Hi, My name is Wei Wu and I would like to participate in Mozilla during GSoC 2013 as a student. I have been studying IonMonkey for several days, but I haven't found any related ideas about it on Mozilla's idea list pages[1][2]. The following are some ideas I'm interested in and I want to know your opinions/comments:
1. Save Type Information for Continuous Optimizations. SpiderMonkey interprets JavaScript bytecodes in the first place and collects type information. IonMonkey will not be called until the type information is stable. If we save type information on disks and reuse it next time, it will shorten the interpreting time and enable further optimizations. ( This idea comes from [3]. ) 2. Cache compiled asm.js codes and IR to avoid runtime overhead. Mozilla proposed asm.js as a low-level, efficient target language for compilers[4]. Emscripten is the main producer of asm.js codes currently. The asm.js codes translated by Emscripten are not changed frequently, compare with typical JavaScript codes written by programmers, and it will never bailout. It may be feasible to store both IR and binary for load time speedup and further optimizations. 3. Improve the JIT Inspector. JIT Inspector is a very useful tool for profiling. But its UI is relatively simple and the interpretation of the results is not easy for beginners. I can improve the UI to make it more user-friendly. Also, I am glad to implement some features for it, e.g. saving results in a file, etc. Are there any mentors interested in one of the ideas and would like to mentor me during the summer? thanks. [1]: https://wiki.mozilla.org/Community:SummerOfCode13 [2]: https://wiki.mozilla.org/Community:SummerOfCode13:Brainstorming [3]: http://markmail.org/message/y6vbgptiull4w7yh [4]: http://asmjs.org -- Wei Wu _______________________________________________ dev-tech-js-engine-internals mailing list [email protected] https://lists.mozilla.org/listinfo/dev-tech-js-engine-internals
