public inbox for linux-kernel@vger.kernel.org
 help / color / mirror / Atom feed
* Possible regression with 2.6.0-test6
@ 2003-09-29 10:52 Stephen Anthony
  2003-09-30  1:17 ` Nick Piggin
  2003-09-30  1:27 ` Joshua Kwan
  0 siblings, 2 replies; 3+ messages in thread
From: Stephen Anthony @ 2003-09-29 10:52 UTC (permalink / raw)
  To: linux-kernel

I've noticed that test6 isn't as 'smooth' as test5.  Specifically, if I 
open an nedit window and use the cursor down key to scroll down through 
text, there is some jerkiness.  At this point I've noticed that CPU use 
goes to 1% or more.  A similar problem happens with Opera (web browser), 
but of course the CPU usage is higher, which is expected.

When I go back to test5, scrolling in nedit is smoother, and CPU never 
goes above 0.6%.  Note that the speed that text moves is the same in 
both, but the test6 version 'jumps' and 'stops' for a little bit.  Hence 
my describing it as jerky.

I know its not much to go on, but it is a noticable difference, at least 
to me.  Since this new release has tweaked interactivity patches, I'm 
wondering if this has anything to do with it.

System is a fully updated Mandrake 9.1, but I don't think that would 
matter.  CPU is P4-2.4GHz with 768 MB RAM, and NVidia video card.

Steve

^ permalink raw reply	[flat|nested] 3+ messages in thread

* Re: Possible regression with 2.6.0-test6
  2003-09-29 10:52 Possible regression with 2.6.0-test6 Stephen Anthony
@ 2003-09-30  1:17 ` Nick Piggin
  2003-09-30  1:27 ` Joshua Kwan
  1 sibling, 0 replies; 3+ messages in thread
From: Nick Piggin @ 2003-09-30  1:17 UTC (permalink / raw)
  To: Stephen Anthony; +Cc: linux-kernel



Stephen Anthony wrote:

>I've noticed that test6 isn't as 'smooth' as test5.  Specifically, if I 
>open an nedit window and use the cursor down key to scroll down through 
>text, there is some jerkiness.  At this point I've noticed that CPU use 
>goes to 1% or more.  A similar problem happens with Opera (web browser), 
>but of course the CPU usage is higher, which is expected.
>
>When I go back to test5, scrolling in nedit is smoother, and CPU never 
>goes above 0.6%.  Note that the speed that text moves is the same in 
>both, but the test6 version 'jumps' and 'stops' for a little bit.  Hence 
>my describing it as jerky.
>
>I know its not much to go on, but it is a noticable difference, at least 
>to me.  Since this new release has tweaked interactivity patches, I'm 
>wondering if this has anything to do with it.
>
>System is a fully updated Mandrake 9.1, but I don't think that would 
>matter.  CPU is P4-2.4GHz with 768 MB RAM, and NVidia video card.
>

Hi Stephen, the new CPU scheduler is likely to be the change. You need
to first make sure your X server is not reniced to -10 (should be at 0).

If that fails to help, then you could try my scheduler (with X at -10!)
http://www.kerneltrap.org/~npiggin/v15a/sched-rollup-v15a-260t6.gz


^ permalink raw reply	[flat|nested] 3+ messages in thread

* Re: Possible regression with 2.6.0-test6
  2003-09-29 10:52 Possible regression with 2.6.0-test6 Stephen Anthony
  2003-09-30  1:17 ` Nick Piggin
@ 2003-09-30  1:27 ` Joshua Kwan
  1 sibling, 0 replies; 3+ messages in thread
From: Joshua Kwan @ 2003-09-30  1:27 UTC (permalink / raw)
  To: linux-kernel; +Cc: stephena

[-- Attachment #1: Type: text/plain, Size: 669 bytes --]

On Mon, Sep 29, 2003 at 08:22:09AM -0230, Stephen Anthony wrote:
> Since this new release has tweaked interactivity patches, I'm 
> wondering if this has anything to do with it.

Yes. Con Kolivas' interactivity patches that have been merged into
mainline mean that if Mandrake's XFree86 runs at a negative nice value,
this will affect interactivity. Please set it to 0.

FWIW for others reading this thread, in Debian it is a matter of running
'dpkg-reconfigure xserver-common'.

(As I was writing this message I realized that Nick Piggin had already
answered, but because of the little Debian factoid I'm deciding to send
it anyway.)

-- 
Joshua Kwan

[-- Attachment #2: Type: application/pgp-signature, Size: 827 bytes --]

^ permalink raw reply	[flat|nested] 3+ messages in thread

end of thread, other threads:[~2003-09-30  1:27 UTC | newest]

Thread overview: 3+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2003-09-29 10:52 Possible regression with 2.6.0-test6 Stephen Anthony
2003-09-30  1:17 ` Nick Piggin
2003-09-30  1:27 ` Joshua Kwan

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox