Invokedynamic slower than reflection?

[John Rose]

Eric's test Hello.java works OK, but your test, Rémi, fails (Fib4.java).  So it is a second bug.

Stephen, I left a comment:   http://gist.github.com/500808

Gist of the gist is: -XX:+UnlockExperimentalVMOptions -XX:+EnableInvokeDynamic

-- John

On Jul 30, 2010, at 8:30 AM, Rémi Forax wrote:

> Le 30/07/2010 08:40, Chanwit Kaewkasi a écrit :
>> Hi Rémi,
>> 
>> I did freshly build mlvm and run Eric's test.
>> Here's my result:
>> 
>> Duration invokedynamic: 103
>> Duration reflection: 1463
>> 
>> The patch seems to be working for me as InvokeDynamic is clearly
>> faster than Reflection on my machine.
>> 
>> Cheers,
>> 
>> Chanwit
>>   
> 
> Hi Chanwit,
> It seems I'm not able to get a mlvm build right :)
> 
> Can you test with this snippet ?
> 
> java -server ... -XX:+PrintCompilation Fib4
> 
> and report if you see lot of lines like "blob made no re-entrant"
> indicating too much deopt.
> 
> cheers,
> Rémi
> <Fib4.java>_______________________________________________
> mlvm-dev mailing list
> mlvm-dev at openjdk.java.net
> http://mail.openjdk.java.net/mailman/listinfo/mlvm-dev