, and then multiplied by value2. The product is subtracted from value1. The result, as a double-precision floating point number, replaces both values on the stack. result=value1−(integral_part(value1/value2)*value2), where integral_part( ) rounds to the nearest integer, with a tie going to the even number.
  • An attempt to divide by zero results in NaN.
  • ineg
  • Integer negate
  • Syntax:
    ineg=116
  • Stack: . . . , value=> . . . , result
  • value must be an integer. It is replaced on the stack by its arithmetic negation.
  • lneg
  • Long integer negate
  • Syntax:
    lneg=117
  • Stack: . . . , value-word1, value-word2=> . . . , result-word1, result-word2
  • value must be a long integer. It is replaced on the stack by its arithmetic negation.
  • fneg
  • Single float negate
  • Syntax:
    fneg=118
  • Stack: . . . , value=> . . . , result
  • value must be a single-precision floating point number. It is replaced on the stack by its arithmetic negation.
  • dneg
  • Double float negate
  • Syntax:
    dneg=119
  • Stack: . . . , value-word1, value-word2=> . . . , result-word1, result-word2
  • value must be a double-precision floating point number. It is replaced on the stack by its arithmetic negation.
  • 3.9 Logical Instructions
  • ishl
  • Integer shift left
  • Syntax:
    ishl=120
  • Stack: . . . ,value1, value2=> . . . , result
  • value1 and value 2 must be integers. value1 is shifted left by the amount indicated by the low five bits of value2. The integer result replaces both values on the stack.
  • ishr
  • Integer arithmetic shift right
  • Syntax:
    ishr=122
  • Stack: . . . , value1, value2=> . . . , result
  • value1 and value 2 must be integers. value1 is shifted right arithmetically (with sign extension) by the amount indicated by the low five bits of value2. The integer result replaces both values on the stack.
  • iushr
  • Integer logical shift right
  • Syntax:
    iushr=124
  • Stack: . . . , value1, value2=> . . . , result
  • value1 and value 2 must be integers. value1 is shifted right logically (with no sign extension) by the amount indicated by the low five bits of value2. The integer result replaces both values on the stack.
  • lshl
  • Long integer shift left
  • Syntax:
    lshl=121
  • Stack: . . . , value1-word1, value1-word2, value2=> . . . , result-word1, result-word2
  • value1 must be a long integer and value 2 must be an integer. value1 is shifted left by the amount indicated by the low six bits of value2. The long integer result replaces both values on the stack.
  • lshr
  • Long integer arithmetic shift right
  • Syntax:
    lshr=123
  • Stack: . . . , value1-word1, value1-word2, value2=> . . . , result-word1, result-word2
  • value1 must be a long integer and value 2 must be an integer. value1 is shifted right arithmetically (with sign extension) by the amount indicated by the low six bits of value2. The long integer result replaces both values on the stack.
  • lushr
  • Long integer logical shift right
  • Syntax:
    lushr=125
  • Stack: . . . , value1-word1, value1-word2, value2-word1, value2-word2=> . . . , result-word1, result-word2
  • value1 must be a long integer and value 2 must be an integer. value1 is shifted right logically (with no sign extension) by the amount indicated by the low six bits of value2. The long integer result replaces both values on the stack.
  • iand
  • Integer boolean AND
  • Syntax:
    iand=126
  • Stack: . . . , value1, value2=> . . . , result
  • value1 and value 2 must both be integers. They are replaced on the stack by their bitwise logical and (conjunction).
  • land
  • Long integer boolean AND
  • Syntax:
    land=127
  • Stack: . . . , value1-word1, value1-word2, value2-word1, value2-word2=> . . . , result-word1, result-word2
  • value1and value 2 must both be long integers. They are replaced on the stack by their bitwise logical and (conjunction).
  • ior
  • Integer boolean OR
  • Syntax:
    ior=128
  • Stack: . . . , value1, value2=> . . . , result
  • value1 and value 2 must both be integers. They are replaced on the stack by their bitwise logical or (disjunction).
  • lor
  • Long integer boolean OR
  • Syntax:
    lor=129
  • Stack: . . . , value1-word1, value1-word2, value2-word1, value2-word2=> . . . , result-word1, result-word2
  • value1 and value 2 must both be long integers. They are replaced on the stack by their bitwise logical or (disjunction).
  • ixor
  • Integer boolean XOR
  • Syntax:
    ixor=130
  • Stack: . . . , value1, value2=> . . . , result
  • value1 and value 2 must both be integers. They are replaced on the stack by their bitwise exclusive or (exclusive disjunction).
  • lxor
  • Long integer boolean XOR
  • Syntax:
    lxor=131
  • Stack: . . . , value1-word1, value1-word2, value2-word1, value2-word2=> . . . result-word1, result-word2
  • value1 and value 2 must both be long integers. They are replaced on the stack by their bitwise exclusive or (exclusive disjunction).
  • 3.10 Conversion Operations
  • i2l
  • Integer to long integer conversion
  • Syntax:
    i2l=133
  • Stack: . . . , value=> . . . result-word1, result-word2
  • value must be an integer. It is converted to a long integer. The result replaces value on the stack.
  • i2f
  • Integer to single float
  • Syntax:
    i2f=134
  • Stack: . . . , value=> . . . ,result
  • value must be an integer. It is converted to a single-precision floating point number. The result replaces value on the stack.
  • i2d
  • Integer to double float
  • Syntax:
    i2d=135
  • Stack: . . . , value=> . . . , result-word1, result-word2
  • value must be an integer. It is converted to a double-precision floating point number. The result replaces value on the stack.
  • l2i
  • Long integer to integer
  • Syntax:
    l2i=136
  • Stack: . . . , value-word1, value-word=> . . . , result
  • value must be a long integer. It is converted to an integer by taking the low-order 32 bits. The result replaces value on the stack.
  • l2f
  • Long integer to single float
  • Syntax:
    l2f=137
  • Stack: . . . value-word1, value-word2=> . . . , result
  • value must be a long integer. It is converted to a single-precision floating point number. The result replaces value on the stack.
  • l2d
  • Long integer to double float
  • Syntax:
    l2d=138
  • Stack: . . . , value-word1, value-word2=> . . . result-word1, result-word2
  • value must be a long integer. It is converted to a double-precision floating point number. The result replaces value on the stack.
  • f2i
  • Single float to integer
  • Syntax:
    f2i=139
  • Stack: . . . , value => . . . , result
  • value must be a single-precision floating point number. It is converted to an integer. The result replaces value on the stack.
  • f2l
  • Single float to long integer
  • Syntax:
    f2l=140
  • Stack: . . . , value=> . . . , result-word1, result-word2
  • value must be a single-precision floating point number. It is converted to a long integer. The result replaces value on the stack.
  • f2d
  • Single float to double float
  • Syntax:
    f2d=141
  • Stack: . . . , value=> . . . , result-word1, result-word2
  • value must be a single-precision floating point number. It is converted to a double-precision floating point number. The result replaces value on the stack.
  • d2i
  • Double float to integer
  • Syntax:
    d2i=142
  • Stack: . . . , value-word1, value-word2=> . . . , result
  • value must be a double-precision floating point number. It is converted to an integer. The result replaces value on the stack.
  • d2l
  • Double float to long integer
  • Syntax:
    d2l=143
  • Stack: . . . , value-word1, value-word2=> . . . , result-word1, result-word2
  • value must be a double-precision floating point number. It is converted to a long integer. The result replaces value on the stack.
  • d2f
  • Double float to single float
  • Syntax:
    d2f=144
  • Stack: . . . , value-word1, value-word2=> . . . , result
  • value must be a double-precision floating point number. It is converted to a single-precision floating point number. If overflow occurs, the result must be infinity with the same sign as value. The result replaces value on the stack.
  • int2byte
  • Integer to signed byte
  • Syntax:
    int2byte=157
  • Stack: . . . , value=> . . . , result
  • value must be an integer. It is truncated to a signed 8-bit result, then sign extended to an integer. The result replaces value on the stack.
  • int2char
  • Integer to char
  • Syntax:
    int2char=146
  • Stack: . . . , value=> . . . , result
  • value must be an integer. It is truncated to an unsigned 16-bit result, then zero extended to an integer. The result replaces value on the stack.
  • int2short
  • Integer to short
  • Syntax:
    int2short=147
  • Stack: . . . , value=> . . . , result
  • value must be an integer. It is truncated to a signed 16-bit result, then sign extended to an integer. The result replaces value on the stack.
  • 3.11 Control Transfer Instructions
  • ifeq
  • Branch if equal to 0
  • Syntax:
    ifeq = 153
    branchbyte1
    branchbyte2
  • Stack: . . . , value=> . . .
  • value must be an integer. It is popped from the stack. If value is zero, branchbyte1 and branchbyte2 are used to construct a signed 16-bit offset. Execution proceeds at that offset from the address of this instruction. Otherwise execution proceeds at the instruction following the ifeq.
  • ifnull
  • Branch if null
  • Syntax:
    ifnull = 198
    branchbyte1
    branchbyte2
  • Stack: . . . , value=> . . .
  • value must be a reference to an object. It is popped from the stack. If value is null, branchbyte1 and branchbyte2 are used to construct a signed 16-bit offset. Execution proceeds at that offset from the address of this instruction. Otherwise execution proceeds at the instruction following the ifnull.
  • iflt
  • Branch if less than 0
  • Syntax:
    iflt = 155
    branchbyte1
    branchbyte2
  • Stack: . . . , value=> . . .
  • value must be an integer. It is popped from the stack. If value is less than zero, branchbyte1 and branchbyte2 are used to construct a signed 16-bit offset. Execution proceeds at that offset from the address of this instruction. Otherwise execution proceeds at the instruction following the iflt.
  • ifle
  • Branch if less than or equal to 0
  • Syntax:
    ifle=158
    branchbyte1
    branchbyte2
  • Stack: . . . , value=> . . .
  • value must be an integer. It is popped from the stack. If value is less than or equal to zero, branchbyte1 and branchbyte2 are used to construct a signed 16-bit offset. Execution proceeds at that offset from the address of this instruction. Otherwise execution proceeds at the instruction following the ifle.
  • ifne
  • Branch if not equal to 0
  • Syntax:
    ifne=154
    branchbyte1
    branchbyte2
  • Stack: . . . value=> . . .
  • value must be an integer. It is popped from the stack. If value is not equal to zero, branchbyte1 and branchbyte2 are used to construct a signed 16-bit offset. Execution proceeds at that offset from the address of this instruction. Otherwise execution proceeds at the instruction following the ifne.
  • ifnonnull
  • Branch if not null
  • Syntax:
    ifnonnull=199
    branchbyte1
    branchbyte2
  • Stack: . . . , value=> . . .
  • value must be a reference to an object. It is popped from the stack. If value is notnull, branchbyte1 and branchbyte2 are used to construct a signed 16-bit offset. Execution proceeds at that offset from the address of this instruction. Otherwise execution proceeds at the instruction following the ifnonnull.
  • ifgt
  • Branch if greater than 0
  • Syntax:
    ifft=157
    branchbyte1
    branchbyte2
  • Stack: . . . , value=> . . .
  • value must be an integer. It is popped from the stack. If value is greater than zero, branchbyte1 and branchbyte2 are used to construct a signed 16-bit offset. Execution proceeds at that offset from the address of this instruction. Otherwise execution proceeds at the instruction following the ifgt.
  • ifge
  • Branch if greater than or equal to 0
  • Syntax:
    ifge=156
    branchbyte1
    branchbyte2
  • Stack: . . . , value=> . . .
  • value must be an integer, It is popped from the stack. If value is greater than or equal to zero, branchbyte1 and branchbyte2 are used to construct a signed 16-bit offset. Execution proceeds at that offset from the address of this instruction. Otherwise execution proceeds at the instruction following instruction ifge.
  • if_icmpeq
  • Branch if integers equal
  • Syntax:
    if_icmpeq=159
    branchbyte1
    branchbyte2
  • Stack: . . . , value1, value2=> . . .
  • value1 and value2 must be integers. They are both popped from the stack. If value1 is equal to value2, branchbyte1 and branchbyte2 are used to construct a signed 16-bit offset. Execution proceeds, at that offset from the address of this instruction. Otherwise execution proceeds at the instruction following instruction if_icmpeq.
  • if_icmpne
  • Branch if integers not equal
  • Syntax:
    if'icmpne=160
    branchbyte1
    branchbyte2
  • Stack: . . . , value1, value2=> . . .
  • value1 and value2 must be integers. They are both popped from the stack. If value1 is not equal to value2, branchbyte1 and branchbyte2 are used to construct a signed 16-bit offset. Execution proceeds at that offset from the address of this instruction. Otherwise execution proceeds at the instruction following instruction if_icmpne.
  • if_icmplt
  • Branch if integer less than
  • Syntax:
    if_icmplt=161
    branchbyte1
    branchbyte2
  • Stack: . . . , value1, value2=> . . .
  • value1 and value2 must be integers. They are both popped from the stack. If value1 is less than value2, branchbyte1 and branchbyte2 are used to construct a signed 16-bit offset. Execution proceeds at that offset from the address of this instruction. Otherwise execution proceeds at the instruction following instruction if_icmplt.
  • if_icmpgt
  • Branch if integer greater than
  • Syntax:
    if_icmpgt=163
    branchbyte1
    branchbyte2
  • Stack: . . . , value1, value2=> . . .
  • value1 and value2 must be integers. They are both. popped from the stack. If value1 is greater than value2, branchbyte1 and branchbyte2 are used to construct a signed 16-bit offset. Execution proceeds at that offset from the address of this instruction. Otherwise execution proceeds at the instruction following instruction if_icmpgt.
  • if_icmple
  • Branch if integer less than or equal to
  • Syntax:
    if_icmple=164
    branchbyte1
    branchbyte2
  • Stack: . . . , value1, value2=> . . .
  • value1 and value2 must be integers. They are both popped from the stack. If value1 is less than or equal to value2, branchbyte1 and branchbyte2 are used to construct a signed 16-bit offset. Execution proceeds at that offset from the address of this instruction. Otherwise execution proceeds at the instruction following instruction if_icmple.
  • if_icmpge
  • Branch if integer greater than or equal to
  • Syntax:
    if_icmpge=162
    branchbyte1
    branchbyte2
  • Stack: . . . , value1, value2=> . . .
  • value1 and value2 must be integers. They are both popped from the stack. If value1 is greater than or equal to value2, branchbyte1 and branchbyte2 are used to construct a signed 16-bit offset Execution proceeds at that offset from the address of this instruction. Otherwise execution proceeds at the instruction following instruction if_icmpge.
  • lcmp
  • Long integer compare
  • Syntax:
  • Stack: . . . , value1-word1,
  • value1-word2,value2-word1, value2-word1=> . . . , result
  • value1 and value2 must be long integers. They are both popped from the stack and compared. If value1 is greater than value2, the integer value1 is pushed onto the stack. If value1 is equal to value2, the value 0 is pushed onto the stack. If value1 is less than value2, the value -1 is pushed onto the stack.
  • fcmpl
  • Single float compare (1 on NaN)
  • Syntax:
    fcmpl=149
  • Stack: . . . ; value1, value2=> . . . ,result
  • value1 and value2 must be single-precision floating point numbers. They are both popped from the stack and compared. If value1 is greater than value2, the integer value 1 is pushed onto the stack. If value1 is equal to value2, the value 0 is pushed onto the stack. If value1 is less than value2, the value −1 is pushed onto the stack.
  • If either value1 or value2 is NaN, the value −1 is pushed onto the stack.
  • fcmpg
  • Single float compare (1 on NaN)
  • Syntax:
    fcmpg=150
  • Stack: . . . ,value1 , value2=> . . . , result
  • value1 and value2 must be single-precision floating point numbers. They are both popped from the stack and compared. If value1 is greater than value2, the integer value 1 is pushed onto the stack. If value1 is equal to value2, the value 0 is pushed onto the stack. If value1 is less than value2, the value −1 is pushed onto the stack.
  • If either value1 or value2 is NaN, the value 1 is pushed onto the stack.
  • dcmpl
  • Double float compare (−1 on NaN)
  • Syntax:
    dcmpl=151
  • Stack: . . . , value1-word1, value1-word2, value2-word1, value2-word1=> . . . , result
  • value1 and value2 must be double-precision floating point numbers. They are both popped from the stack and compared. If value1 is greater than value2, the integer value 1 is pushed onto the stack. If value1 is equal to value2, the value 0 is pushed onto the stack. If value1 is less than value2, the value 1 is pushed onto the stacks
  • If either value1 or value2 is NaN, the value 1 is pushed onto the stack.
  • dcmpg
  • Double float compare (1 on NaN)
  • Syntax:
    dcmpg=152
  • Stack: . . . , value1-word1, value1-word2, value2-word1, value2-word1=> . . . , result
  • value1 and value2 must be double-precision floating point numbers. They are both popped from the stack and compared. If value1 is greater than value2, the integer value 1 is pushed onto the stack. If value1 is equal to value2, the value 0 is pushed onto the stack. If value1 is less than value2, the value −1 is pushed onto the stack.
  • If either value1 or value2 is NaN, the value 1 is pushed onto the stack.
  • if_acmpeq
  • Branch if object references are equal
  • Syntax:
    if_acmpeq=165
    branchbyte1
    branchbyte2
  • Stack: . . . ,value1 ,value2=> . . .
  • value1 and value2 must be references to objects. They are both popped from the stack. If the objects referenced are not the same, branchbyte1 and branchbyte2 are used to construct a signed 16-bit offset.
  • Execution proceeds at that offset from the Address of this instruction. Otherwise execution proceeds at the instruction following the if_acmpeq.
  • if_acmpne
  • Branch if object references not equal
  • Syntax:
    if_acmpne=166
    branchbyte1
    branchbyte2
  • Stack: . . . , value1 , value2=>. . .
  • value1 and value2 must be references to objects. They are both popped from the stack. If the objects referenced are not the same, branchbyte1 and branchbyte2 are used to construct a signed 16-bit offset.
  • Execution proceeds at that offset from the address of this instruction. Otherwise execution proceeds at the instruction following instruction if_acmpne.
  • goto
  • Branch always
  • Syntax:
    goto=167
    branchbyte1
    branchbyte2
  • Stack: no change
  • branchbyte1 and branchbyte2 are used to construct a signed 16-bit offset. Execution proceeds at that offset from the address of this instruction.
  • goto_w
  • Branch always (wide index)
  • Syntax:
    goto_w=200
    branchbyte1
    branchbyte2
    branchbyte3
    branchbyte4
  • Stack: no change
  • branchbyte1, branchbyte2, branchbyte3, and branchbyte4 are used to construct a signed 32-bit offset.
  • Execution proceeds at that offset from the address of this instruction.
  • jsr
  • Jump subroutine
  • Syntax:
    jsr=168
    branchbyte1
    branchbyte2
  • Stack: . . . => . . . , return-address
  • branchbyte1 and branchbyte2 are used to construct a signed 16-bit offset. The address of the instruction immediately following the jar is pushed onto the stack. Execution proceeds at the offset from the address of this instruction.
  • jsr_w
  • Jump subroutine (wide index)
  • Syntax:
    jsr_w=201
    branchbyte1
    branchbyte2
    branchbyte3
    branchbyte4
  • Stack: . . . => . . . , return-address
  • branchbyte1, branchbyte2, branchbyte3, and branchbyte4 are used to construct a signed 32-bit offset. The address of the instruction immediately following the jsr_w is pushed onto the stack. Execution proceeds at the offset from the address of this instruction.
  • ret
  • Return from subroutine
  • Syntax:
    ret=169
    vindex
  • Stack: no change
  • Local variable vindex in the current JAVA frame must contain a return address. The contents of the local variable are written into the pc.
  • Note that jsr pushes the address onto the stack, and ret gets it out of a local variable. This asymmetry is intentional.
  • ret_w
  • Return from subroutine (wide index)
  • Syntax:
    ret_w=209
    vindexbyte1
    vindexbyte2
  • Stack: no change
  • vindexbyte1 and vindexbyte2 are assembled into an unsigned 16-bit index to a local variable in the current JAVA frame. That local variable must contain a return address. The contents of the local variable are written into the pc. See the ret instruction for more information.
  • 3.12 Function Return
  • ireturn
  • Return integer from function
  • Syntax:
    ireturn=172
  • Stack: . . . , value=> [empty]
  • value must be an integer. The value value is pushed onto the stack of the previous execution environment. Any other values on the operand stack are discarded. The interpreter then returns control to its caller.
  • lreturn
  • Return long integer from function
  • Syntax:
    lreturn=173
  • Stack: . . . , value-word1, value-word2=> [empty]
  • value must be a long integer. The value value is pushed onto the stack of the previous execution environment. Any other values on the operand stack are discarded. The interpreter then returns control to its caller.
  • freturn
  • Return single float from function
  • Syntax:
    freturn=174
  • Stack: . . . , value=> [empty]
  • value must be a single-precision floating point number. The value value is pushed onto the stack of the previous execution environment. Any other values on the operand stack are discarded. The interpreter then returns control to its caller.
  • dreturn
  • Return double float from function
  • Syntax:
    dreturn=175
  • Stack: . . . , value-word1, value-word2=> [empty]
  • value must be a double-precision floating point number. The value value is pushed onto the stack of the previous execution environment. Any other values on the operand stack are discarded. The interpreter. then returns control to its caller.
  • areturn
  • Return object reference from function
  • Syntax:
    areturn=176
  • Stack: . . . , value=> [empty]
  • value must be a reference to an object. The value value is pushed onto the stack of the previous execution environment. Any other values on the operand stack are discarded. The interpreter then returns control to its caller.
  • return
  • Return (void) from procedure
  • Syntax:
    return=177
  • Stack: . . . => [empty]
  • All values on the operand stack are discarded. The interpreter then returns control to its caller.
  • breakpoint
  • Stop and pass control to breakpoint handler
  • Syntax:
    breakpoint=202
  • Stack: no change
  • 3.13 Table Jumping
  • tableswitch
  • Access jump table by index and jump
  • Syntax:
    tableswitch=170
    ...0-3 byte pad...
    default-offset1
    default-offset2
    default-offset3
    default-offset4
    low1
    low2
    low3
    low4
    high1
    high2
    high3
    high4
    ...jump offsets...
  • Stack: . . . , index=> . . .
  • tableswitch is a variable length instruction. Immediately after the tableswitch opcode, between zero and three 0's are inserted as padding so that the next byte begins at an address that is a multiple of four. After the padding follow a series of signed 4-byte quantities: default-offset, low, high, and then high-low+1 further signed 4-byte offsets. The high-low+1 signed 4-byte offsets are treated as a 0-based jump table.
  • The index must be an integer. If index is less than low or index is greater than high, then default-offset is added to the address of this instruction. Otherwise, low is subtracted from index, and the index-low'th element of the jump table is extracted, and added to the address of this instruction.
  • lookupswitch
  • Access jump table by key match and jump
  • Syntax:
    lookupswitch=171
    ...0-3 byte pad..
    default-offset1
    default-offset2
    default-offset3
    default-offset4
    npairs1
    npairs2
    npairs3
    npairs4
    ...match-offset pairs...
  • Stack: . . . , key=> . . .
  • lookupswitch is a variable length instruction. Immediately after the lookupswitch opcode, between zero and three 0's are inserted as padding so that the next byte begins at an address that is a multiple of four.
  • Immediately after the padding are a series of pairs of signed 4-byte quantities. The first pair is special. The first item of that pair is the default offset, and the second item of that pair gives the number of pairs that follow. Each subsequent pair consists of a match and an offset.
  • The key must be an integer. The integer key on the stack is compared against each of the matches. If it is equal to one of them, the offset is added to the address of this instruction. If the key does not match any of the matches, the default offset is added to the address of this instruction.
  • 3.14 Manipulating Object Fields
  • putfield
  • Set field in object
  • Syntax:
    putfield=181
    indexbyte1
    indexbyte2
  • Stack: . . . , objectref, value=> . . .
      • OR
        Stack: . . . , objectref, value-word1, value-word2=> . . .
  • indexbyte1 and indexbyte2 are used to construct an index into the constant pool of the current class. The constant pool item will be a field reference to a class name and a field name. The item is resolved to a field block pointer which has both the field width (in bytes) and the field offset (in bytes).
  • The field at that offset from the start of the object referenced by object refwill be set to the value on the top of the stack.
  • This instruction deals with both 32-bit and 64-bit wide fields.
  • If object ref is null, aNullPointerException is generated.
  • If the specified field is a static field, anIncompatibleClassChangeError is thrown.
  • getfield
  • Fetch field from object
  • Syntax:
    getfield=180
    indexbyte1
    indexbyte2
  • Stack: . . . , objectref=> . . . ,value
      • OR
        Stack: . . . , objectref=> . . . , value-word1, value-word2
  • indexbyte1 and indexbyte2 are used to construct an index into the constant pool of the current class. The constant pool item will be a field reference to a class name and a field name. The item is resolved to a field block pointer which has both the field width (in bytes) and the field offset (in bytes).
  • objectref must be a reference to an object. The value at offset into the object referenced by objectref replaces objectref on the top of the stack.
  • This instruction deals with both 32-bit and 64-bit wide fields.
  • If objectref is null, a NullPointerException is generated.
  • If the specified field is a static field, an IncompatibleClassChangeError is thrown.
  • putstatic
  • Set static field in class
  • Syntax:
    putstatic=179
    indexbyte1
    indexbyte2

    Stack: . . . , value=> . . .
      • OR
  • Stack: . . . , value-word1, value-word2=> . . .
  • indexbyte1 and indexbyte2 are used to construct an index into the constant pool of the current class. The constant pool item will be a field reference to a static field of a class. That field will be set to have the value on the top of the stack.
  • This instruction works for both 32-bit and 64-bit wide fields.
  • If the specified field is a dynamic field, an IncompatibleClassChangeError is thrown.
  • getstatic
  • Get static field from class
  • Syntax:
    getstatic=178
    indexbyte1
    indexbyte2

    Stack: . . . , => . . . , value
      • OR
        Stack: . . . , => . . . value-word1, value-word2
  • indexbyte1 and indexbyte2 are used to construct an index into the constant pool of the current class. The constant pool item will be a field reference to a static field of a class.
  • This instruction deals with both 32-bit and 64-bit wide fields.
  • If the specified field is a dynamic field, an IncompatibleClassChangeError is generated.
  • 3.15 Method Invocation
  • There are four instructions that implement method invocation.
    • invokevirtual Invoke an instance method of an object, dispatching based on the runtime (virtual) type of the object. This is the normal method dispatch in JAVA.
    • invokenonvirtual Invoke an instance method of an object, dispatching based on the compile-time (non-virtual) type of the object. This is used, for example, when the keywordsuper or the name of a superclass is used as a method qualifier.
    • invokestatic Invoke a class (static) method in a named class.
    • invokeinterface Invoke a method which is implemented by an interface, searching the methods implemented by the particular run-time object to find the appropriate method.
      invokevirtual
  • Invoke instance method, dispatch based on run-time type
  • Syntax:
    invokevirtual=182
    indexbyte1
    indexbyte2
  • Stack: . . . objectref, [arg1, [arg2 . . . ]], . . . => . . .
  • The operand stack must contain a reference to an object and some number of arguments.indexbyte1 and indexbyte2 are used to construct an index into the constant pool of the current class. The item at that index in the constant pool contains the complete method signature. A pointer to the object's method table is retrieved from the object reference. The method signature is looked up in the method table. The method signature is guaranteed to exactly match one of the method signatures in the table.
  • The result of the lookup is an index into the method table of the named class, which is used with the object's dynamic type to look in the method table of that type, where a pointer to the method block for the matched method is found. The method block indicates the type of method (native, synchronized, and so on) and the number of arguments expected on the operand stack.
  • If the method is marked synchronized the monitor associated with objectref is entered.
  • The objectref and arguments are popped off this method's stack and become the initial values of the local variables of the new method. Execution continues with the first instruction of the new method.
  • If the object reference on the operand stack is null, a NullPointerException is thrown. If during the method invocation a stack overflow is detected, a StackOverflowError is thrown.
  • invokenonvirtual
  • Invoke instance method, dispatching based on compile-time type
  • Syntax:
    invokenonvirtual = 183
    indexbyte1
    indexbyte2
  • Stack: . . . , objectref, [arg1, [arg2 . . . ]], . . . => . . .
  • The operand stack must contain a reference to an object and some number of arguments.indexbyte1 and indexbyte2 are used to construct an index into the constant pool of the current class. The item at that index in the constant pool contains a complete method signature and class. The method signature is looked up in the method table of the class indicated. The method signature is guaranteed to exactly match one of the method signatures in the table.
  • The result of the lookup is a method block. The method block indicates the type of method (native, synchronized, and so on) and the number of arguments (nargs) expected on the operand stack.
  • If the method is marked synchronized the monitor associated with objectref is entered.
  • The objectref and arguments are popped off this method's stack and become the initial values of the local variables of the new method. Execution continues with the first instruction of the new method.
  • If the object reference on the operand stack is null, a NullPointerException is thrown. If during the method invocation a stack overflow is detected, a StackOverflowError is thrown.
  • invokestatic
  • Invoke a class (static) method
  • Syntax:
    invokestatis = 184
    indexbyte1
    indexbyte2
  • Stack: . . . , [arg1, [arg2 . . .]]. . . => . . .
  • The operand stack must contain some number of arguments.indexbyte1 and indexbyte2 are used to construct an index into the constant pool of the current class. The item at that index in the constant pool contains the complete method signature and class. The method signature is looked up in the method table of the class indicated. The method signature is guaranteed to exactly match one of the method signatures in the class's method table.
  • The result of the lookup is a method block. The method block indicates the type of method (native, synchronized, and so on) and the number of arguments (nargs) expected on the operand stack.
  • If the method is marked synchronized the monitor associated with the class is entered.
  • The arguments are popped off this method's stack and become the initial values of the local variables of the new method. Execution continues with the first instruction of the new method.
  • If during the method invocation a stack overflow is detected, a StackOverflowError is thrown.
  • invokeinterface
  • Invoke interface method
  • Syntax:
    invokeinterface = 185
    indexbyte1
    indexbyte2
    nargs
    reserved

    Stack: . . . , objectref, [arg1, [arg2 . . . ]], . . . => . . .
  • The operand stack must contain a reference to an object and nargs-1 arguments. indexbyte1 and indexbyte2 are used to construct an index into the constant pool of the current class. The item at that index in the constant pool contains the complete method signature. A pointer to the object's method table is retrieved from the object reference. The method signature is looked up in the method table. The method signature is guaranteed to exactly match one of the method signatures in the table.
  • The result of the lookup is a method block. The method block indicates the type of method (native, synchronized, and so on) but unlike invokevirtual and invokenonvirtual, the number of available arguments (nargs) is taken from the bytecode.
  • If the method is markedsynchronized the monitor associated with objectref is entered.
  • The objectref and arguments are popped off this method's stack and become the initial values of the local variables of the new method. Execution continues with the first instruction of the new method.
  • If the objectref on the operand stack is null, a NullPointerException is thrown. If during the method invocation a stack overflow is detected, a StackOverflowError is thrown.
  • 3.16 Exception Handling
  • athrow
  • Throw exception or error
  • Syntax:
    athrow=191
  • Stack: . . . , objectref=> [undefined]
  • objectref must be a reference to an object which is a subclass of Throwable, which is thrown. The current JAVA stack frame is searched for the most recent catch clause that catches this class or a superclass of this class. If a matching catch list entry is found, the pc is reset to the address indicated by the catch-list entry, and execution continues there.
  • If no appropriate catch clause is found in the current stack frame, that frame is popped and the object is rethrown. If one is found, it contains the location of the code for this exception. The pc is reset to that location and execution continues. If no appropriate catch is found in the current stack frame, that frame is popped and the objectref is rethrown.
  • If objectref is null, then a NullPointerException is thrown instead.
  • 3.17 Miscellaneous Object Operations
  • new
  • Create new object
  • Syntax:
    new = 187
    indexbyte1
    indexbyte2
  • Stack: . . . => . . . , objectref
  • indexbyte1 and indexbyte2 are used to construct an index into the constant pool of the current class. The item at that index must be a class name that can be resolved to a class pointer, class. A new instance of that class is then created and a reference to the object is pushed on the stack.
  • checkcast
  • Make sure object is of given type
  • Syntax:
    checkcast = 192
    indexbyte1
    indexbyte2
  • Stack: . . . , objectref=> . . . , objectref
  • indexbyte1 and indexbyte2 are used to construct an index into the constant pool of the current class. The string at that index of the constant pool is presumed to be a class name which can be resolved to a class pointer, class. objectref must be a reference to an object.
  • checkcast determines whether objectref can be cast to be a reference to an object of class class. A null objectref can be cast to any class. Otherwise the referenced object must be an instance of class or one of its superclasses. If objectref can be cast to class execution proceeds at the next instruction, and the objectref remains on the stack.
  • If objectref cannot be cast to class, a ClassCastException is thrown.
  • instanceof
  • Determine if an object is of given type
  • Syntax:
    instanceof = 193
    indexbyte1
    indexbyte2
  • Stack: . . . , objectref=> . . . , result
  • indexbyte1 and indexbyte2 are used to construct an index into the constant pool of the current class. The string at that index of the constant pool is presumed to be a class name which can be resolved to a class pointer, class. objectref must be a reference to an object.
  • instanceof determines whether objectref can be cast to be a reference to an object of the class class. This instruction will overwrite objectref with 1 if objectref is an instance of class or one of its superclasses. Otherwise, objectref is overwritten by 0. If objectref is null, it's overwritten by 0.
  • 3.18 Monitors
  • monitorenter
  • Enter monitored region of code
  • Syntax:
    monitorenter=194
  • Stack: . . . , objectref=> . . .
  • objectref must be a reference to an object,
  • The interpreter attempts to obtain exclusive access via a lock mechanism to objectref. If another thread already has objectref locked, than the current thread waits until the object is unlocked. If the current thread already has the object locked, then continue execution. If the object is not locked, then obtain an exclusive lock.
  • If objectref is null, then a NullPointerException is thrown instead.
  • monitorexit
  • Exit monitored region of code
  • Syntax:
    monitorexit=195
  • Stack: . . . , objectref=> . . .
  • objectref must be a reference to an object. The lock on the object released. If this is the last lock that this thread has on that object (one thread is allowed to have multiple locks on a single object), then other threads that are waiting for the object to be available are allowed to proceed.
  • If objectref is null, then a NullPointerException is thrown instead.
  • Appendix A: An Optimization
  • The following set of pseudo-instructions suffixed by _quick are variants of JAVA virtual machine instructions. They are used to improve the speed of interpreting bytecodes. They are not part of the virtual machine specification or instruction set, and are invisible outside of an JAVA virtual machine implementation. However, inside a virtual machine implementation they have proven to be an effective optimization.
  • A compiler from JAVA source code to the JAVA virtual machine instruction set emits only non-_quick instructions. If the _quick pseudo-instructions are used, each instance of a non-_quick instruction with a _quick variant is overwritten on execution by its_quick variant. Subsequent execution of that instruction instance will be of the_quick variant.
  • In all cases, if an instruction has an alternative version with the suffix_quick, the instruction references the constant pool. If the_quick optimization is used, each non-_quick instruction with a_quick variant performs the following:
      • Resolves the specified item in the constant pool;
      • Signals an error if the item in the constant pool could not be resolved for some reason;
      • Turns itself into the _quick version of the instruction. The instructions putstatic, getstatic, putfield, and getfield each have two_quick versions; and
      • Performs its intended operation,
  • This is identical to the action of the instruction without the _quick optimization, except for the additional step in which the instruction overwrites itself with its _quick variant.
  • The _quick variant of an instruction assumes that the item in the constant pool has already been resolved, and that this resolution did not generate any errors. It simply performs the intended operation on the resolved item.
  • Note: some of the invoke methods only support a single-byte offset into the method table of the object; for objects with 256 or more methods some invocations cannot be “quicked” with only these bytecodes.
  • This Appendix doesn't give the opcode values of the pseudo-instructions, since they are invisible and subject to change.
  • A.1 Constant Pool Resolution
  • When the class is read in, an array constant_pool [ ] of size n constants is created and assigned to a field in the class.constant_pool [0] is set to point to a dynamically allocated array which indicates which fields in the constant_pool have already been resolved.constant_pool [1] through constant pool [nconstants−1] are set to point at the “type” field that corresponds to this constant item.
  • When an instruction is executed that references the constant pool, an index is generated, and constant_pool[0] is checked to see if the index has already been resolved. If so, the value of constant_pool [index] is returned. If not, the value of constant_pool [index] is resolved to be the actual pointer or data, and overwrites whatever value was already in constant_pool [index].
  • A.2 Pushing Constants onto the Stack (_quick variants)
  • ldcl_quick
  • Push item from constant pool onto stack
  • Syntax:
    ldc1_quick
    indexbyte1
  • Stack: . . . => . . . ,item
  • indexbyte1 is used as an unsigned 8-bit index into the constant pool of the current class. The item at that index is pushed onto the stack.
  • ldc2_quick
  • Push item from constant pool onto stack
  • Syntax:
    ldc2_quick
    indexbyte1
    indexbyte2
  • Stack: . . . => . . . , item
  • indexbyte1 and indexbyte2 are used to construct an index into the constant pool of the current class. The constant at that index is resolved and the item at that index is pushed onto the stack.
  • ldc2w_quick
  • Push long integer or double float from constant pool onto stack
  • Syntax:
    ldc2w_quick
    indexbyte1
    indexbyte2
  • Stack: . . . => . . . ,constant-word1,constant-word2
  • indexbyte1 and indexbyte2 are used to construct an index into the constant pool of the current class. The constant at that index is pushed onto the stack.
  • A.3 Managing Arrays (_quick variants)
  • anewarray_quick
  • Allocate new array of references to objects
  • Syntax:
    anewarray_quick
    indexbyte1
    indexbyte2
  • Stack: . . . ,size=>result
  • size must be an integer. It represents the number of elements in the new array.
  • indexbyte1 and indexbyte2 are used to construct an index into the constant pool of the current class. The entry must be a class.
  • A new array of the indicated class type and capable of holding size elements is allocated, and result is a reference to this new array. Allocation of an array large enough to contain size items of the given class type is attempted. All elements of the array are initialized to zero.
  • If size is less than zero, a NegativeArraySizeException is thrown. If there is not enough memory to allocate the; array, an OutOfMemoryError is thrown.
  • multianewarray_quick
  • Allocate new multi-dimensional array
  • Syntax:
    multianewarray_quick
    indexbyte1
    indexbyte2
    dimensions
  • Stack: . . . ,size1,size2, . . . sizen=>result
  • Each size must be an integer. Each represents the number of elements in a dimension of the array.
  • indexbyte1 and indexbyte2 are used to construct an index into the constant pool of the current class. The resulting entry must be a class.
  • dimensions has the following aspects:
  • It must be an integer ≧1.
      • It represents the number of dimensions being created. It must be ≦ the number of dimensions of the array class.
      • It represents the number of elements that are popped off the stack. All must be integers greater than or equal to zero. These are used as the sizes of the dimension.
  • If any of the size arguments on the stack is less than zero, a NegativeArraySizeException is thrown. If there is not enough memory to allocate the array, an OutOfMemoryError is thrown.
  • The result is a reference to the new array object.
  • A.4 Manipulating Object Fields (_quick variants)
  • putfield_quick
  • Set field in object
  • Syntax:
    putfield2_quick
    offset
    unused
  • Stack: . . . ,objectref,value=> . . .
  • objectref must be a reference to an object. value must be a value of a type appropriate for the specified field. offset is the offset for the field in that object. value is written at offset into the object. Both objectref and value are popped from the stack.
  • If objectref is null, a NullPointerException is generated.
  • putfield2_quick
  • Set long integer or double float field in object
  • Syntax:
    putfield2_quick
    offset
    unused
  • Stack: . . . ,objectref,value-word1,value-word2=> . . .
  • objectref must be a reference to an object. value must be a value of a type appropriate for the specified field. offset is the offset for the field in that object. value is written at offset into the object. Both objectref and value are popped from the stack.
  • If objectref is null, a NullPointerException is generated.
  • getfield_quick
  • Fetch field from object
  • Syntax:
    getfield2_quick
    offset
    unused
  • Stack: . . . ,objectref=> . . . ,value
  • objectref must be a handle to an object. The value at offset into the object referenced by objectref replaces objectref on the top of the stack.
  • If objectref is null, a NullPointerException is generated.
  • getfield2_quick
  • Fetch field from object
  • Syntax:
    getfield2_quick
    offset
    unused
  • Stack: . . . ,objectref=> . . . ,value-word1,value-word2
  • objectref must be a handle to an object. The value at offset into the object referenced by objectref replaces objectref on the top of the stack.
  • If objectref is null, a NullPointerException is generated.
  • putstatic_quick
  • Set static field in class
  • Syntax:
    putstatic_quick
    indexbyte1
    indexbyte2
  • Stack: . . . ,value=> . . .
  • indexbyte1 and indexbyte2 are used to construct an index into the constant pool of the current class. The constant pool item will be a field reference to a static field of a class.value must be the type appropriate to that field. That field will be set to have the value value.
  • putstatic2_quick
  • Set static field in class
  • Syntax:
    putstatic2_quick
    indexbyte1
    indexbyte2
  • Stack: . . . ,value-word1,value-word2=> . . .
  • indexbyte1 and indexbyte2 are used to construct an index into the constant pool of the current class. The constant pool item will be a field reference to a static field of a class. That field must either be a long integer or a double precision floating point number. value must be the type appropriate to that field. That field will be set to have the value value.
  • getstatic_quick
  • Get static field from class
  • Syntax:
    getstatic_quick
    indexbyte1
    indexbyte2
  • Stack: . . . ,=> . . . ,value
  • indexbyte1 and indexbyte2 are used to construct an index into the constant pool of the current class. The constant pool item will be a field reference to a static field of a class. The value of that field will replace handle on the stack.
  • getstatic2_quick
  • Get static field from class
  • Syntax:
    getstatic2_quick
    indexbyte1
    indexbyte2
  • Stack: . . . ,=> . . . ,value-word1,value-word2
  • indexbyte1 and indexbyte2 are used to construct an index into the constant pool of the current class. The constant pool item will be a field reference to a static field of a class. The field must be a long integer or a double precision floating point number. The value of that field will replace handle on the stack
  • A.5 Method Invocation (_quick variants)
  • invokevirtual_quick
      • Invoke instance method, dispatching based on run-time type
  • Syntax:
    invokevirtual_quick
    offset
    nargs
      • Stack: . . . ,objectref,[arg1,[arg2 . . . ]]=> . . .
  • The operand stack must contain objectref, a reference to an object and nargs-1 arguments. The method block at offset in the object's method table, as determined by the object's dynamic type, is retrieved. The method block indicates the type of method (native, synchronized, etc.).
  • If the method is marked synchronized the monitor associated with the object is entered.
  • The base of the local variables array for the new JAVA stack frame is set to point to objectref on the stack, making objectref and the supplied arguments (arg1,arg2, . . . ) the first nargs local variables of the new frame. The total number of local variables used by the method is determined, and the execution environment of the new frame is pushed after leaving sufficient room for the locals. The base of the operand stack for this method invocation is set to the first word after the execution environment. Finally, execution continues with the first instruction of the matched method.
  • If objectref is null, a NullPointerException is thrown. If during the method invocation a stack overflow is detected, a StackOverflowError is thrown.
  • invokevirtualobject_quick
  • Invoke instance method of class JAVA.lang.Object, specifically for benefit of arrays
  • Syntax:
    invokevirtualobject_quick
    offset
    nargs
  • Stack: . . . ,objectref, [arg1, [arg2 . . . ]]=> . . .
  • The operand stack must contain objectref, a reference to an object or to an array and nargs-1 arguments. The method block at offset in JAVA.lang.Object's method table is retrieved. The method block indicates the type of method (native, synchronized, etc.).
  • If the method is marked synchronized the monitor associated with handle is entered.
  • The base of the local variables array for the new JAVA stack frame is set to point to objectref on the stack, making objectref and the supplied arguments (arg1,arg2, . . . ) the first nargs local variables of the new frame. The total number of local variables used by the method is determined, and the execution environment of the new frame is pushed after leaving sufficient room for the locals. The base of the operand stack for this method invocation is set to the first word after the execution environment. Finally, execution continues with the first instruction of the matched method.
  • If objectref is null, a NullPointerException is thrown. If during the method invocation a stack overflow is detected, a StackOverflowError is thrown.
  • invokenonvirtual_quick
  • Invoke instance method, dispatching based on compile-time type
  • Syntax:
    invokenonvirtual_quick
    indexbyte1
    indexbyte2
  • Stack: . . . ,objectref,[arg1, [arg2 . . . ]]=> . . .
  • The operand stack must contain objectref, a reference to an object and some number of arguments. indexbyte1 and indexbyte2 are used to construct an index into the constant pool of the current class. The item at that index in the constant pool contains a method slot index and a pointer to a class. The method block at the method slot index in the indicated class is retrieved. The method block indicates the type of method (native, synchronized, etc.) and the number of arguments (nargs) expected on the operand stack.
  • If the method is marked synchronized the monitor associated with the object is entered.
  • The base of the local variables array for the new JAVA stack frame is set to point to objectref on the stack, making objectref and the supplied arguments (arg1, arg2, . . . ) the first nargs local variables of the new frame. The total number of local variables used by the method is determined, and the execution environment of the new frame is pushed after leaving sufficient room for the locals. The base of the operand stack for this method invocation is set to the first word after the execution environment. Finally, execution continues with the first instruction of the matched method.
  • If objectref is null, a NullPointerException is thrown. If during the method invocation a stack overflow is detected, a StackOverflowError is thrown.
  • invokestatic_quick
  • Invoke a class (static) method
  • Syntax:
    invokestatic_quick
    indexbyte1
    indexbyte2

    Stack: . . . , [arg1, [arg2 . . . ]]=> . . .
  • The operand stack must contain some number of arguments. indexbyte1 and indexbyte2 are used to construct an index into the constant pool of the current class. The item at that index in the constant pool contains a method slot index and a pointer to a class. The method block at the method slot index in the indicated class is retrieved. The method block indicates the type of method (native, synchronized, etc.) and the number of arguments (nargs) expected on the operand stack.
  • If the method is marked synchronized the monitor associated with the method's class is entered.
  • The base of the local variables array for the new JAVA stack frame is set to point to the first argument on the stack, making the supplied arguments (arg1,arg2, . . . ) the first nargs local variables of the new frame. The total number of local variables used by the method is determined, and the execution environment of the new frame is pushed after leaving sufficient room for the locals. The base of the operand stack for this method invocation is set to the first word after the execution environment. Finally, execution continues with the first instruction of the matched method.
  • If the object handle on the operand stack is null, a NullPointerException is thrown. If during the method invocation a stack overflow is detected, a StackOverflowError is thrown.
  • invokeinterface_quick
  • Invoke interface method
  • Syntax:
    invokeinterface_quick
    idbyte1
    idbyte2
    nargs
    guess
  • Stack: . . . ,objectref,[arg1,[arg2 . . . ]]=> . . .
  • The operand stack must contain objectref, a reference to an object, and nargs-1 arguments. idbyte1 and idbyte2 are used to construct an index into the constant pool of the current class. The item at that index in the constant pool contains the complete method signature. A pointer to the object's method table is retrieved from the object handle.
  • The method signature is searched for in the object's method table. As a short-cut, the method signature at slot guess is searched first. If that fails, a complete search of the method table is performed. The method signature is guaranteed to exactly match one of the method signatures in the table.
  • The result of the lookup is a method block. The method block indicates the type of method (native, synchronized, etc.) but the number of available arguments (nargs) is taken from the bytecode.
  • If the method is marked synchronized the monitor associated with handle is entered.
  • The base of the local variables array for the new JAVA stack frame is set to point to handle on the stack, making handle and the supplied arguments (arg1,arg2, . . . ) the first nargs local variables of the new frame. The total number of local variables used by the method is determined, and the execution environment of the new frame is pushed after leaving sufficient room for the locals. The base of the operand stack for this method invocation is set to the first word after the execution environment. Finally, execution continues with the first instruction of the matched method.
  • If objectref is null, a NullPointerException is thrown. If during the method invocation a stack overflow is detected, a StackOverflowError is thrown.
  • guess is the last guess. Each time through, guess is set to the method offset that was used.
  • A.6 Miscellaneous Object Operations (_quick variants)
  • new_quick
  • Create new object
  • Syntax:
    new_quick
    indexbyte1
    indexbyte2
  • Stack: . . . => . . . ,objectref
  • indexbyte1 and indexbyte2 are used to construct an index into the constant pool of the current class. The item at that index must be a class. A new instance of that class is then created and objectref, a reference to that object is pushed on the stack.
  • checkcast_quick
  • Make sure object is of given type
  • Syntax:
    checkcast_quick
    indexbyte1
    indexbyte2
  • Stack: . . . ,objectref=> . . . ,objectref
  • objectref must be a reference to an object. indexbyte1 and indexbyte2 are used to construct an index into the constant pool of the current class. The object at that index of the constant pool must have already been resolved.
  • checkcast then determines whether objectref can be cast to a reference to an object of class class. A null reference can be cast to any class, and otherwise the superclasses of objectref's type are searched for class. If class is determined to be a superclass of objectref's type, or if objectref is null, it can be cast to objectref cannot be cast to class, a ClassCastException is thrown.
  • instanceof_quick
  • Determine if object is of given type
  • Syntax:
    instanceof_quick
    indexbyte1
    indexbyte2
  • Stack: . . . ,objectref=> . . . ,result
  • objectref must be a reference to an object. indexbyte1 and indexbyte2 are used to construct an index into the constant pool of the current class. The item of class class at that index of the constant pool must have already been resolved.
  • Instance of determines whether objectref can be cast to an object of the class class. A null objectref can be cast to any class and otherwise the superclasses of objectref's type are searched for class. If class is determined to be a superclass of objectref's type, result is 1 (true). Otherwise, result is 0 (false). If handle is null, result is 0 (false).

Claims (20)

1-56. (canceled)
57. An apparatus comprising:
an execution unit operable to manipulate data residing in register storage; and
an instruction decoder operable to decode plural successive stack-based instructions and to cause the execution unit to perform a single, register-based operation that corresponds to the plural decoded stack-based instructions.
58. The apparatus of claim 57,
wherein the single, register-based operation explicitly identifies a result location in the register storage and at least two source locations in the register storage.
59. The apparatus of claim 57,
wherein the instruction decoder is further operable to select plural entries of the register storage, the selected register storage entries corresponding to a subset of stack and local variable storage locations explicitly and implicitly targeted by the plural decoded stack-based instructions.
60. The apparatus of claim 57, further comprising:
stack and local variable portions of the register storage, wherein at least one of the plural decoded stack-based instructions defines an information transfer from a local variable to an top position of the stack.
61. The apparatus of claim 57,
wherein the instruction decoder is part of an instruction folding just-in-time (JIT) compiler.
62. The apparatus of claim 57,
wherein the instruction decoder is part of an instruction folding bytecode interpreter that implements a virtual machine.
63. The apparatus of claim 57,
wherein the instruction decoder includes instruction folding hardware.
64. The apparatus of claim 57, further comprising:
a just-in-time (JIT) compiler executable to produce object code native to the execution unit.
65. A method comprising:
fetching a sequence of stack-based instructions;
decoding plural successive stack-based instructions; and
causing an execution unit to perform a single, register-based operation that corresponds to the plural decoded stack-based instructions.
66. The method of claim 65,
wherein the decoding is performed in a just-in-time (JIT) compiler to produce object code native to a processor that includes the execution unit.
67. The method of claim 65,
wherein the decoding is performed using an instruction folding just-in-time (JIT) compiler.
68 . The method of claim 65,
wherein the decoding is performed using an instruction folding bytecode interpreter.
69. The method of claim 65,
wherein the decoding is performed using instruction folding hardware.
70. The method of claim 65, further comprising:
selecting plural entries of register storage, the selected register storage entries corresponding to a subset of stack and local variable storage locations explicitly and implicitly targeted by the plural decoded stack-based instructions.
71. A computational system comprising:
an instruction source;
register storage;
a processor; and
means for receiving stack-based instructions from the instruction source and for causing the processor to operate upon data residing in the register storage using explicit identifiers for source and result locations in the register storage, the means translating plural successive ones of the stack-based instructions into a single corresponding register-based operation.
72. The computational system of claim 71,
wherein the means for translating includes means for compiling the stack-based instructions to produce object code native to the processor.
73. The computational system of claim 71,
wherein the means for translating includes means for folding instructions using a just-in-time (JIT) compiler.
74. The computational system of claim 71,
wherein the means for translating includes means for folding instructions using a bytecode interpreter.
75. The computational system of claim 71,
wherein the means for translating includes means for folding instructions using a hardware decoder.
US11/461,340 1996-01-24 2006-07-31 Instruction folding for a stack-based machine Abandoned US20070277021A1 (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
US11/461,340 US20070277021A1 (en) 1996-01-24 2006-07-31 Instruction folding for a stack-based machine

Applications Claiming Priority (7)

Application Number Priority Date Filing Date Title
US1052796P 1996-01-24 1996-01-24
US64225396A 1996-05-02 1996-05-02
US64710396A 1996-05-07 1996-05-07
US08/787,617 US6532531B1 (en) 1996-01-24 1997-01-23 Method frame storage using multiple memory circuits
US10/346,886 US6950923B2 (en) 1996-01-24 2003-01-17 Method frame storage using multiple memory circuits
US11/096,183 US20050267996A1 (en) 1996-01-24 2005-03-30 Method frame storage using multiple memory circuits
US11/461,340 US20070277021A1 (en) 1996-01-24 2006-07-31 Instruction folding for a stack-based machine

Related Parent Applications (1)

Application Number Title Priority Date Filing Date
US11/096,183 Continuation US20050267996A1 (en) 1996-01-24 2005-03-30 Method frame storage using multiple memory circuits

Publications (1)

Publication Number Publication Date
US20070277021A1 true US20070277021A1 (en) 2007-11-29

Family

ID=27359254

Family Applications (5)

Application Number Title Priority Date Filing Date
US08/787,617 Expired - Lifetime US6532531B1 (en) 1996-01-24 1997-01-23 Method frame storage using multiple memory circuits
US10/346,886 Expired - Fee Related US6950923B2 (en) 1996-01-24 2003-01-17 Method frame storage using multiple memory circuits
US10/442,410 Expired - Fee Related US6961843B2 (en) 1996-01-24 2003-05-20 Method frame storage using multiple memory circuits
US11/096,183 Abandoned US20050267996A1 (en) 1996-01-24 2005-03-30 Method frame storage using multiple memory circuits
US11/461,340 Abandoned US20070277021A1 (en) 1996-01-24 2006-07-31 Instruction folding for a stack-based machine

Family Applications Before (4)

Application Number Title Priority Date Filing Date
US08/787,617 Expired - Lifetime US6532531B1 (en) 1996-01-24 1997-01-23 Method frame storage using multiple memory circuits
US10/346,886 Expired - Fee Related US6950923B2 (en) 1996-01-24 2003-01-17 Method frame storage using multiple memory circuits
US10/442,410 Expired - Fee Related US6961843B2 (en) 1996-01-24 2003-05-20 Method frame storage using multiple memory circuits
US11/096,183 Abandoned US20050267996A1 (en) 1996-01-24 2005-03-30 Method frame storage using multiple memory circuits

Country Status (6)

Country Link
US (5) US6532531B1 (en)
EP (1) EP0976034B1 (en)
JP (1) JP3634379B2 (en)
KR (1) KR100584964B1 (en)
DE (1) DE69734399D1 (en)
WO (1) WO1997027539A1 (en)

Cited By (30)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20060029053A1 (en) * 2000-05-03 2006-02-09 At&T Laboratories-Cambridge Ltd. Data transfer, synchronising applications, and low latency networks
US20070094474A1 (en) * 2005-10-26 2007-04-26 James Wilson Lookup table addressing system and method
US20070204257A1 (en) * 2005-11-28 2007-08-30 Ntt Docomo, Inc. Software operation modeling device, software operation monitoring device, software operation modeling method, and software operation monitoring method
US20090089649A1 (en) * 2007-09-27 2009-04-02 James Wilson Programmable compute unit with internal register and bit FIFO for executing Viterbi code
US20090150863A1 (en) * 2005-12-30 2009-06-11 Peng Guo Type checking for object-oriented programming languages
US20100158245A1 (en) * 2008-12-24 2010-06-24 Microsoft Corporation Implicit iteration of keyed array symbol
US20100228661A1 (en) * 2009-03-06 2010-09-09 Lutnick Howard W Method and apparatus for exchange-based condition processing
US20100241550A1 (en) * 2009-03-18 2010-09-23 Lutnick Howard W Interprogram communication using messages related to events and actions on an exchange
US20110153307A1 (en) * 2009-12-23 2011-06-23 Sebastian Winkel Transitioning From Source Instruction Set Architecture (ISA) Code To Translated Code In A Partial Emulation Environment
US20110306374A1 (en) * 2009-02-23 2011-12-15 Mitsubishi Electric Corporation Wireless communication system, wireless communication device, and wireless communication method
WO2013039481A1 (en) * 2011-09-13 2013-03-21 Empire Technology Development Llc Operation transfer from an origin virtual machine to a destination virtual machine
US20130132709A1 (en) * 2011-11-22 2013-05-23 International Business Machines Corporation Method and system for processing instruction information
US8458445B2 (en) 2005-10-26 2013-06-04 Analog Devices Inc. Compute units using local luts to reduce pipeline stalls
WO2013133846A1 (en) * 2012-03-09 2013-09-12 Empire Technology Development Llc Compilation of code in a data center
US20140379986A1 (en) * 2013-06-20 2014-12-25 Advanced Micro Devices, Inc. Stack access tracking
US20150007125A1 (en) * 2013-06-28 2015-01-01 Sap Ag Type instances
US20150186168A1 (en) * 2013-12-30 2015-07-02 Unisys Corporation Dedicating processing resources to just-in-time compilers and instruction processors in a dynamic translator
US20150254185A1 (en) * 2011-10-31 2015-09-10 Stec, Inc. System and Method to Cache Hypervisor Data
US20150293855A1 (en) * 2014-04-15 2015-10-15 International Business Machines Corporation Page table including data fetch width indicator
US20160283257A1 (en) * 2015-03-25 2016-09-29 Vmware, Inc. Parallelized virtual machine configuration
WO2016126523A3 (en) * 2015-02-02 2016-09-29 Qualcomm Incorporated Authenticated control stacks
US9891936B2 (en) 2013-09-27 2018-02-13 Intel Corporation Method and apparatus for page-level monitoring
US10369451B2 (en) 2012-07-05 2019-08-06 Golfzondeca, Inc. Golf GPS device with automatic hole recognition and playing hole selection
US10572252B2 (en) 2013-08-08 2020-02-25 Movidius Limited Variable-length instruction buffer management
US10621092B2 (en) 2008-11-24 2020-04-14 Intel Corporation Merging level cache and data cache units having indicator bits related to speculative execution
US10649746B2 (en) 2011-09-30 2020-05-12 Intel Corporation Instruction and logic to perform dynamic binary translation
US10725755B2 (en) 2008-11-24 2020-07-28 Intel Corporation Systems, apparatuses, and methods for a hardware and software system to automatically decompose a program to multiple parallel threads
US20200371803A1 (en) * 2019-05-24 2020-11-26 Texas Instruments Incorporated System and method for addressing data in memory
EP3136228B1 (en) * 2015-08-24 2023-08-09 Glenfly Tech Co., Ltd. Methods for combining instructions and apparatuses having multiple data pipes
US11768689B2 (en) 2013-08-08 2023-09-26 Movidius Limited Apparatus, systems, and methods for low power computational imaging

Families Citing this family (106)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO1997027539A1 (en) * 1996-01-24 1997-07-31 Sun Microsystems, Inc. Methods and apparatuses for stack caching
EP0825506B1 (en) 1996-08-20 2013-03-06 Invensys Systems, Inc. Methods and apparatus for remote process control
US6263401B1 (en) * 1997-09-30 2001-07-17 Institute For The Development Of Emerging Architectures, L.L.C. Method and apparatus for transferring data between a register stack and a memory resource
EP1135727A1 (en) * 1998-11-25 2001-09-26 Sun Microsystems, Inc. A method for enabling comprehensive profiling of garbage-collected memory systems
US7225436B1 (en) 1998-12-08 2007-05-29 Nazomi Communications Inc. Java hardware accelerator using microcode engine
US6332215B1 (en) 1998-12-08 2001-12-18 Nazomi Communications, Inc. Java virtual machine hardware for RISC and CISC processors
JP3355602B2 (en) * 1999-01-27 2002-12-09 インターナショナル・ビジネス・マシーンズ・コーポレーション Method and apparatus for processing multidimensional array objects
US7089530B1 (en) 1999-05-17 2006-08-08 Invensys Systems, Inc. Process control configuration system with connection validation and configuration
WO2000070531A2 (en) 1999-05-17 2000-11-23 The Foxboro Company Methods and apparatus for control configuration
US6788980B1 (en) 1999-06-11 2004-09-07 Invensys Systems, Inc. Methods and apparatus for control using control devices that provide a virtual machine environment and that communicate via an IP network
US6665793B1 (en) * 1999-12-28 2003-12-16 Institute For The Development Of Emerging Architectures, L.L.C. Method and apparatus for managing access to out-of-frame Registers
KR20010076328A (en) * 2000-01-19 2001-08-11 이정태 Aparatus for processing tcp/ip by hardware, and operating method thereof
GB2367654B (en) * 2000-10-05 2004-10-27 Advanced Risc Mach Ltd Storing stack operands in registers
GB2367653B (en) * 2000-10-05 2004-10-20 Advanced Risc Mach Ltd Restarting translated instructions
KR20020028814A (en) 2000-10-10 2002-04-17 나조미 커뮤니케이션즈, 인코포레이티드 Java hardware accelerator using microcode engine
WO2002045385A2 (en) * 2000-11-20 2002-06-06 Zucotto Wireless, Inc. Methods and devices for caching method frame segments in a low-power stack-based processor
US8769508B2 (en) 2001-08-24 2014-07-01 Nazomi Communications Inc. Virtual machine hardware for RISC and CISC processors
US6772292B2 (en) * 2002-06-04 2004-08-03 Isaak Garber Two area stack
US7228532B1 (en) * 2002-06-26 2007-06-05 Sun Microsystems, Inc. Method and apparatus to facilitate code verification and garbage collection in a platform-independent virtual machine
EP1387249B1 (en) * 2002-07-31 2019-03-13 Texas Instruments Incorporated RISC processor having a stack and register architecture
EP1387274A3 (en) 2002-07-31 2004-08-11 Texas Instruments Incorporated Memory management for local variables
GB0220282D0 (en) * 2002-08-31 2002-10-09 Ibm Improved just in time compilation of java software methods
US7367022B2 (en) * 2002-09-05 2008-04-29 Intel Corporation Methods and apparatus for optimizing the operating speed and size of a computer program
US7801120B2 (en) * 2003-01-13 2010-09-21 Emulex Design & Manufacturing Corporation Method and system for efficient queue management
US7136990B2 (en) * 2003-01-16 2006-11-14 Ip-First, Llc. Fast POP operation from RAM cache using cache row value stack
US7191291B2 (en) * 2003-01-16 2007-03-13 Ip-First, Llc Microprocessor with variable latency stack cache
US7139877B2 (en) * 2003-01-16 2006-11-21 Ip-First, Llc Microprocessor and apparatus for performing speculative load operation from a stack memory cache
US7139876B2 (en) * 2003-01-16 2006-11-21 Ip-First, Llc Microprocessor and apparatus for performing fast speculative pop operation from a stack memory cache
US7694301B1 (en) * 2003-06-27 2010-04-06 Nathan Laredo Method and system for supporting input/output for a virtual machine
US20050066305A1 (en) * 2003-09-22 2005-03-24 Lisanke Robert John Method and machine for efficient simulation of digital hardware within a software development environment
US7444632B2 (en) * 2003-09-25 2008-10-28 International Business Machines Corporation Balancing computational load across a plurality of processors
US7516456B2 (en) * 2003-09-25 2009-04-07 International Business Machines Corporation Asymmetric heterogeneous multi-threaded operating system
US7415703B2 (en) * 2003-09-25 2008-08-19 International Business Machines Corporation Loading software on a plurality of processors
US7496917B2 (en) * 2003-09-25 2009-02-24 International Business Machines Corporation Virtual devices using a pluarlity of processors
US7523157B2 (en) * 2003-09-25 2009-04-21 International Business Machines Corporation Managing a plurality of processors as devices
US7389508B2 (en) * 2003-09-25 2008-06-17 International Business Machines Corporation System and method for grouping processors and assigning shared memory space to a group in heterogeneous computer environment
US20050071828A1 (en) * 2003-09-25 2005-03-31 International Business Machines Corporation System and method for compiling source code for multi-processor environments
US7549145B2 (en) * 2003-09-25 2009-06-16 International Business Machines Corporation Processor dedicated code handling in a multi-processor environment
US7478390B2 (en) * 2003-09-25 2009-01-13 International Business Machines Corporation Task queue management of virtual devices using a plurality of processors
US7257665B2 (en) * 2003-09-29 2007-08-14 Intel Corporation Branch-aware FIFO for interprocessor data sharing
US20050071606A1 (en) * 2003-09-30 2005-03-31 Roman Talyansky Device, system and method of allocating spill cells in binary instrumentation using one free register
US7707389B2 (en) * 2003-10-31 2010-04-27 Mips Technologies, Inc. Multi-ISA instruction fetch unit for a processor, and applications thereof
FR2864411B1 (en) * 2003-12-23 2006-03-03 Cit Alcatel TERMINAL WITH MEANS OF PROTECTION AGAINST THE DYSFUNCTION OF CERTAIN JAVA APPLICATIONS
AT413739B (en) * 2004-02-09 2006-05-15 Ge Jenbacher Gmbh & Co Ohg METHOD FOR REGULATING A COMBUSTION ENGINE
DE102004025419A1 (en) * 2004-05-24 2005-12-22 Infineon Technologies Ag Controller and method for processing instructions
DE102004025418A1 (en) * 2004-05-24 2005-12-22 Infineon Technologies Ag Controller with a decoder
US7278122B2 (en) * 2004-06-24 2007-10-02 Ftl Systems, Inc. Hardware/software design tool and language specification mechanism enabling efficient technology retargeting and optimization
US20060095675A1 (en) * 2004-08-23 2006-05-04 Rongzhen Yang Three stage hybrid stack model
US7594081B2 (en) 2004-09-10 2009-09-22 Cavium Networks, Inc. Direct access to low-latency memory
WO2006031551A2 (en) 2004-09-10 2006-03-23 Cavium Networks Selective replication of data structure
US7941585B2 (en) * 2004-09-10 2011-05-10 Cavium Networks, Inc. Local scratchpad and data caching system
US7526502B2 (en) * 2004-09-10 2009-04-28 Microsoft Corporation Dynamic call site binding
US7314491B2 (en) * 2004-12-29 2008-01-01 Bull Hn Information Systems Inc. Encapsulation of large native operating system functions as enhancements of the instruction set in an emulated central processor system
US7478224B2 (en) * 2005-04-15 2009-01-13 Atmel Corporation Microprocessor access of operand stack as a register file using native instructions
KR100725393B1 (en) * 2005-05-19 2007-06-07 삼성전자주식회사 System and method for reducing execution time of byte code at java virtual machine
US7676796B2 (en) * 2005-09-29 2010-03-09 Intel Corporation Device, system and method for maintaining a pre-defined number of free registers within an instrumented program
US7728744B2 (en) * 2005-10-26 2010-06-01 Analog Devices, Inc. Variable length decoder system and method
US7454572B2 (en) * 2005-11-08 2008-11-18 Mediatek Inc. Stack caching systems and methods with an active swapping mechanism
US7366842B1 (en) * 2005-12-15 2008-04-29 Nvidia Corporation Creating permanent storage on the fly within existing buffers
US7502029B2 (en) * 2006-01-17 2009-03-10 Silicon Integrated Systems Corp. Instruction folding mechanism, method for performing the same and pixel processing system employing the same
DE102006041002B4 (en) * 2006-08-31 2009-01-02 Infineon Technologies Ag Method to adapt a program to a cache, and circuitry
GB2442495B (en) * 2006-10-02 2009-04-01 Transitive Ltd Method and apparatus for handling dynamically linked function cells with respect to program code conversion
DE102007051345A1 (en) * 2007-10-26 2009-04-30 Fraunhofer-Gesellschaft zur Förderung der angewandten Forschung e.V. Explosive charge
CN104407518B (en) 2008-06-20 2017-05-31 因文西斯系统公司 The system and method interacted to the reality and Simulation Facility for process control
KR101493076B1 (en) 2009-04-07 2015-02-12 삼성전자 주식회사 Apparatus and method of preventing virus code execution through buffer overflow control
US8463964B2 (en) 2009-05-29 2013-06-11 Invensys Systems, Inc. Methods and apparatus for control configuration with enhanced change-tracking
US8127060B2 (en) 2009-05-29 2012-02-28 Invensys Systems, Inc Methods and apparatus for control configuration with control objects that are fieldbus protocol-aware
US9268945B2 (en) * 2010-03-19 2016-02-23 Contrast Security, Llc Detection of vulnerabilities in computer systems
US20130086359A1 (en) * 2011-09-29 2013-04-04 Qualcomm Incorporated Processor Hardware Pipeline Configured for Single-Instruction Address Extraction and Memory Access Operation
US20130091331A1 (en) * 2011-10-11 2013-04-11 Iulian Moraru Methods, apparatus, and articles of manufacture to manage memory
US9471344B1 (en) * 2012-03-27 2016-10-18 Marvell International Ltd. Hardware support for processing virtual machine instructions
US9569211B2 (en) 2012-08-03 2017-02-14 International Business Machines Corporation Predication in a vector processor
US9575755B2 (en) 2012-08-03 2017-02-21 International Business Machines Corporation Vector processing in an active memory device
US9003160B2 (en) 2012-08-03 2015-04-07 International Business Machines Corporation Active buffered memory
US9632777B2 (en) 2012-08-03 2017-04-25 International Business Machines Corporation Gather/scatter of multiple data elements with packed loading/storing into/from a register file entry
US9594724B2 (en) 2012-08-09 2017-03-14 International Business Machines Corporation Vector register file
US9298395B2 (en) 2012-10-22 2016-03-29 Globalfoundries Inc. Memory system connector
US8972782B2 (en) 2012-11-09 2015-03-03 International Business Machines Corporation Exposed-pipeline processing element with rollback
US9189399B2 (en) 2012-11-21 2015-11-17 Advanced Micro Devices, Inc. Stack cache management and coherence techniques
US20140143498A1 (en) * 2012-11-21 2014-05-22 Advanced Micro Devices, Inc. Methods and apparatus for filtering stack data within a cache memory hierarchy
US9734059B2 (en) 2012-11-21 2017-08-15 Advanced Micro Devices, Inc. Methods and apparatus for data cache way prediction based on classification as stack data
US9032157B2 (en) * 2012-12-11 2015-05-12 International Business Machines Corporation Virtual machine failover
US9069701B2 (en) * 2012-12-11 2015-06-30 International Business Machines Corporation Virtual machine failover
KR101893814B1 (en) 2014-03-26 2018-10-04 인텔 코포레이션 Three source operand floating point addition processors, methods, systems, and instructions
US9965185B2 (en) 2015-01-20 2018-05-08 Ultrata, Llc Utilization of a distributed index to provide object memory fabric coherency
EP3248097B1 (en) 2015-01-20 2022-02-09 Ultrata LLC Object memory data flow instruction execution
US9886210B2 (en) 2015-06-09 2018-02-06 Ultrata, Llc Infinite memory fabric hardware implementation with router
US9971542B2 (en) 2015-06-09 2018-05-15 Ultrata, Llc Infinite memory fabric streams and APIs
US10698628B2 (en) 2015-06-09 2020-06-30 Ultrata, Llc Infinite memory fabric hardware implementation with memory
US9705620B2 (en) * 2015-09-18 2017-07-11 Qualcomm Incorporated Synchronization of endpoints using tunable latency
CA3006773A1 (en) 2015-12-08 2017-06-15 Ultrata, Llc Memory fabric software implementation
CN115061971A (en) 2015-12-08 2022-09-16 乌尔特拉塔有限责任公司 Memory fabric operation and consistency using fault tolerant objects
US10241676B2 (en) 2015-12-08 2019-03-26 Ultrata, Llc Memory fabric software implementation
US10235063B2 (en) 2015-12-08 2019-03-19 Ultrata, Llc Memory fabric operations and coherency using fault tolerant objects
US10908910B2 (en) * 2018-07-27 2021-02-02 Oracle International Corporation Lazy copying of runtime-managed stack frames
US10942852B1 (en) 2019-09-12 2021-03-09 Advanced New Technologies Co., Ltd. Log-structured storage systems
CN111566611B (en) * 2019-09-12 2023-08-04 创新先进技术有限公司 Log structured storage system
SG11202002588RA (en) 2019-09-12 2020-04-29 Alibaba Group Holding Ltd Log-structured storage systems
SG11202002587TA (en) 2019-09-12 2020-04-29 Alibaba Group Holding Ltd Log-structured storage systems
SG11202002614XA (en) 2019-09-12 2020-04-29 Alibaba Group Holding Ltd Log-structured storage systems
WO2019228575A2 (en) 2019-09-12 2019-12-05 Alibaba Group Holding Limited Log-structured storage systems
WO2019228572A2 (en) 2019-09-12 2019-12-05 Alibaba Group Holding Limited Log-structured storage systems
EP3695586A4 (en) 2019-09-12 2020-12-09 Alibaba Group Holding Limited Log-structured storage systems
CN111837113A (en) 2019-09-12 2020-10-27 创新先进技术有限公司 Log structure storage system
CN113326020A (en) * 2020-02-28 2021-08-31 北京百度网讯科技有限公司 Cache device, cache, system, data processing method, device and medium
US11809839B2 (en) 2022-01-18 2023-11-07 Robert Lyden Computer language and code for application development and electronic and optical communication

Citations (38)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US4439828A (en) * 1981-07-27 1984-03-27 International Business Machines Corp. Instruction substitution mechanism in an instruction handling unit of a data processing system
US5187793A (en) * 1989-01-09 1993-02-16 Intel Corporation Processor with hierarchal memory and using meta-instructions for software control of loading, unloading and execution of machine instructions stored in the cache
US5274818A (en) * 1992-02-03 1993-12-28 Thinking Machines Corporation System and method for compiling a fine-grained array based source program onto a course-grained hardware
US5313614A (en) * 1988-12-06 1994-05-17 At&T Bell Laboratories Method and apparatus for direct conversion of programs in object code form between different hardware architecture computer systems
US5359507A (en) * 1989-04-07 1994-10-25 Mitsubishi Denki Kabushiki Kaisha Sequence controller
US5367650A (en) * 1992-07-31 1994-11-22 Intel Corporation Method and apparauts for parallel exchange operation in a pipelined processor
US5479643A (en) * 1985-11-18 1995-12-26 John Fluke Mfg. Co., Inc. Virtual machine programming system
US5499352A (en) * 1993-09-30 1996-03-12 Intel Corporation Floating point register alias table FXCH and retirement floating point register array
US5522051A (en) * 1992-07-29 1996-05-28 Intel Corporation Method and apparatus for stack manipulation in a pipelined processor
US5524262A (en) * 1993-09-30 1996-06-04 Intel Corporation Apparatus and method for renaming registers in a processor and resolving data dependencies thereof
US5581183A (en) * 1994-09-30 1996-12-03 Siemens Aktiengesellschaft Method for the operation of a magnetic resonance apparatus
US5596732A (en) * 1993-10-19 1997-01-21 Fujitsu Limited Method of optimizing instruction sequence of compiler
US5600726A (en) * 1995-04-07 1997-02-04 Gemini Systems, L.L.C. Method for creating specific purpose rule-based n-bit virtual machines
US5619655A (en) * 1993-02-26 1997-04-08 Siemens Aktiengesellschaft System for administration and management of network from remote or central station through an operator interface which prepares and stores the management orders for execution
US5634118A (en) * 1995-04-10 1997-05-27 Exponential Technology, Inc. Splitting a floating-point stack-exchange instruction for merging into surrounding instructions by operand translation
US5638525A (en) * 1995-02-10 1997-06-10 Intel Corporation Processor capable of executing programs that contain RISC and CISC instructions
US5657486A (en) * 1995-12-07 1997-08-12 Teradyne, Inc. Automatic test equipment with pipelined sequencer
US5699537A (en) * 1995-12-22 1997-12-16 Intel Corporation Processor microarchitecture for efficient dynamic scheduling and execution of chains of dependent instructions
US5740441A (en) * 1994-12-20 1998-04-14 Sun Microsystems, Inc. Bytecode program interpreter apparatus and method with pre-verification of data type restrictions and object initialization
US5862370A (en) * 1995-09-27 1999-01-19 Vlsi Technology, Inc. Data processor system with instruction substitution filter for deimplementing instructions
US5903761A (en) * 1997-10-31 1999-05-11 Preemptive Solutions, Inc. Method of reducing the number of instructions in a program code sequence
US5953741A (en) * 1996-11-27 1999-09-14 Vlsi Technology, Inc. Stack cache for stack-based processor and method thereof
US6026485A (en) * 1996-01-24 2000-02-15 Sun Microsystems, Inc. Instruction folding for a stack-based machine
US6151703A (en) * 1996-05-20 2000-11-21 Inprise Corporation Development system with methods for just-in-time compilation of programs
US6205578B1 (en) * 1998-08-14 2001-03-20 Ati International Srl Interpreter for stack-based languages
US6332215B1 (en) * 1998-12-08 2001-12-18 Nazomi Communications, Inc. Java virtual machine hardware for RISC and CISC processors
US6338160B1 (en) * 1998-12-08 2002-01-08 Nazomi Communications, Inc. Constant pool reference resolution method
US20020019976A1 (en) * 1998-12-08 2002-02-14 Patel Mukesh K. Java hardware accelerator using thread manager
US6349383B1 (en) * 1998-09-10 2002-02-19 Ip-First, L.L.C. System for combining adjacent push/pop stack program instructions into single double push/pop stack microinstuction for execution
US6711667B1 (en) * 1996-06-28 2004-03-23 Legerity, Inc. Microprocessor configured to translate instructions from one instruction set to another, and to store the translated instructions
US6832307B2 (en) * 2001-07-19 2004-12-14 Stmicroelectronics, Inc. Instruction fetch buffer stack fold decoder for generating foldable instruction status information
US6904517B2 (en) * 2000-11-27 2005-06-07 Arm Limited Data processing apparatus and method for saving return state
US6907515B2 (en) * 2001-05-31 2005-06-14 Arm Limited Configuration control within data processing systems
US20050149694A1 (en) * 1998-12-08 2005-07-07 Mukesh Patel Java hardware accelerator using microcode engine
US6965984B2 (en) * 2001-05-31 2005-11-15 Arm Limited Data processing using multiple instruction sets
US7000094B2 (en) * 2000-10-05 2006-02-14 Arm Limited Storing stack operands in registers
US7003652B2 (en) * 2000-10-05 2006-02-21 Arm Limited Restarting translated instructions
US7076771B2 (en) * 2000-12-01 2006-07-11 Arm Limited Instruction interpretation within a data processing system

Family Cites Families (52)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US3878513A (en) 1972-02-08 1975-04-15 Burroughs Corp Data processing method and apparatus using occupancy indications to reserve storage space for a stack
US3810117A (en) 1972-10-20 1974-05-07 Ibm Stack mechanism for a data processor
GB1441816A (en) 1973-07-18 1976-07-07 Int Computers Ltd Electronic digital data processing systems
US3889243A (en) 1973-10-18 1975-06-10 Ibm Stack mechanism for a data processor
JPS5474651A (en) 1977-11-28 1979-06-14 Toshiba Corp Stack control system
US4354232A (en) 1977-12-16 1982-10-12 Honeywell Information Systems Inc. Cache memory command buffer circuit
US4325118A (en) 1980-03-03 1982-04-13 Western Digital Corporation Instruction fetch circuitry for computers
US4524416A (en) 1980-04-15 1985-06-18 Honeywell Information Systems Inc. Stack mechanism with the ability to dynamically alter the size of a stack in a data processing system
US4375678A (en) 1980-08-25 1983-03-01 Sperry Corporation Redundant memory arrangement providing simultaneous access
US4530049A (en) 1982-02-11 1985-07-16 At&T Bell Laboratories Stack cache with fixed size stack frames
US5043870A (en) 1982-02-24 1991-08-27 At&T Bell Laboratories Computer with automatic mapping of memory contents into machine registers during program execution
US4600986A (en) 1984-04-02 1986-07-15 Sperry Corporation Pipelined split stack with high performance interleaved decode
US4674032A (en) 1984-04-02 1987-06-16 Unisys Corporation High-performance pipelined stack with over-write protection
JPS6133546A (en) 1984-07-25 1986-02-17 Nec Corp Information processor
US4761733A (en) 1985-03-11 1988-08-02 Celerity Computing Direct-execution microprogrammable microprocessor system
DE3689595T2 (en) 1985-04-08 1994-05-19 Hitachi Ltd Data processing system.
JP2826309B2 (en) 1985-11-20 1998-11-18 日本電気株式会社 Information processing device
JPH0687221B2 (en) 1986-04-02 1994-11-02 日本電気株式会社 Information processing equipment
JP2545789B2 (en) 1986-04-14 1996-10-23 株式会社日立製作所 Information processing device
US4811208A (en) 1986-05-16 1989-03-07 Intel Corporation Stack frame cache on a microprocessor chip
JPS63242243A (en) 1987-03-31 1988-10-07 株式会社東芝 Ultrasonic doppler diagnostic apparatus
US5115500A (en) 1988-01-11 1992-05-19 International Business Machines Corporation Plural incompatible instruction format decode method and apparatus
US5210874A (en) 1988-03-22 1993-05-11 Digital Equipment Corporation Cross-domain call system in a capability based digital data processing system
US4951194A (en) 1989-01-23 1990-08-21 Tektronix, Inc. Method for reducing memory allocations and data copying operations during program calling sequences
US5107457A (en) * 1989-04-03 1992-04-21 The Johns Hopkins University Stack data cache having a stack management hardware with internal and external stack pointers and buffers for handling underflow and overflow stack
US5142635A (en) 1989-04-07 1992-08-25 Intel Corporation Method and circuitry for performing multiple stack operations in succession in a pipelined digital computer
US5093777A (en) 1989-06-12 1992-03-03 Bull Hn Information Systems Inc. Method and apparatus for predicting address of a subsequent cache request upon analyzing address patterns stored in separate miss stack
US5157777A (en) * 1989-12-22 1992-10-20 Intel Corporation Synchronous communication between execution environments in a data processing system employing an object-oriented memory protection mechanism
JP2818249B2 (en) * 1990-03-30 1998-10-30 株式会社東芝 Electronic computer
US5471591A (en) * 1990-06-29 1995-11-28 Digital Equipment Corporation Combined write-operand queue and read-after-write dependency scoreboard
JP3027627B2 (en) 1991-02-25 2000-04-04 松下電工株式会社 Arithmetic processor of programmable controller
US5701417A (en) 1991-03-27 1997-12-23 Microstar Laboratories Method and apparatus for providing initial instructions through a communications interface in a multiple computer system
JP3204323B2 (en) 1991-07-05 2001-09-04 エヌイーシーマイクロシステム株式会社 Microprocessor with built-in cache memory
US5634027A (en) 1991-11-20 1997-05-27 Kabushiki Kaisha Toshiba Cache memory system for multiple processors with collectively arranged cache tag memories
US5438668A (en) 1992-03-31 1995-08-01 Seiko Epson Corporation System and method for extraction, alignment and decoding of CISC instructions into a nano-instruction bucket for execution by a RISC computer
US5471602A (en) * 1992-07-31 1995-11-28 Hewlett-Packard Company System and method of scoreboarding individual cache line segments
WO1994003860A1 (en) * 1992-08-07 1994-02-17 Thinking Machines Corporation Massively parallel computer including auxiliary vector processor
JPH0793216A (en) 1993-09-27 1995-04-07 Hitachi Ltd Cache memory controller
US5481684A (en) 1994-01-11 1996-01-02 Exponential Technology, Inc. Emulating operating system calls in an alternate instruction set using a modified code segment descriptor
EP0676691A3 (en) 1994-04-06 1996-12-11 Hewlett Packard Co Apparatus for register saving and restoring in a digital computer.
US5485572A (en) 1994-04-26 1996-01-16 Unisys Corporation Response stack state validation check
US5481693A (en) 1994-07-20 1996-01-02 Exponential Technology, Inc. Shared register architecture for a dual-instruction-set CPU
US5598546A (en) 1994-08-31 1997-01-28 Exponential Technology, Inc. Dual-architecture super-scalar pipeline
US5636362A (en) 1994-09-28 1997-06-03 Intel Corporation Programmable high watermark in stack frame cache using second region as a storage if first region is full and an event having a predetermined minimum priority
US6496922B1 (en) 1994-10-31 2002-12-17 Sun Microsystems, Inc. Method and apparatus for multiplatform stateless instruction set architecture (ISA) using ISA tags on-the-fly instruction translation
US5630066A (en) 1994-12-20 1997-05-13 Sun Microsystems, Inc. System and method for locating object view and platform independent object
US6076155A (en) 1995-10-24 2000-06-13 S3 Incorporated Shared register architecture for a dual-instruction-set CPU to facilitate data exchange between the instruction sets
US5687336A (en) 1996-01-11 1997-11-11 Exponential Technology, Inc. Stack push/pop tracking and pairing in a pipelined processor
US5761408A (en) * 1996-01-16 1998-06-02 Parasoft Corporation Method and system for generating a computer program test suite using dynamic symbolic execution
US5784553A (en) 1996-01-16 1998-07-21 Parasoft Corporation Method and system for generating a computer program test suite using dynamic symbolic execution of JAVA programs
US5925123A (en) 1996-01-24 1999-07-20 Sun Microsystems, Inc. Processor for executing instruction sets received from a network or from a local memory
WO1997027539A1 (en) * 1996-01-24 1997-07-31 Sun Microsystems, Inc. Methods and apparatuses for stack caching

Patent Citations (41)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US4439828A (en) * 1981-07-27 1984-03-27 International Business Machines Corp. Instruction substitution mechanism in an instruction handling unit of a data processing system
US5479643A (en) * 1985-11-18 1995-12-26 John Fluke Mfg. Co., Inc. Virtual machine programming system
US5313614A (en) * 1988-12-06 1994-05-17 At&T Bell Laboratories Method and apparatus for direct conversion of programs in object code form between different hardware architecture computer systems
US5187793A (en) * 1989-01-09 1993-02-16 Intel Corporation Processor with hierarchal memory and using meta-instructions for software control of loading, unloading and execution of machine instructions stored in the cache
US5359507A (en) * 1989-04-07 1994-10-25 Mitsubishi Denki Kabushiki Kaisha Sequence controller
US5274818A (en) * 1992-02-03 1993-12-28 Thinking Machines Corporation System and method for compiling a fine-grained array based source program onto a course-grained hardware
US5522051A (en) * 1992-07-29 1996-05-28 Intel Corporation Method and apparatus for stack manipulation in a pipelined processor
US5367650A (en) * 1992-07-31 1994-11-22 Intel Corporation Method and apparauts for parallel exchange operation in a pipelined processor
US5619655A (en) * 1993-02-26 1997-04-08 Siemens Aktiengesellschaft System for administration and management of network from remote or central station through an operator interface which prepares and stores the management orders for execution
US5499352A (en) * 1993-09-30 1996-03-12 Intel Corporation Floating point register alias table FXCH and retirement floating point register array
US5524262A (en) * 1993-09-30 1996-06-04 Intel Corporation Apparatus and method for renaming registers in a processor and resolving data dependencies thereof
US5596732A (en) * 1993-10-19 1997-01-21 Fujitsu Limited Method of optimizing instruction sequence of compiler
US5581183A (en) * 1994-09-30 1996-12-03 Siemens Aktiengesellschaft Method for the operation of a magnetic resonance apparatus
US5740441A (en) * 1994-12-20 1998-04-14 Sun Microsystems, Inc. Bytecode program interpreter apparatus and method with pre-verification of data type restrictions and object initialization
US5638525A (en) * 1995-02-10 1997-06-10 Intel Corporation Processor capable of executing programs that contain RISC and CISC instructions
US5600726A (en) * 1995-04-07 1997-02-04 Gemini Systems, L.L.C. Method for creating specific purpose rule-based n-bit virtual machines
US5634118A (en) * 1995-04-10 1997-05-27 Exponential Technology, Inc. Splitting a floating-point stack-exchange instruction for merging into surrounding instructions by operand translation
US5862370A (en) * 1995-09-27 1999-01-19 Vlsi Technology, Inc. Data processor system with instruction substitution filter for deimplementing instructions
US5657486A (en) * 1995-12-07 1997-08-12 Teradyne, Inc. Automatic test equipment with pipelined sequencer
US5699537A (en) * 1995-12-22 1997-12-16 Intel Corporation Processor microarchitecture for efficient dynamic scheduling and execution of chains of dependent instructions
US6026485A (en) * 1996-01-24 2000-02-15 Sun Microsystems, Inc. Instruction folding for a stack-based machine
US6151703A (en) * 1996-05-20 2000-11-21 Inprise Corporation Development system with methods for just-in-time compilation of programs
US6711667B1 (en) * 1996-06-28 2004-03-23 Legerity, Inc. Microprocessor configured to translate instructions from one instruction set to another, and to store the translated instructions
US5953741A (en) * 1996-11-27 1999-09-14 Vlsi Technology, Inc. Stack cache for stack-based processor and method thereof
US5903761A (en) * 1997-10-31 1999-05-11 Preemptive Solutions, Inc. Method of reducing the number of instructions in a program code sequence
US6205578B1 (en) * 1998-08-14 2001-03-20 Ati International Srl Interpreter for stack-based languages
US6349383B1 (en) * 1998-09-10 2002-02-19 Ip-First, L.L.C. System for combining adjacent push/pop stack program instructions into single double push/pop stack microinstuction for execution
US6826749B2 (en) * 1998-12-08 2004-11-30 Nazomi Communications, Inc. Java hardware accelerator using thread manager
US20050149694A1 (en) * 1998-12-08 2005-07-07 Mukesh Patel Java hardware accelerator using microcode engine
US20020066083A1 (en) * 1998-12-08 2002-05-30 Patel Mukesh K. Java virtual machine hardware for RISC and CISC processors
US6338160B1 (en) * 1998-12-08 2002-01-08 Nazomi Communications, Inc. Constant pool reference resolution method
US6332215B1 (en) * 1998-12-08 2001-12-18 Nazomi Communications, Inc. Java virtual machine hardware for RISC and CISC processors
US7080362B2 (en) * 1998-12-08 2006-07-18 Nazomi Communication, Inc. Java virtual machine hardware for RISC and CISC processors
US20020019976A1 (en) * 1998-12-08 2002-02-14 Patel Mukesh K. Java hardware accelerator using thread manager
US7000094B2 (en) * 2000-10-05 2006-02-14 Arm Limited Storing stack operands in registers
US7003652B2 (en) * 2000-10-05 2006-02-21 Arm Limited Restarting translated instructions
US6904517B2 (en) * 2000-11-27 2005-06-07 Arm Limited Data processing apparatus and method for saving return state
US7076771B2 (en) * 2000-12-01 2006-07-11 Arm Limited Instruction interpretation within a data processing system
US6907515B2 (en) * 2001-05-31 2005-06-14 Arm Limited Configuration control within data processing systems
US6965984B2 (en) * 2001-05-31 2005-11-15 Arm Limited Data processing using multiple instruction sets
US6832307B2 (en) * 2001-07-19 2004-12-14 Stmicroelectronics, Inc. Instruction fetch buffer stack fold decoder for generating foldable instruction status information

Cited By (67)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US8843655B2 (en) 1999-05-04 2014-09-23 At&T Investments Uk Llc Data transfer, synchronising applications, and low latency networks
US20080228946A1 (en) * 1999-05-04 2008-09-18 At&T Laboratories-Cambridge Ltd. Data transfer, synchronising applications, and low latency networks
US8423675B2 (en) 1999-05-04 2013-04-16 At&T Intellectual Property I, L.P. Data transfer, synchronising applications, and low latency networks
US9769274B2 (en) 1999-05-04 2017-09-19 At&T Investments Uk Llc Data transfer, synchronising applications, and low latency networks
US8346971B2 (en) 1999-05-04 2013-01-01 At&T Intellectual Property I, Lp Data transfer, synchronising applications, and low latency networks
US8725903B2 (en) 1999-05-04 2014-05-13 At&T Intellectual Property I, L.P. Data transfer, synchronising applications, and low latency networks
US8073994B2 (en) * 2000-05-03 2011-12-06 At&T Laboratories Data transfer, synchronising applications, and low latency networks
US20060029053A1 (en) * 2000-05-03 2006-02-09 At&T Laboratories-Cambridge Ltd. Data transfer, synchronising applications, and low latency networks
US20070094474A1 (en) * 2005-10-26 2007-04-26 James Wilson Lookup table addressing system and method
US8458445B2 (en) 2005-10-26 2013-06-04 Analog Devices Inc. Compute units using local luts to reduce pipeline stalls
US8285972B2 (en) 2005-10-26 2012-10-09 Analog Devices, Inc. Lookup table addressing system and method
US8015551B2 (en) * 2005-11-28 2011-09-06 Ntt Docomo, Inc. Software operation modeling device, software operation monitoring device, software operation modeling method, and software operation monitoring method
US20070204257A1 (en) * 2005-11-28 2007-08-30 Ntt Docomo, Inc. Software operation modeling device, software operation monitoring device, software operation modeling method, and software operation monitoring method
US20090150863A1 (en) * 2005-12-30 2009-06-11 Peng Guo Type checking for object-oriented programming languages
US9009688B2 (en) * 2005-12-30 2015-04-14 Intel Corporation Type checking for object-oriented programming languages
US8301990B2 (en) * 2007-09-27 2012-10-30 Analog Devices, Inc. Programmable compute unit with internal register and bit FIFO for executing Viterbi code
US20090089649A1 (en) * 2007-09-27 2009-04-02 James Wilson Programmable compute unit with internal register and bit FIFO for executing Viterbi code
US10621092B2 (en) 2008-11-24 2020-04-14 Intel Corporation Merging level cache and data cache units having indicator bits related to speculative execution
US10725755B2 (en) 2008-11-24 2020-07-28 Intel Corporation Systems, apparatuses, and methods for a hardware and software system to automatically decompose a program to multiple parallel threads
US20100158245A1 (en) * 2008-12-24 2010-06-24 Microsoft Corporation Implicit iteration of keyed array symbol
US8453114B2 (en) * 2008-12-24 2013-05-28 Microsoft Corporation Implicit iteration of keyed array symbol
US20110306374A1 (en) * 2009-02-23 2011-12-15 Mitsubishi Electric Corporation Wireless communication system, wireless communication device, and wireless communication method
US8676239B2 (en) * 2009-02-23 2014-03-18 Mitsubishi Electric Corporation Wireless communication system, wireless communication device, and wireless communication method
US20190318429A1 (en) * 2009-03-06 2019-10-17 Bgc Partners, Inc. Method and apparatus for exchange-based condition processing
US11544790B2 (en) * 2009-03-06 2023-01-03 Bgc Partners, Inc. Method and apparatus for exchange-based condition processing
US20100228661A1 (en) * 2009-03-06 2010-09-09 Lutnick Howard W Method and apparatus for exchange-based condition processing
US10380689B2 (en) 2009-03-06 2019-08-13 Bgc Partners, Inc. Method and apparatus for exchange-based condition processing
US11250508B2 (en) 2009-03-18 2022-02-15 Bgc Partners, Inc. Interprogram communication using messages related to events and actions on an exchange
US20100241550A1 (en) * 2009-03-18 2010-09-23 Lutnick Howard W Interprogram communication using messages related to events and actions on an exchange
US10453130B2 (en) * 2009-03-18 2019-10-22 Bgc Partners, Inc. Electronic exchange system using messages related to events and actions on an exchange
US8775153B2 (en) * 2009-12-23 2014-07-08 Intel Corporation Transitioning from source instruction set architecture (ISA) code to translated code in a partial emulation environment
US20110153307A1 (en) * 2009-12-23 2011-06-23 Sebastian Winkel Transitioning From Source Instruction Set Architecture (ISA) Code To Translated Code In A Partial Emulation Environment
US9189271B2 (en) 2011-09-13 2015-11-17 Empire Technology Development, Llc Operation transfer from an origin virtual machine to a destination virtual machine while continue the execution of the operation on the origin virtual machine
CN103782276A (en) * 2011-09-13 2014-05-07 英派尔科技开发有限公司 Operation transfer from an origin virtual machine to a destination virtual machine
WO2013039481A1 (en) * 2011-09-13 2013-03-21 Empire Technology Development Llc Operation transfer from an origin virtual machine to a destination virtual machine
US10649746B2 (en) 2011-09-30 2020-05-12 Intel Corporation Instruction and logic to perform dynamic binary translation
US20150254185A1 (en) * 2011-10-31 2015-09-10 Stec, Inc. System and Method to Cache Hypervisor Data
US10248566B2 (en) * 2011-10-31 2019-04-02 Western Digital Technologies, Inc. System and method for caching virtual machine data
US9021422B2 (en) * 2011-11-22 2015-04-28 International Business Machines Corporation Method and system for processing instruction information
US20130132709A1 (en) * 2011-11-22 2013-05-23 International Business Machines Corporation Method and system for processing instruction information
WO2013133846A1 (en) * 2012-03-09 2013-09-12 Empire Technology Development Llc Compilation of code in a data center
US8990788B2 (en) 2012-03-09 2015-03-24 Empire Technology Development Llc Compilation of code in a data center
US11590403B2 (en) 2012-07-05 2023-02-28 GolfzonDeca Inc. Golf GPS device with hole recognition and hole selection
US11045708B2 (en) 2012-07-05 2021-06-29 Golfzondeca, Inc. Golf GPS device with hole recognition and hole selection
US10369451B2 (en) 2012-07-05 2019-08-06 Golfzondeca, Inc. Golf GPS device with automatic hole recognition and playing hole selection
US20140379986A1 (en) * 2013-06-20 2014-12-25 Advanced Micro Devices, Inc. Stack access tracking
US9292292B2 (en) * 2013-06-20 2016-03-22 Advanced Micro Devices, Inc. Stack access tracking
US20150007125A1 (en) * 2013-06-28 2015-01-01 Sap Ag Type instances
US8943462B2 (en) * 2013-06-28 2015-01-27 Sap Se Type instances
US10572252B2 (en) 2013-08-08 2020-02-25 Movidius Limited Variable-length instruction buffer management
US11768689B2 (en) 2013-08-08 2023-09-26 Movidius Limited Apparatus, systems, and methods for low power computational imaging
US11579872B2 (en) 2013-08-08 2023-02-14 Movidius Limited Variable-length instruction buffer management
US9891936B2 (en) 2013-09-27 2018-02-13 Intel Corporation Method and apparatus for page-level monitoring
US20150186168A1 (en) * 2013-12-30 2015-07-02 Unisys Corporation Dedicating processing resources to just-in-time compilers and instruction processors in a dynamic translator
US9513805B2 (en) * 2014-04-15 2016-12-06 International Business Machines Corporation Page table including data fetch width indicator
US20150293855A1 (en) * 2014-04-15 2015-10-15 International Business Machines Corporation Page table including data fetch width indicator
US9524100B2 (en) * 2014-04-15 2016-12-20 International Business Machines Corporation Page table including data fetch width indicator
US9672351B2 (en) 2015-02-02 2017-06-06 Qualcomm Incorporated Authenticated control stacks
WO2016126523A3 (en) * 2015-02-02 2016-09-29 Qualcomm Incorporated Authenticated control stacks
CN107209826A (en) * 2015-02-02 2017-09-26 高通股份有限公司 Certified control storehouse
US11327779B2 (en) * 2015-03-25 2022-05-10 Vmware, Inc. Parallelized virtual machine configuration
US20160283257A1 (en) * 2015-03-25 2016-09-29 Vmware, Inc. Parallelized virtual machine configuration
EP3136228B1 (en) * 2015-08-24 2023-08-09 Glenfly Tech Co., Ltd. Methods for combining instructions and apparatuses having multiple data pipes
EP4258110A3 (en) * 2015-08-24 2024-01-03 Glenfly Tech Co., Ltd. Methods for combining instructions and apparatuses having multiple data pipes
US11106463B2 (en) * 2019-05-24 2021-08-31 Texas Instruments Incorporated System and method for addressing data in memory
US20200371803A1 (en) * 2019-05-24 2020-11-26 Texas Instruments Incorporated System and method for addressing data in memory
US11836494B2 (en) 2019-05-24 2023-12-05 Texas Instruments Incorporated System and method for addressing data in memory

Also Published As

Publication number Publication date
US20030200351A1 (en) 2003-10-23
KR100584964B1 (en) 2006-05-29
WO1997027539A1 (en) 1997-07-31
US6961843B2 (en) 2005-11-01
JP3634379B2 (en) 2005-03-30
US20030115238A1 (en) 2003-06-19
US20050267996A1 (en) 2005-12-01
EP0976034A1 (en) 2000-02-02
DE69734399D1 (en) 2006-03-02
JP2000513464A (en) 2000-10-10
US6950923B2 (en) 2005-09-27
EP0976034B1 (en) 2005-10-19
KR20050052529A (en) 2005-06-02
US6532531B1 (en) 2003-03-11

Similar Documents

Publication Publication Date Title
US6532531B1 (en) Method frame storage using multiple memory circuits
EP0976030B1 (en) Instruction folding for a stack-based machine
KR100513138B1 (en) A processor for executing instruction sets received from a network or from a local memory
KR100466722B1 (en) An array bounds checking method and apparatus, and computer system including this
US6038643A (en) Stack management unit and method for a processor having a stack
US6076141A (en) Look-up switch accelerator and method of operating same
US6148391A (en) System for simultaneously accessing one or more stack elements by multiple functional units using real stack addresses
US5970242A (en) Replicating code to eliminate a level of indirection during execution of an object oriented computer program
US6065108A (en) Non-quick instruction accelerator including instruction identifier and data set storage and method of implementing same
US20060026370A1 (en) Method and system for accessing indirect memories
Yellin et al. The java virtual machine specification
KR100618718B1 (en) Caching Methods and Devices in Stacked Memory Structures

Legal Events

Date Code Title Description
STCB Information on status: application discontinuation

Free format text: ABANDONED -- FAILURE TO RESPOND TO AN OFFICE ACTION