CN103136029A - Real-time compiling system self-adapting adjusting and optimizing method - Google Patents
Real-time compiling system self-adapting adjusting and optimizing method Download PDFInfo
- Publication number
- CN103136029A CN103136029A CN2013100791291A CN201310079129A CN103136029A CN 103136029 A CN103136029 A CN 103136029A CN 2013100791291 A CN2013100791291 A CN 2013100791291A CN 201310079129 A CN201310079129 A CN 201310079129A CN 103136029 A CN103136029 A CN 103136029A
- Authority
- CN
- China
- Prior art keywords
- information
- time system
- time
- resources bank
- adaption
- 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.)
- Pending
Links
Images
Abstract
The invention provides a real-time compiling system self-adapting adjusting and optimizing method which comprises the following steps: providing a resource library which is across application programs and is used repeatedly, wherein information in the resource library is separated to each running instance, and therefore the informations of all running instances in the resource library are not mutually affected; judging if the current running instance needs to gather relevant information of a real-time compiling process in an on-line mode, wherein a real-time compiling system writes in untreated runtime information to the resource library through a runtime information output, and / or reading the runtime information which is run across the application programs from the resource library; statically analyzing the gathered untreated runtime information in an off-line mode in the resource library to form a precalculated online running strategy; reading the precalculated online running strategy which is read and calculated from the resource library through a runtime optimizing strategy; confirming an optimization method of the real-time compiling system according to the precalculated online running strategy; and running application programs in an on-line mode combining an existing self-adapting optimizing system according to the confirmed optimization method.
Description
Technical field
The present invention relates to field of computer technology, more particularly, the present invention relates to a kind of Just-In-Time system self-adaption tuning method.
Background technology
In Java programming language and environment, Just-In-Time (Just-In-Time compiler, just-in-time compiler) system is one the bytecode of Java (program that comprises the instruction that need to be explained) is converted to the program that can directly send to the instruction of processor.After writing what a java applet, the statement of source language will be compiled into bytecode by the Java compiler, rather than be compiled into the instruction code corresponding with certain specific processor hardware platform (such as, the Pentium microprocessor of Intel or the System/390 processor of IBM).Bytecode is the code that is independent of platform that can send to any platform and can move on that platform.
Environmental Support when the Just-In-Time optimization system generally has dynamic operation operates in the Java Virtual Machine environment as java applet.Compare with the program of traditional static state compiling, Just-In-Time optimization is all when occurring in the program operation, information in the time of accessing the complicated running orbit of program and overall situation operation, and these information have been brought more multimachine meeting to the compile optimization of Just-In-Time system.
The information when program of dynamic perfromance has the operation that static compiling do not have, characteristic when utilizing compiling and optimization all to occur in the program operation, optimisation technique when adopting the overall self-adapting operation which kind of optimization comparatively complicated tutorial program carry out, collect on-the-flier compiler information and performance monitoring information, obtain better to carry out efficient, higher performance.
But, existing Just-In-Time system ubiquity following defective: on the one hand, the information that Just-In-Time relies on is all that monitoring is collected when program is moved, but these information will directly be dropped after program is complete, when program is carried out again next time, Just-In-Time will be monitored the information that run-time optimizing needs of collecting again again, during the operation of collection in the past, information can not be effectively utilized, and has wasted existing information and has improved the chance that improves program feature.Because the Just-In-Time local code need to take program runtime, compile out the higher code of degree of optimization, the time that spends may be longer; And want to compile out the higher code of degree of optimization, and information when interpreter wants help compiler monitoring collection operation, this is also influential to explaining speed of carrying out.
On the other hand, the Just-In-Time system all can provide considerable optimization characteristics and parameter regulating measure.These options impact effect to performance on different platform is also different, many users often can be at a loss as to what to do in the face of so many optional parameter, the optimization option that domestic consumer uses is usually in 3, the optimization option that advanced level user uses is usually in 5, seldom can use more than 8, seldom the user is ready to attempt different optimization options, and more fine-grained optimal control hardly may.
Summary of the invention
Technical matters to be solved by this invention is for there being defects in prior art, provide a kind of in conjunction with when operation dynamic collecting informations and the self-adaptation tuning method of static information off-line, in order to further improve the Just-In-Time system self-adaption tuning method of the performance of Just-In-Time system operation.
According to the present invention, a kind of Just-In-Time system self-adaption tuning method is provided, and it comprises: provide one across the nonexpondable resources bank of application program, the wherein information in resources bank, separate for each running example, the information of each running example in resources bank is independent of each other thus; Off-line, analyze when operation information of being untreated collected in resources bank to form the on-line operation strategy of precomputation statically; Read the on-line operation strategy that reads the precomputation of calculating from resources bank by the run-time optimizing strategy, and the on-line operation strategy of calculating is on the estimation determined the optimization method of Just-In-Time system; Optimization method according to determining runs application online in conjunction with original adaptive optimal system.
Preferably, described Just-In-Time system self-adaption tuning method also comprises: judge whether current running example needs to collect online the relevant information of Just-In-Time process, and, the Just-In-Time system during by operation information output write information when being untreated operation in the resources bank.
Preferably, described Just-In-Time system self-adaption tuning method also comprises: judge whether current running example needs to collect online the relevant information of Just-In-Time process, and, information when the Just-In-Time system reads across the operation of application program operation from resources bank.
Preferably, described Just-In-Time system self-adaption tuning method also comprises: termly the information in resources bank is cleared up.
Preferably, during the precomputation on-line operation, strategy comprises: which takes optimize option, when compile certain by the use of thermal means, and inline or forbid inline ad hoc approach clearly.
Preferably, described Just-In-Time system self-adaption tuning method is used for the focus method of Java Virtual Machine.
Preferably, the relevant information of Just-In-Time process comprises: the number of times that method is carried out, the expense that when joins compiling formation, method compiling needs, the income that the optimization execution obtains, the objectives of virtual method call.
In Just-In-Time system self-adaption tuning method of being association of activity and inertia disclosed by the invention, take full advantage of the information of collected offline, the performance bottleneck of auto-analyzer procedure operation, the expense of Dynamic Information Gathering when having reduced some focus method operations, in advance these methods are carried out high performance optimization, the waste problem of the performance information data of having collected in process when having solved the operation of existing Just-In-Time system has reached the effect that further improves and optimizates application program efficient.
Description of drawings
By reference to the accompanying drawings, and by with reference to following detailed description, will more easily to the present invention, more complete understanding be arranged and more easily understand its advantage of following and feature, wherein:
Fig. 1 schematically shows the configuration schematic diagram of Just-In-Time system self-adaption tuning method according to the preferred embodiment of the invention.
Fig. 2 schematically shows the process flow diagram of Just-In-Time system self-adaption tuning method according to the preferred embodiment of the invention.
Fig. 3 schematically shows the process flow diagram of a concrete example of Just-In-Time system self-adaption tuning method according to the preferred embodiment of the invention.
Need to prove, accompanying drawing is used for explanation the present invention, and unrestricted the present invention.Note, the accompanying drawing of expression structure may not be to draw in proportion.And in accompanying drawing, identical or similar element indicates identical or similar label.
Embodiment
In order to make content of the present invention more clear and understandable, below in conjunction with specific embodiments and the drawings, content of the present invention is described in detail.
The invention provides a kind of self-adaptation tuning method of being association of activity and inertia, this framework considers " cost benefit " problem and the already present virtual machine adaptive optimal system of Just-In-Time system optimization process, the information of successfully off-line being obtained and online information of collecting combine, and then instruct the focus program whether can optimize in advance, adopt which kind of optimization means, when compile, whether the short and small method that guidance is called needs inline etc. immediately, thereby makes the adaptive optimal system of Just-In-Time obtain higher performance.
By the self-adaptation tuning framework of being association of activity and inertia, multidate information when multidate information and online operation during in the process of implementation in conjunction with the off-line operation collected across program operation, instruct adaptively application program more early to implement corresponding effectively optimizing, thereby choose best path optimizing and method for virtual machine quickly, the operational efficiency of the self-adaptation tuning of raising program operation.
The below will specifically describe the preferred embodiments of the present invention.
Fig. 2 schematically shows the process flow diagram of Just-In-Time system self-adaption tuning method according to the preferred embodiment of the invention.
As shown in Figure 2, Just-In-Time system self-adaption tuning method comprises according to the preferred embodiment of the invention:
First step S11: provide one across the nonexpondable resources bank 100 of application program.
Particularly, resources bank 100 can repeatedly use across application program, and the information of collecting between the assurance application program can not wasted.
Second step S12: judge whether current running example needs to collect online the relevant information of Just-In-Time process.And when the Just-In-Time system can be as required by operation, information output 10 writes information 101 when being untreated operation in the resources bank 100, and information can read as required operation across the application program operation from resources bank 100 time.
Information in resources bank 100 is separated for each running example; The information of each running example in resources bank 100 is independent of each other.
When writing or reading operation in the resources bank 100 from resources bank 100, information is optional, and special situation can be left in the basket, and described second step S12 is optional thus.
Third step S13: off-line, analyze when operation information 101 of being untreated collected in resources bank 100 to form the on-line operation strategy 102 of precomputation statically.
Particularly, the information 101 that is untreated during operation of exporting from the Just-In-Time system is enriched, so as when the resource analysis phase analysis effectively to move information.
And resource analysis is off-line analysis, does not take the expense of Just-In-Time system.
The 4th step S14: read 20 by the run-time optimizing strategy and read the on-line operation strategy 102 of the precomputation that third step S13 calculates from resources bank 100, and the on-line operation strategy 102 of calculating on the estimation, determine the optimization method of Just-In-Time system.
Particularly, what during the precomputation on-line operation, strategy 102 offered the Just-In-Time system is short and small and clear and definite optimisation strategy, which takes optimize option, when compile certain by the use of thermal means such as comprising, inline or forbid inline certain (short and small) method etc. clearly.
The 5th step S15: the optimization method according to the 4th step S14 determines runs application online.
Preferably, can clear up the information in resources bank 100 termly, thus the garbage in removing resources bank 100.
When the self-adaptation tuning framework of being association of activity and inertia that the above embodiment of the present invention proposes does not affect Just-In-Time system on-line operation, the continuation of information is not collected.And the self-adaptation tuning method of the dynamic bind that the above embodiment of the present invention proposes is improved the performance that has improved the program adaptive optimization significantly.
Fig. 3 schematically shows the process flow diagram of a concrete example of Just-In-Time system self-adaption tuning method according to the preferred embodiment of the invention.Wherein survey as example take the focus method of Java Virtual Machine Just-In-Time system self-adaption tuning method has been described.
The compiler object of Java Virtual Machine calls the number of times that the counters count method is carried out when generally carrying out by explanation, carry out when surpassing a definite threshold values when method, will submit to instant compiler the requirement for compiler an of the method to.In the Java Virtual Machine implementation, if can move the Information Statistics of focus method last time in resources bank according to program, the operation of lower secondary program determines that according to the information of adding up in resources bank the method carries out as early as possible, and the statistical information process of the method will be saved.If can determine a plurality of clear and definite Compilation Methods, the efficient of program execution will improve greatly.The inline reason of method is like this equally, if some can be needed inline code Java method shorter, that the execution frequency is higher inline as early as possible, can cause the inline method of method inverse optimization to be forbidden those inline, must also can improve the performance of Java Virtual Machine.
Concrete flowage structure as shown in Figure 3, comprising:
Initially, as mentioned above, provide one across the nonexpondable resources bank 100 of application program.Particularly, resources bank 100 can repeatedly use across application program, and the information of collecting between the assurance application program can not wasted.Carry out subsequently following step.
Step S21: information when carrying out the operation of at least one time, and the information of these collections is merged store being untreated during operation in information 101 of resources bank 100 into.
Particularly, in program process, information (relevant information of Just-In-Time process) during the operation of exporting correlation technique by virtual machine according to the different needs of optimizing direction, the number of times of carrying out as: method, when join expense that compiling formation, method compiling need, optimize and carry out the income that obtains, the objectives of virtual method call etc., and the information of these collections is merged store in resources bank.
Step S22: static and analyze off-line resource information bank.Specifically, off-line, analyze when operation information 101 of being untreated collected in resources bank 100 to form the on-line operation strategy 102 of precomputation statically.
Particularly, the resource information analysis is based on cost-benefit model and builds on-line operation strategy (the on-line operation strategy 102 of precomputation), and it is static state and off-line, does not take the expense of application program operation.Continuation instructs strategy as example take the compiling of focus method opportunity.Determining of Java Virtual Machine compiler object is generally the execution number of times that calls the counters count method when carrying out by explanations, during over a definite threshold values, will submit to instant compiler the requirement for compiler an of the method when the method execution to.In the Java Virtual Machine implementation, if in the time of can moving last time according to program, the Information Statistics of focus method are in resources bank, the operation of lower secondary program determines that according to the information of adding up in resources bank the method carries out as early as possible, and the Information Statistics process of the method can be saved in operational process.If can determine a plurality of clear and definite Compilation Methods, the efficient of program execution will improve greatly.
To determine as early as possible Compilation Method, the expense of saving the collection of correlation technique information monitoring is purpose, and based on Java Virtual Machine, the present invention has designed and Implemented a cover off-line selection algorithm.
At first the focus method based on compiling builds a model, and it is based on three data:
These three data are based on the data that off-line test is estimated.Wherein, Speedup (M) is that the data estimation of mobile phone in a basis method operational process comes, and estimation equation is as follows:
speedup=(N
I-N
N)*(T
I/T
N)-C
T
In this formula, N
1And N
NRefer to that respectively certain method is at the number of times of interpretive scheme and pattern compiler execution, T
1And N
NRefer to that respectively certain method carries out the averaging time of each needs under interpretive scheme and pattern compiler, and C
TRepresent the compiling expense of the method.
Based on the quantity of these definition, in the time of just can being untreated operation according to the complexity of collecting, when information judgement focus method carries out compile optimization, and the method is described below.In this arthmetic statement, Timeall is the execution time of representation program execution pre-estimation.S1, s2 and s3 are the less constant empirical values that pre-sets, and be used for to control the quantity of method to be compiled in the compiling formation, because if the compiling expense is overweight, its income of bringing will lose more than gain.The Threshold that uses in algorithm triggers focus method Just-In-Time and the default threshold that arranges in Java Virtual Machine.
Step S23: determine clear and definite and short and small on-line optimization strategy.
It is short and small and clear and definite that the self-adaptation tuning of being association of activity and inertia particularly, is used for instructing the precomputation on-line operation strategy of Java Virtual Machine optimization.For example, be used for instructing and carry out which kind of optimization, when be optimized.Instruct the tactful cocoa when the focus method is optimized to be divided into 3 kinds: normal makes a comment or criticism and just compiles when often reaching counter threshold, earlycompile and exclude represent respectively just-ahead-of-time compilation and do not need to compile, when also having further the nation method of the current compilings of control and display such as reinterpret, recompile that uncommon_trap occurs, reselect according to additional information and compile immediately or will again carry out adaptive optimization according to online information.Optimisation strategy can also instruct the storehouse of different target machine to distribute; The method concrete inline which virtual of guidance method calls etc.
Step S24: preferably, termly the information in resources bank 100 is cleared up, thus the garbage in removing resources bank 100.
Particularly, in the self-adaptation tuning framework of being association of activity and inertia, the information of resources bank and the on-line operation strategy of calculating move under different machines or different time, and the information that needs to collect may change.Therefore, old, the static information that are of little use in resources bank need regularly to remove, and prevent that resources bank from becoming too fat to move.
In Just-In-Time system self-adaption tuning method of being association of activity and inertia disclosed by the invention, take full advantage of the information of collected offline, the performance bottleneck of auto-analyzer procedure operation, the expense of Dynamic Information Gathering when having reduced some focus method operations, in advance these methods are carried out high performance optimization, the waste problem of the performance information data of having collected in process when having solved the operation of existing Just-In-Time system has reached the effect that further improves and optimizates application program efficient.
In addition, need to prove, unless otherwise indicated, otherwise the term in instructions " first ", " second ", " the 3rd " etc. describe each assembly of only being used for distinguishing instructions, element, step etc., rather than are used for logical relation between each assembly of expression, element, step or ordinal relation etc.
Be understandable that, although the present invention with the preferred embodiment disclosure as above, yet above-described embodiment is not to limit the present invention.For any those of ordinary skill in the art, do not breaking away from technical solution of the present invention scope situation, all can utilize the technology contents of above-mentioned announcement to make many possible changes and modification to technical solution of the present invention, or be revised as the equivalent embodiment of equivalent variations.Therefore, every content that does not break away from technical solution of the present invention, all still belongs in the scope of technical solution of the present invention protection any simple modification made for any of the above embodiments, equivalent variations and modification according to technical spirit of the present invention.
Claims (7)
1. Just-In-Time system self-adaption tuning method is characterized in that comprising:
Provide one across the nonexpondable resources bank of application program, wherein the information in resources bank, separate for each running example, and the information of each running example in resources bank is independent of each other thus;
Off-line, analyze when operation information of being untreated collected in resources bank to form the on-line operation strategy of precomputation statically;
Read the on-line operation strategy that reads the precomputation of calculating from resources bank by the run-time optimizing strategy, and the on-line operation strategy of calculating is on the estimation determined the optimization method of Just-In-Time system;
Optimization method according to determining runs application online in conjunction with original adaptive optimal system.
2. Just-In-Time system self-adaption tuning method according to claim 1, characterized by further comprising: judge whether current running example needs to collect online the relevant information of Just-In-Time process, and, the Just-In-Time system during by operation information output write information when being untreated operation in the resources bank.
3. Just-In-Time system self-adaption tuning method according to claim 1 and 2, characterized by further comprising: judge whether current running example needs to collect online the relevant information of Just-In-Time process, and, information when the Just-In-Time system reads across the operation of application program operation from resources bank.
4. Just-In-Time system self-adaption tuning method according to claim 1 and 2, characterized by further comprising: termly the information in resources bank is cleared up.
5. Just-In-Time system self-adaption tuning method according to claim 1 and 2, it is characterized in that, during the precomputation on-line operation, strategy comprises: which takes optimize option, when compile certain by the use of thermal means, and inline or forbid inline ad hoc approach clearly.
6. Just-In-Time system self-adaption tuning method according to claim 2, is characterized in that, described Just-In-Time system self-adaption tuning method is used for the focus method of Java Virtual Machine.
7. Just-In-Time system self-adaption tuning method according to claim 6, it is characterized in that, the relevant information of Just-In-Time process comprises: the number of times that method is carried out, the expense that when joins compiling formation, method compiling needs, the income that the optimization execution obtains, the objectives of virtual method call.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN2013100791291A CN103136029A (en) | 2013-03-12 | 2013-03-12 | Real-time compiling system self-adapting adjusting and optimizing method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN2013100791291A CN103136029A (en) | 2013-03-12 | 2013-03-12 | Real-time compiling system self-adapting adjusting and optimizing method |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN103136029A true CN103136029A (en) | 2013-06-05 |
Family
ID=48495895
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN2013100791291A Pending CN103136029A (en) | 2013-03-12 | 2013-03-12 | Real-time compiling system self-adapting adjusting and optimizing method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN103136029A (en) |
Cited By (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104461711A (en) * | 2014-12-15 | 2015-03-25 | 北京奇虎科技有限公司 | Adaptive optimization method and adaptive optimization device of computing equipment |
| CN105589729A (en) * | 2015-12-28 | 2016-05-18 | 北京锐安科技有限公司 | Dynamic compiling method and device based on embedded virtual machine |
| CN105786586A (en) * | 2014-12-23 | 2016-07-20 | 龙芯中科技术有限公司 | Hot method identification method and apparatus |
| CN106325964A (en) * | 2015-06-18 | 2017-01-11 | 龙芯中科技术有限公司 | Dynamic compiling and scheduling method and device |
| CN108446119A (en) * | 2017-12-28 | 2018-08-24 | 北京奇虎科技有限公司 | Inline control method and device |
| CN109240793A (en) * | 2017-05-16 | 2019-01-18 | 龙芯中科技术有限公司 | Recognition methods, device, electronic equipment and the storage medium of hot-spots |
| CN109542443A (en) * | 2017-07-27 | 2019-03-29 | 阿里巴巴集团控股有限公司 | Compilation Method and device, terminal, the data processing method of application program |
| CN111176654A (en) * | 2019-11-18 | 2020-05-19 | 浙江大学 | Internet of things application online compiling method based on multi-user cache |
| CN112433706A (en) * | 2020-11-27 | 2021-03-02 | 海光信息技术股份有限公司 | Compiling option tuning method and device, processor chip and server |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US6971091B1 (en) * | 2000-11-01 | 2005-11-29 | International Business Machines Corporation | System and method for adaptively optimizing program execution by sampling at selected program points |
| CN1922575A (en) * | 2004-02-20 | 2007-02-28 | 英特尔公司 | Methods and apparatus to optimize application program interfaces in a virtual machine environment |
| US20100241600A1 (en) * | 2009-03-20 | 2010-09-23 | Nokia Corporation | Method, Apparatus and Computer Program Product for an Instruction Predictor for a Virtual Machine |
-
2013
- 2013-03-12 CN CN2013100791291A patent/CN103136029A/en active Pending
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US6971091B1 (en) * | 2000-11-01 | 2005-11-29 | International Business Machines Corporation | System and method for adaptively optimizing program execution by sampling at selected program points |
| CN1922575A (en) * | 2004-02-20 | 2007-02-28 | 英特尔公司 | Methods and apparatus to optimize application program interfaces in a virtual machine environment |
| US20100241600A1 (en) * | 2009-03-20 | 2010-09-23 | Nokia Corporation | Method, Apparatus and Computer Program Product for an Instruction Predictor for a Virtual Machine |
Non-Patent Citations (3)
| Title |
|---|
| CHANDRA KRINTZ: "Coupling on-line and off-line profile information to improve program performance", 《CODE GENERATION AND OPTIMIZATION》 * |
| MATTHEW ARNOLD等: "Adaptive Optimization in the Jalapeno JVM", 《PROCEEDINGS OF THE 15TH ACM SIGPLAN CONFERENCE ON OBJECT-ORIENTED PROGRAMMING,SYSTEMS,LANGUAGES, AND APPLICATIONS》 * |
| ZHANG: "JVM基础之The Java HotSpot Performance Engine Architecture", 《HTTP://SISHUOK.COM/FORUM/BLOGPOST/LIST/338.HTML》 * |
Cited By (14)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104461711B (en) * | 2014-12-15 | 2017-05-17 | 北京奇虎科技有限公司 | Adaptive optimization method and adaptive optimization device of computing equipment |
| CN104461711A (en) * | 2014-12-15 | 2015-03-25 | 北京奇虎科技有限公司 | Adaptive optimization method and adaptive optimization device of computing equipment |
| CN105786586B (en) * | 2014-12-23 | 2018-10-30 | 龙芯中科技术有限公司 | Hotspot approach recognition methods and device |
| CN105786586A (en) * | 2014-12-23 | 2016-07-20 | 龙芯中科技术有限公司 | Hot method identification method and apparatus |
| CN106325964A (en) * | 2015-06-18 | 2017-01-11 | 龙芯中科技术有限公司 | Dynamic compiling and scheduling method and device |
| CN106325964B (en) * | 2015-06-18 | 2019-09-27 | 龙芯中科技术有限公司 | On-the-flier compiler dispatching method and device |
| CN105589729A (en) * | 2015-12-28 | 2016-05-18 | 北京锐安科技有限公司 | Dynamic compiling method and device based on embedded virtual machine |
| CN109240793A (en) * | 2017-05-16 | 2019-01-18 | 龙芯中科技术有限公司 | Recognition methods, device, electronic equipment and the storage medium of hot-spots |
| CN109542443A (en) * | 2017-07-27 | 2019-03-29 | 阿里巴巴集团控股有限公司 | Compilation Method and device, terminal, the data processing method of application program |
| CN108446119A (en) * | 2017-12-28 | 2018-08-24 | 北京奇虎科技有限公司 | Inline control method and device |
| CN111176654A (en) * | 2019-11-18 | 2020-05-19 | 浙江大学 | Internet of things application online compiling method based on multi-user cache |
| CN111176654B (en) * | 2019-11-18 | 2021-04-27 | 浙江大学 | Internet of things application online compiling method based on multi-user cache |
| CN112433706A (en) * | 2020-11-27 | 2021-03-02 | 海光信息技术股份有限公司 | Compiling option tuning method and device, processor chip and server |
| CN112433706B (en) * | 2020-11-27 | 2023-03-14 | 海光信息技术股份有限公司 | Compiling option tuning method and device, processor chip and server |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN103136029A (en) | Real-time compiling system self-adapting adjusting and optimizing method | |
| US9043788B2 (en) | Experiment manager for manycore systems | |
| KR101942518B1 (en) | Two pass automated application instrumentation | |
| US5787285A (en) | Apparatus and method for optimizing applications for multiple operational environments or modes | |
| Kwon et al. | Mantis: Automatic performance prediction for smartphone applications | |
| US20130080760A1 (en) | Execution Environment with Feedback Loop | |
| Hazelwood et al. | Adaptive online context-sensitive inlining | |
| US20150161385A1 (en) | Memory Management Parameters Derived from System Modeling | |
| CN100405294C (en) | System, method and program product to optimize code during run time | |
| Suganuma et al. | Design and evaluation of dynamic optimizations for a Java just-in-time compiler | |
| US20050166207A1 (en) | Self-optimizing computer system | |
| CN102063328B (en) | System for detecting interrupt-driven type program data competition | |
| US9128747B2 (en) | Methods and systems for optimizing the performance of software applications at runtime | |
| CN110543338A (en) | dynamic loading method and device for files | |
| Suganuma et al. | An Empirical Study of Method In-lining for a Java {Just-in-Time} Compiler | |
| US8056061B2 (en) | Data processing device and method using predesignated register | |
| Colombo et al. | poly Larva: runtime verification with configurable resource-aware monitoring boundaries | |
| CN102099786A (en) | Program optimization method | |
| CN104462943A (en) | Non-intrusive performance monitoring device and method for service system | |
| Arnold et al. | Improving virtual machine performance using a cross-run profile repository | |
| CN101395581A (en) | Optimised profile-driven compilation method for conditional code for a processor with predicated execution | |
| CN101493767A (en) | Pile pitching method of explicit releasing object in instant compiler-aid refuse collection | |
| Gu et al. | Phase-based adaptive recompilation in a JVM | |
| CN103019865B (en) | Virtual machine monitoring method and system | |
| US9417856B2 (en) | Efficient interpreter profiling to obtain accurate call-path information |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| C12 | Rejection of a patent application after its publication | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20130605 |