> Interesting but what this has todo with the
> Logger change? There
> are no volatile fields, transient yes...
>
> --jason
I'm looking for a rock to crawl back under :-)
Regards,
Adrian
_
View thread online: http://main.jboss.org/t
I'm beginning to doubt my sanity.
This is the second time this week!
I tried this about 5 times in different orders and
changed the test slightly (included the preprocess)
I just swapped it back and yes, it is the JIT!
Sorry to waste your time. :-(
Regards,
Adrian
> Tried it all ways :-)
>
>
Interesting but what this has todo with the Logger change? There
are no volatile fields, transient yes...
--jason
Adrian Brock wrote:
>Weird,
>
>I expected the volatile to make it run slower.
>The JVM has to check the reference to Category
>on every operation.
>
>hprof says this is true,
Tried it all ways :-)
Regards,
Adrian
> >public static void main(String[] args)
> >{
> >
> org.apache.log4j.BasicConfigurator.configure();
> >
> > testWrapper();
> > testVolatile();
> >}
>
>
> What if you run the volatile test first and then the
> wrapper test?
>public static void main(String[] args)
>{
> org.apache.log4j.BasicConfigurator.configure();
>
> testWrapper();
> testVolatile();
>}
What if you run the volatile test first and then the wrapper test?
Maybe you are seeing jit stuff. I suggest running both, throwin
Weird,
I expected the volatile to make it run slower.
The JVM has to check the reference to Category
on every operation.
hprof says this is true, but for some reason
the method invoked through the Category takes less
time?
[pre]
CPU TIME (ms) BEGIN (total = 41544) Wed Feb 27 13:42:01 2002
rank
I did not mean to check this is. It is a serializable Logger. I sent
email before about if there were any objections and did not hear any.
Should I leave it or rollback?
--jason
___
Jboss-development mailing list
[EMAIL PROTECTED]
https://lists.s
