John Cowan wrote:
> On 10/13/07, Miles Sabin <[EMAIL PROTECTED]> wrote:
> 
>> Actually, google does seem to know about it, and it's just the obvious
>> restriction to an inline cache for exactly one <class, method> pair.
> 
> Quite so, and the idea is that on highly pipelined modern CPUs,
> you don't want to pay the price for more than a single conditional
> branch.

JRuby's inline cache is currently monomorphic, and after removing a few 
roadblocks it appears that HotSpot has really picked that up and ran 
with it. However I have a PIC patch hanging around that improved 
polymorphic dispatch by almost 50%, and didn't appear to impact 
monomorphic inline caching perf at all. We will probably put it in place 
some time soon, along with a "hotness" measure to allow re-sorting the 
cache occasionally.

- 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
-~----------~----~----~----~------~----~------~--~---

Reply via email to