John Rose wrote: > For language runtimes the JVMs should supply more hooks for affecting > JIT-level inlining, maybe something as simple (in HotSpot's case) as > @sun.misc.Inline. I'm not aware of a credible effort to standardize > on it, though.
If I could simply tell hotspot to inline through all my dyncall plumbing, it would save me a world of pain trying to bring unique code all the way back to the original call site. Of course invokedynamic has better ways around this like TCO, but for now my plumbing totally defeats inlining (as discussed in Cliff Click's presentation last fall). In theory if I could tell Hotspot to immediately inline my "CachingCallSite" logic everywhere it was encountered, I'd be that much closer to getting the method handles to inline, wouldn't I? BTW, I did also try making CachingCallSite's logic static methods that manipulate a common data structure. It did not help. The problem I run into now trying to get dyncalls to inline is that it usually adds so much extra bytecode to the calling method (because the dyncall plumbing is complex) that it actually slows things down more than it speeds them up :( - Charlie --~--~---------~--~----~------------~-------~--~----~ You received this message because you are subscribed to the Google Groups "JVM Languages" group. To post to this group, send email to [email protected] To unsubscribe from this group, send email to [email protected] For more options, visit this group at http://groups.google.com/group/jvm-languages?hl=en -~----------~----~----~----~------~----~------~--~---
