tomcat-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Casey Lucas" <clu...@armassolutions.com>
Subject RE: first cut at tag pooling implementation
Date Wed, 14 Mar 2001 06:07:39 GMT

> > Great!
> > 
> > You will do before/after performance comparisons?
> 
> In the process...

Ok, here are some basic numbers.

For simple performance testing I wrote a jsp that
uses an outer tag (BodyTag iteration style) and an
inner tag.  I placed timings around the outer tag.
Based on 3.3 (nightly from a couple of days ago)
and running 10000 iterations, over 10 runs each
from one browser...

tag processing took avg of:
pre tag pooling: 7362 milliseconds
tag pooling patches (pooling enabled): 1586 milli
tag pooling patches (pooling disabled): 6494 milli

Yes, I know most pages don't have 10000 tags and my
tests are far from rigorous, but I just wanted to
get a rough estimation.

I found it interesting that the patched (but pooling
disabled) runs were faster than the originals.  The
only thing that I can think of that might be different
is that the old code had some references to a static
variable that I removed.  I'm not sure if that could/
would make a difference.  Maybe there was something
else going on.

> 
> > 
> > Have you tested your changes with Watchdog?
> > 
> 

Well I tried to run the patches with watchdog and became
worried when some of the tests failed.  But the ones
that failed appeared to be the same ones that failed with
the originals.  Are the nightly builds supposed to pass
watchdog tests?

Hopefully the 3.3m2 stuff will work cleanly with watchdog
and I can test against that.

-Casey

---------------------------------------------------------------------
To unsubscribe, e-mail: tomcat-dev-unsubscribe@jakarta.apache.org
For additional commands, email: tomcat-dev-help@jakarta.apache.org


Mime
View raw message