I already sent out a few e-mails on working on a concurrent logger and
possibility of performance gains using a different locking strategy. I
have been trying to reconcile if my design change had any impact on
performance. I felt that I wasn't measuring the impact of my changes
correctly.
Have any of you used OptimizeIt or JProbe (or any real code profiler) on
log4j 1.3?
I have spent some time working with the code on JBoss Profiler, and have
found some interesting things. (No, you don't need JBoss, it's just the
name.) I attached some example output.
Although, it's a bit rough around the edges, it gives accurate results
and has a pretty good interface. If any of you would like assistance on
setting it up, I would be glad to help.
I know the log4j developers have worked hard on optimizing time and
memory, but I think some of the optimizations may have been
"premature." In an effort to optimize some things, I think some of the
changes that worsen the overall cleanliness of the code base.
|
Processes executed from Sun Jan 23 17:42:07 UTC 2005 till Sun Jan 23 17:44:20 UTC 2005
|
|
---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]