[prev in list] [next in list] [prev in thread] [next in thread] 

List:       kwin
Subject:    Re: [Kwin] kwin performance (kde 3.2b2)
From:       aviv bergman <aviv_brg () yahoo ! com>
Date:       2004-01-22 19:18:01
Message-ID: 20040122191801.52470.qmail () web13122 ! mail ! yahoo ! com
[Download RAW message or body]

just a quick update on my system:
running kernel 2.6.1 (though previously 2.4.22 felt
the same)
athlonXp 2500/geforce3

i've tested the (rate limiting) patch and i can't say
it worked for me - while the cpu usage was defenetly
lower - the movement was _very_ jerkey with the
default values.
when i tried higher values (30,60,125) the movement
was better/same as default.
i also tried both patches together, and i belive (but
cannot say for sure) that draging was not as smooth as
with my patch alone - probably because the timers
cannot be synced to the mouse interrupts.

based on your default values for rate limiting, i see
why you cannot see the problem - dragging with these
values was _very_ bad for me - the window and the
pointer should always be in total sync. since even on
low end computers (~500Mhz) i get better performance,
it might be that you have an underperforming X
driver/graphics card/unaccelerated server?

anyway, if you merge the rate limiting patch into
kwin, _please_ add a way to disable it - it make
performance worse for more modern systems (any modern
computer should be able to handle the 125 events/sec
from the usb mouse, and by modern computer i mean
1Ghz+, geforce+).

b.t.w. - i belive that one of the 
    delete moveresize_timer;
    moveresize_timer = NULL;
in finishMoveResize/leaveMoveResize is a leftover from
previous version of the code.

aviv




__________________________________
Do you Yahoo!?
Yahoo! SiteBuilder - Free web site building tool. Try it!
http://webhosting.yahoo.com/ps/sb/
_______________________________________________
Kwin mailing list
Kwin@kde.org
https://mail.kde.org/mailman/listinfo/kwin
[prev in list] [next in list] [prev in thread] [next in thread]