Re: another question about megamorphic call sites in combination with MethodHandles

Hi Jochen, At least with the current tip version of HotSpot, the mh.invokeWithArguments() callsite is not likely to get its actual target inlined into the caller; we depended a lot on being able to prove that the MethodHandle is a (JIT-)compiled-time constant to be able to do inlining across t

Re: Looking for comments on paper draft "DynaMate: Simplified and optimized invokedynamic dispatch"

So do I. Thanks, Kris On 2013/2/19 13:55, Remi Forax wrote: > On 02/19/2013 03:51 PM, Attila Szegedi wrote: >> Sure, I'd be interested in reading this. >> >> Thanks, >> Attila. > So am I. > > Rémi > >> On Feb 19, 2013, at 2:37 PM, Eric Bodden wrote: >> >>> Hi all. >>> >>> Kamil Erhard, a stu

Re: hotspot-comp OS X builds

Hi Duncan, If the JDK8 promotion goes well this week (still in progress now) then yes, all changes in Charlie's build (plus more bugfixes from GC and Runtime teams) will be in this week's JDK8 ea snapshot. - Kris On 2013/1/25 19:56, MacGregor, Duncan (GE Energy Management) wrote: > Can I just

Re: Instrumenting call sites.

Hi Aleksey, I don't think TraceTypeProfile works in this case. First, only invokevirtual and invokeinterface actually have type profiles; other invokes don't. You can see that from the interpreter code that only invokevirtual and invokeinterface do profile_virtual_call() (which keeps track of