Re: Re: Re: Resulting class performance

I know.  Just saying, it's still tricky -- for example if the problem is
really the iconst/ldc difference it's not really going to make a major
difference to fix it, although it will seem like it for your test.

On Fri, May 7, 2010 at 1:01 PM, Jeremy Cowgar <[email protected]> wrote:

> On 5/7/2010 3:57 PM, Scott Vachalek wrote:
>
>> Look up "Java vs. C++ performance" to see all the problems that can occur
>> with this sort of micro-benchmark.  It could all come down to differences in
>> measurement (whether both VMs are running in the same mode, whether startup
>> time is counted, etc. etc.)  But off the top of my head, I would try
>> replacing that ldc with an iconst.  I'd be surprised if the JIT doesn't
>> produce equivalent code in both cases but that seems to be the most
>> prominent difference.
>>
>
> I understand micro benchmarks are bad and very deceiving but I also
> understand that I should be able to achieve the same performance as Java. If
> something as fundamental as my loop construct is wrong, it can have drastic
> effects on every part of the language and since one takes 3 seconds and the
> other 250 ms, I was thinking something is wrong. If one was 250 ms and the
> other 270 ms, I would have kept on chugging along.
>
> BTW... The difference between the existing Josl interpreter and the new JVM
> target was not my real concern. The concern was the Java compiled code was 2
> 3/4 seconds faster on a 3 second long test.
>
> Jeremy
>
>
>
> --
> You receive this message as a subscriber of the [email protected] mailing list.
> To unsubscribe: mailto:[email protected]
> For general help: mailto:[email protected]?subject=help
> OW2 mailing lists service home page: http://www.ow2.org/wws
>
>

-- 
You receive this message as a subscriber of the [email protected] mailing list.
To unsubscribe: mailto:[email protected]
For general help: mailto:[email protected]?subject=help
OW2 mailing lists service home page: http://www.ow2.org/wws