Alexander Yermolovich
Christian Wimmer
Michael Franz
ABSTRACT
Creating an interpreter is a simple and fast way to implement a dynamic programming language. With this ease also come major drawbacks. Interpreters are significantly slower than compiled machine code because they have a high dispatch overhead and cannot perform optimizations. To overcome these limitations, interpreters are commonly combined with just-in-time compilers to improve the overall performance. However, this means that a just-in-time compiler has to be implemented for each language.
We explore the approach of taking an interpreter of a dynamic language and running it on top of an optimizing trace-based virtual machine, i.e., we run a guest VM on top of a host VM. The host VM uses trace recording to observe the guest VM executing the application program. Each recorded trace represents a sequence of guest VM bytecodes corresponding to a given execution path through the application program. The host VM optimizes and compiles these traces to machine code, thus eliminating the need for a custom just-in-time compiler for the guest VM. The guest VM only needs to provide basic information about its interpreter loop to the host VM.
References
- Adobe Systems Inc. ActionScript 3 Language Specification, 2006. http://livedocs.adobe.com/specs/actionscript/3/wwhelp/wwhimpl/js/html/wwhelp.htm.Google Scholar
- Adobe Systems Inc. ActionScript Virtual Machine 2 Overview, 2007. http://www.adobe.com/devnet/actionscript/articles/avm2overview.pdf.Google Scholar
- Adobe Systems Inc. Adobe FlaCC, 2008. http://llvm.org/devmtg/2008-08/Petersen_FlashCCompiler.pdf.Google Scholar
- Adobe Systems Inc. Adobe Flex SDK, 2008. http://www.adobe.com/products/flex/flexdownloads/.Google Scholar
- Adobe Systems Inc. Adobe Alchemy, 2009. http://labs.adobe.com/technologies/alchemy/.Google Scholar
- V. Bala, E. Duesterwald, and S. Banerjia. Dynamo: A transparent dynamic optimization system. In Proceedings of the ACM SIGPLAN Conference on Programming Language Design and Implementation, pages 1--12. ACM Press, 2000. Google ScholarDigital Library
- C. F. Bolz, A. Cuni, M. Fijalkowski, and A. Rigo. Tracing the meta-level: PyPy's tracing JIT compiler. In Proceedings of the Workshop on the Implementation, Compilation, Optimization of Object-Oriented Languages and Programming Systems, pages 18--25. ACM Press, 2009. Google ScholarDigital Library
- M. Chang, E. Smith, R. Reitmaier, M. Bebenita, A. Gal, C. Wimmer, B. Eich, and M. Franz. Tracing for Web 3.0: Trace compilation for the next generation web applications. In Proceedings of the International Conference on Virtual Execution Environments, pages 71--80. ACM Press, 2009. Google ScholarDigital Library
- B. Chiles. CLR inside out: IronPython and the dynamic language runtime. MSDN Magazine, October 2007.Google Scholar
- R. Cytron, J. Ferrante, B. Rosen, M. Wegman, and F. Zadeck. Efficiently computing static single assignment form and the control dependence graph. ACM Transactions on Programming Languages and Systems, 13(4):451--490, 1991. Google ScholarDigital Library
- The Da Vinci Machine Project, 2008. http://openjdk.java.net/projects/mlvm/.Google Scholar
- Dynamic Language Runtime, 2008. http://www.codeplex.com/dlr/.Google Scholar
- ECMA. Standard ECMA-262: ECMAScript Language Specification, 3rd edition, 1999. http://www.ecma-international.org/publications/standards/Ecma-262.htm.Google Scholar
- A. Gal, B. Eich, M. Shaver, D. Anderson, D. Mandelin, M. R. Haghighat, B. Kaplan, G. Hoare, B. Zbarsky, J. Orendorff, J. Ruderman, E. Smith, R. Reitmaier, M. Bebenita, M. Chang, and M. Franz. Trace-based just-in-time type specialization for dynamic languages. In Proceedings of the ACM SIGPLAN Conference on Programming Language Design and Implementation, pages 465--478. ACM Press, 2009. Google ScholarDigital Library
- A. Gal, C. W. Probst, and M. Franz. HotpathVM: An effective JIT compiler for resource-constrained devices. In Proceedings of the International Conference on Virtual Execution Environments, pages 144--153. ACM Press, 2006. Google ScholarDigital Library
- J. Ha, M. R. Haghighat, S. Cong, and K. S. McKinley. A concurrent trace-based just-in-time compiler for JavaScript. Technical Report TR-09-06, University of Texas, Austin, 2009.Google Scholar
- R. Ierusalimschy, L. H. de Figueiredo, and W. C. Filho. Lua - an extensible extension language. Software: Practice and Experience, 26(6):635--652, 1996. Google ScholarDigital Library
- IronPython, 2009. http://www.codeplex.com/Wiki/View.aspx?ProjectName=IronPython.Google Scholar
- Java Specification Request 223: Scripting for the Java™ Platform, 2006. http://www.jcp.org/en/jsr/detail?id=223.Google Scholar
- The Jython Project, 2009. http://www.jython.org/Project/.Google Scholar
- The LLVM Compiler Infrastructure, 2009. http://llvm.org/.Google Scholar
- Lua Benchmarks, 2009. http://shootout.alioth.debian.org/gp4/lua.php.Google Scholar
- Lua Programming Language, 2009. http://www.lua.org/. Google ScholarDigital Library
- D. Mandelin. Tamarin Tracing Internals, Part I to V, 2008. http://blog.mozilla.com/dmandelin/2008/05/.Google Scholar
- PyPy, 2009. http://codespeak.net/pypy/.Google Scholar
- Scripting Engines for Java, 2008. https://scripting.dev.java.net/.Google Scholar
- G. T. Sullivan, D. L. Burening, I. Baron, T. Garnett, and S. Amarasinghe. Dynamic native optimization of interpreters. In Proceedings of the Workshop on Interpreters, Virtual Machines, and Emulators, pages 50--57. ACM Press, 2003. Google ScholarDigital Library
- R. Uhlig, G. Neiger, D. Rodgers, A. L. Santoni, F. C. M. Martins, A. V. Anderson, S. M. Bennett, A. Kagi, F. H. Leung, and L. Smith. Intel virtualization technology. Computer, 38(5):48--56, 2005. Google ScholarDigital Library
Index Terms
Optimization of dynamic languages using hierarchical layering of virtual machines
Reviews
R. Clayton
What happens when a host virtual machine (VM) that uses trace-driven just-in-time (JIT) compilation executes another VM, the guest VM__?__ How much do the programs that execute on the guest VM benefit from the optimizations on the host VM__?__ How can the host-guest interactions be altered to provide greater benefits to guest-executed programs__?__ In this paper, the authors address these questions. Trace-driven optimization focuses on frequently executed program sections. VMs use trace-driven optimization to restrict JIT compilation to program sections that are most likely to benefit from the expensive, complex optimization. Host VM benefits are delivered to the guest programs through the guest VM. The host VM uses a combined guest program-VM program counter to clarify loops in the guest program, and relies on the guest VM to identify immutable code regions, in order to enable more rigorous host VM optimizations. The implementation is a bit jury-rigged. The host trace-driven JIT VM executes ActionScript 3 (AS3). Since the guest VM for the Lua dynamic programming language was written in C, it had to be translated to AS3. The results, however, are impressive: "This approach leads to a considerable speedup ... when a well-established interpreter-based dynamic language is executed on top of a VM that already uses a trace-based JIT compiler." The paper is well written. While it is not necessary for readers to be familiar with the languages discussed, they should have basic interpreter knowledge. Online Computing Reviews Service
Access critical reviews of Computing literature here
Become a reviewer for Computing Reviews.
Comments