[lkml]   [2003]   [Jul]   [31]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
Messages in this thread
SubjectRe: 2.6.0-test2-mm2
* Andrew Morton <> [31-07-2003 14:52]:
> . CPU scheduler changes
> . Several changes to the synaptics and PS/2 drivers. People who have had
> problems with keyboards and mice, please test and report.
> . Lots of other things, mainly bugfixes.

I have noticed that while the system remains responsive when there is
heavy CPU load alone, the unresponsiveness comes when there is heavy disk
i/o. In light of this, I subjected my system to the following tests :

1) Untarred linux-2.6.0-test2.tar.bz2 (To create a stream of writes)
2) did a `find / -name "foobar" -print` as root (To create a stream of reads)
3) bzip2'ed a 30 MB file to 3.2 MB ( this ensures 100 % CPU usage).

All the above were working on the same disk.

As the above three things were going on, I was browsing, playing music
on xmms, reading a pdf file in acrobat and generally switching between

I did the above test on both 2.6.0-test2-mm1 + O11int + O11.1int ( no
O11.2 int) and on 2.6.0-test2-mm2. I find that mm1 + O11 patches are
better than mm2. The music never skipped on mm1, while it did on
mm2. Also, overall resposiveness of the various windows was better on

However, when I did a `rpm --rebuilddb` on mm1 + O11int patches, I
still got quite severe skipping toward the end of the 8 min process. I
could not repeat the skipping again, even on the same kernel, because
I guess there was not much rebuilding to do again..

If there are tools which I can use to produce helpful numbers, please
let me know. I will post the required numbers ASAP.

Also, if you have any other tests which I could perform to create
heavy disk I/O, please let me know about that too.


- Apurva
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to
More majordomo info at
Please read the FAQ at

 \ /
  Last update: 2005-03-22 13:47    [W:0.125 / U:1.684 seconds]
©2003-2020 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site