Re: Relieving some ClassLoader contention in PreferredClassProvider

Hi Dan,

Dan Creswell wrote:

> Unless we're going to define a set of platforms across which to make
> judgements/perform benchmarks, I don't see how to get a useful indication.

I indicated there is no official benchmark so I'm well aware of all the
things you mentioned. But I can hardly belief this change went in
without doing any benchmarking to see whether it didn't make things
worse for some typical use cases.

I left it up to the person(s) responsible for this improvement (?) to
share with us what he knows, dismissing such a question on forehand and
to keep myself wondering about the positiveness of this change or to
spend a significant amount of time on something that might be lying on
somebody's desk seems less effective.

Also I think that if you make improvements to Blitz in the concurrency
area you will test to get a feeling of the perceived improvement. And of
course you should make the standard disclaimer that everybody must check
on its typical scenario/hardware/OS, etc. whether he or she gets the
same/less or better numbers.
--
Mark

--------------------------------------------------------------------------
Getting Started:     http://www.jini.org/wiki/Category:Getting_Started
Community Web Site:  http://jini.org
jini-users Archive:  http://archives.java.sun.com/archives/jini-users.html
Unsubscribing:       email "signoff JINI-USERS"  to [email protected]