Ralf Ertzinger wrote:
On Wed, Jan 04, 2006 at 03:41:38PM +0530, Rahul Sundaram wrote:
Thats pretty vague. Benchmarks, reproducible regressions, specific bug
reports etc is the way to move forward with this.
Yes, I know. Working on that. One thing I have established is that Clearlooks
is significantly slower than bluecurce (although I can not say whether
FC4-CL is faster/slower than RC-CL right now).
My other gripe is Xv performance, I am writing a benchmark to get to the
bottom of that one.
For fun, here are some gtkperf numbers for current RH, RH default theme
(mainly CL, I think) vs. bluecurve. Machine is a PII400, Number Nine T2R
(yay!). 100 gtkperf iterations. The Pixbufs numbers are bogus, ignore.
Ok So I created his page for tracking such efforts on a ongoing basis.
What we need a list of tools and how to produce these benchmarks and
document the results between various version of Fedora under different
settings. We can start by tracking down and avoiding regressions.
http://fedoraproject.org/wiki/Performance
If you do not have edit access to the wiki already see
http://fedoraproject.org/wiki/WikiEditing. Register in the
FirstnameLastname format (case sensitive for wiki linking) and send me a
offlist mail if you need me to add you to the edit group.
--
fedora-devel-list mailing list
fedora-devel-list@xxxxxxxxxx
https://www.redhat.com/mailman/listinfo/fedora-devel-list
