Terrible interactivity with 2.6.0-t9-mm3
I just noticed major interactivity problems whilst ogging one of my cds. X (which is running at a nice of 0) stuttered in its display uptdates (eg: ogging on two seperate machines but displaying on one, the progress display of the much faster box would come in spurts until I ^z the ogg on the box running X). Keyboard input also stutters and the nfs connection (over which I was encoding from the X running box) eventually gave up the ghost and I got lots of 'server not responding, timed out' msgs.
Playing Heroes3 was also impossible until I ran it under nice -n 1.
Doh. :/ This is the first time this has been so bad that I've felt it was worth writing about. :/
Systems:
X: P3-700 with 256MB RAM, half of it free running Debian sarge (libc
2.3.2 and X 4.2.1.1) and 2.6.0-t9-mm3.
SSH: Athlon XP 2500+ with 512MB of RAM with all but 46MB free running
the same version of Debian but with 2.6.0-t9-mm2.
Signature · 3 lines
From the people who brought you burnt villages in Vietnam...
http://news.independent.co.uk/world/middle_east/story.jsp?story=452375Quoted 1 line · CaT
I just noticed major interactivity problems whilst ogging one of my
[...]
Quoted 2 lines
Doh. :/ This is the first time this has been so bad that I've felt it was worth writing about. :/
Yup, I was using a patched mm3 so I wanted to try the plain one, but due to your post, I can conclude it really is mm3 which is really bad. I noticed this as well. Using Nick's CPU scheduler things are not thaaat bad, but still far from mm2. I think there is some major problem introduced to mm3. Without Nicks patch doing an emerge/compiling, even my mouse heavily stutters like hell, regradless of the used io scheduler. With Nick's patch the mouse is rather OK, but the rest if the system is still not really usable. It is not a HD problem...
Going back to mm2 (patched mm2) and everything it fine again.
Athlon XP 1900MHz
1GB DDR RAM
NFORCE2 Chipset
Prakash
Quoted 5 lines · Prakash K. Cheemplavam
Quoted 1 line · CaT
I just noticed major interactivity problems whilst ogging one of my
"ogging"?
Quoted 1 line
Going back to mm2 (patched mm2) and everything it fine again.
Two things to try, please:
a) Is the problem from Linus's tree? Try 2.6.0-test9 plus
ftp://ftp.kernel.org/pub/linux/kernel/people/akpm/patches/2.6/2.6.0-test9/2.6.0-test9-mm3/broken-out/linus.patch
b) The only significant scheduler change in mm3 was
ftp://ftp.kernel.org/pub/linux/kernel/people/akpm/patches/2.6/2.6.0-test9/2.6.0-test9-mm3/broken-out/context-switch-accounting-fix.patch
So please try -mm3 with the above patch reverted with
patch -R -p1 < context-switch-accounting-fix.patch
Thanks.
Em Sun, Nov 16, 2003 at 01:42:31PM -0800, Andrew Morton escreveu:
Quoted 8 lines
Quoted 5 lines · Prakash K. Cheemplavam
Quoted 1 line · CaT
I just noticed major interactivity problems whilst ogging one of my
"ogging"?
To Ogg: to oggify, to convert some digital sound file to the .ogg format
8)
- Arnaldo
Quoted 4 lines · Arnaldo Carvalho de Melo
Em Sun, Nov 16, 2003 at 01:42:31PM -0800, Andrew Morton escreveu:
Quoted 2 lines
Quoted 1 line · Prakash K. Cheemplavam
Quoted 9 lines
Quoted 6 lines
Quoted 4 lines
Quoted 1 line · CaT
I just noticed major interactivity problems whilst ogging one of my
"ogging"?
To Ogg: to oggify, to convert some digital sound file to the .ogg format
This much I know. But it's not exactly a tightly defined and repeatable workload. Oh well.
Quoted 5 lines · Andrew Morton
Quoted 4 lines · Arnaldo Carvalho de Melo
Em Sun, Nov 16, 2003 at 01:42:31PM -0800, Andrew Morton escreveu:
Quoted 3 lines
Quoted 2 lines · Prakash K. Cheemplavam
CaT wrote:
Quoted 8 lines
Quoted 5 lines
Quoted 3 lines
Quoted 1 line
Quoted 1 line
I just noticed major interactivity problems whilst ogging one of my
"ogging"?
To Ogg: to oggify, to convert some digital sound file to the .ogg format
This much I know. But it's not exactly a tightly defined and repeatable workload. Oh well.
I can't do much testing atm but the similar symptoms happen whilst running Heroes3 which tends to suck CPU so I think mp3 encoding would work just aswell or anything else that sucks all the CPU it can.
Prakash also reported similar problems whilst compiling the kernel. This is something I can duplicate also. As my kernel compiles the system can't even keep up with my typing.
Signature · 3 lines
From the people who brought you burnt villages in Vietnam...
http://news.independent.co.uk/world/middle_east/story.jsp?story=452375I also had problems with -mm3.
Now I am using the latest tree from linus with no problems.
P.S.
Sorry but I have no time to try mm3 with the patch reverted.
Panagiotis Papadakos
Quoted 31 lines · Andrew Morton
Quoted 5 lines · Prakash K. Cheemplavam
Quoted 1 line · CaT
I just noticed major interactivity problems whilst ogging one of my
"ogging"?
Quoted 1 line
Going back to mm2 (patched mm2) and everything it fine again.
Two things to try, please:
a) Is the problem from Linus's tree? Try 2.6.0-test9 plus
ftp://ftp.kernel.org/pub/linux/kernel/people/akpm/patches/2.6/2.6.0-test9/2.6.0-test9-mm3/broken-out/linus.patchb) The only significant scheduler change in mm3 was
ftp://ftp.kernel.org/pub/linux/kernel/people/akpm/patches/2.6/2.6.0-test9/2.6.0-test9-mm3/broken-out/context-switch-accounting-fix.patchSo please try -mm3 with the above patch reverted with
patch -R -p1 < context-switch-accounting-fix.patch
Thanks.
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to [email protected]
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/
Quoted 12 lines · Andrew Morton
Two things to try, please:
a) Is the problem from Linus's tree? Try 2.6.0-test9 plus
ftp://ftp.kernel.org/pub/linux/kernel/people/akpm/patches/2.6/2.6.0-test9/2.6.0-test9-mm3/broken-out/linus.patchb) The only significant scheduler change in mm3 was
ftp://ftp.kernel.org/pub/linux/kernel/people/akpm/patches/2.6/2.6.0-test9/2.6.0-test9-mm3/broken-out/context-switch-accounting-fix.patchSo please try -mm3 with the above patch reverted with
patch -R -p1 < context-switch-accounting-fix.patch
Hi Andrew,
This is also easily reproducible here with just a kernel compile.
I have tried both a) and b) with b) not changing anything, but a) seems to work... Anything more to try?
Gawain
Quoted 15 lines · Gawain Lynch
Quoted 12 lines · Andrew Morton
Two things to try, please:
a) Is the problem from Linus's tree? Try 2.6.0-test9 plus
ftp://ftp.kernel.org/pub/linux/kernel/people/akpm/patches/2.6/2.6.0-test9/2.6.0-test9-mm3/broken-out/linus.patchb) The only significant scheduler change in mm3 was
ftp://ftp.kernel.org/pub/linux/kernel/people/akpm/patches/2.6/2.6.0-test9/2.6.0-test9-mm3/broken-out/context-switch-accounting-fix.patchSo please try -mm3 with the above patch reverted with
patch -R -p1 < context-switch-accounting-fix.patch
Quoted 7 lines
Hi Andrew,
This is also easily reproducible here with just a kernel compile.
I have tried both a) and b) with b) not changing anything, but a) seems to work... Anything more to try?
Your report has totally confused me. Are you saying that the jerkiness is caused by linus.patch? Or not? Pleas try again ;)
Quoted 18 lines · Andrew Morton
Quoted 17 lines · Gawain Lynch
Quoted 12 lines · Andrew Morton
Two things to try, please:
a) Is the problem from Linus's tree? Try 2.6.0-test9 plus
ftp://ftp.kernel.org/pub/linux/kernel/people/akpm/patches/2.6/2.6.0-test9/2.6.0-test9-mm3/broken-out/linus.patchb) The only significant scheduler change in mm3 was
ftp://ftp.kernel.org/pub/linux/kernel/people/akpm/patches/2.6/2.6.0-test9/2.6.0-test9-mm3/broken-out/context-switch-accounting-fix.patchSo please try -mm3 with the above patch reverted with
patch -R -p1 < context-switch-accounting-fix.patch
Hi Andrew,
Quoted 8 lines
Quoted 5 lines
This is also easily reproducible here with just a kernel compile.
I have tried both a) and b) with b) not changing anything, but a) seems to work... Anything more to try?
Your report has totally confused me. Are you saying that the jerkiness is caused by linus.patch? Or not? Pleas try again ;)
Sorry that would be the "several too many" glasses of wine last night... ;)
I do NOT see the interactivity problems with linus.patch.
I DO see the problem with -mm3
I DO see the problem with context-switch-accounting-fix.patch backed out...
Quoted 19 lines · Andrew Morton
Quoted 18 lines · Gawain Lynch
Quoted 6 lines · Andrew Morton
Two things to try, please:
a) Is the problem from Linus's tree? Try 2.6.0-test9 plus
ftp://ftp.kernel.org/pub/linux/kernel/people/akpm/patches/2.6/2
.6.0-test9/2.6.0-test9-mm3/broken-out/linus.patchQuoted 8 lines
b) The only significant scheduler change in mm3 was
ftp://ftp.kernel.org/pub/linux/kernel/people/akpm/patches/2.6/2.6.0-test9/2.6.0-test9-mm3/broken-out/context-switch-accounting-f ix.patch
So please try -mm3 with the above patch reverted with
patch -R -p1 < context-switch-accounting-fix.patch
Hi Andrew,
Quoted 7 lines
Quoted 4 lines
This is also easily reproducible here with just a kernel compile.
I have tried both a) and b) with b) not changing anything, but a) seems to work... Anything more to try?
Your report has totally confused me. Are you saying that the jerkiness is caused by linus.patch? Or not? Pleas try again ;)
In defense of this code, I ran -mm3 with the deadline elevator for about 3 days and was very happy with the interactivity. Now I've been running with the elevator=cfq for most of the day, and it also seems to be pretty responsive. The default as wasn't, at least for -mm2, and I haven't tried it yet for -mm3. Should I, and report back in a day or so so that you've got reports from an otherwise identical system to compare?
We need a test suite for this :)
Signature · 7 lines
Cheers, Gene AMD K6-III@500mhz 320M Athlon1600XP@1400mhz 512M 99.27% setiathome rank, not too shabby for a WV hillbilly Yahoo.com attornies please note, additions to this message by Gene Heskett are: Copyright 2003 by Maurice Eugene Heskett, all rights reserved.
Quoted 28 lines · Gene Heskett
Quoted 26 lines · Andrew Morton
Quoted 24 lines · Gawain Lynch
Quoted 6 lines · Andrew Morton
Two things to try, please:
a) Is the problem from Linus's tree? Try 2.6.0-test9 plus
ftp://ftp.kernel.org/pub/linux/kernel/people/akpm/patches/2.6/2
.6.0-test9/2.6.0-test9-mm3/broken-out/linus.patchQuoted 9 lines
b) The only significant scheduler change in mm3 was
ftp://ftp.kernel.org/pub/linux/kernel/people/akpm/patches/2.6/2.6.0-test9/2.6.0-test9-mm3/broken-out/context-switch-accounting-f ix.patch
So please try -mm3 with the above patch reverted with
patch -R -p1 < context-switch-accounting-fix.patch
Hi Andrew,
This is also easily reproducible here with just a kernel compile.
I have tried both a) and b) with b) not changing anything, but a) seems to work... Anything more to try?
Quoted 8 lines
Quoted 3 lines
Your report has totally confused me. Are you saying that the jerkiness is caused by linus.patch? Or not? Pleas try again ;)
In defense of this code, I ran -mm3 with the deadline elevator for about 3 days and was very happy with the interactivity. Now I've been running with the elevator=cfq for most of the day, and it also seems to be pretty responsive. The default as wasn't, at least for -mm2, and I haven't tried it yet for -mm3. Should I, and report back
in a day or so so that you've got reports from an otherwise identical system to compare?
I think you might have confused Andrew a bit more ;)
To start with, you are talking about IO schedulers, while the thread is about CPU interactivity.
The problem here looks like something that is caused by something in mm3, not in mm2, not linus.patch, and not context-switch-accounting-fix.patch.
Off topic: it would be good if you could try the as disk scheduler in mm3. I recall you had some problems with it earlier, but they should be fixed in mm3. Thanks.
Quoted 3 lines
We need a test suite for this :)
Subjective reports from our base of beta testers has proven to be the best thing.
Quoted 5 lines · Nick Piggin
I think you might have confused Andrew a bit more ;)
To start with, you are talking about IO schedulers, while the thread is about CPU interactivity.
I wasn't aware that such performance issues were divorced. In my admitted limited view, a laggy mouse is a laggy mouse, and its due to irq latency in achieving the context switch to service the mouses data. To me, its sorta like 2=2. :)
Quoted 8 lines
The problem here looks like something that is caused by something in mm3, not in mm2, not linus.patch, and not context-switch-accounting-fix.patch.
Off topic: it would be good if you could try the as disk scheduler in mm3. I recall you had some problems with it earlier, but they should be fixed in mm3. Thanks.
Ok Nick. I'll reboot tomorrow without the elevator argument. Right now, amanda is fixin to be fired off in about 25 minutes and I want to see how badly its estimate phase hogs the machine using the cfq scheduler. With the -mm2 as, it was almost psychedelic to watch the mouse move.
Also off topic re mouse performance, and I expect this is an X issue, but when its been blanked because I'm typing, it takes about a full seconds worth of hand waving before it becomes visible again. This is an X issue and I should go away, right?
Quoted 5 lines
Quoted 2 lines
We need a test suite for this :)
Subjective reports from our base of beta testers has proven to be the best thing.
So beit. I guess I must be one of those. (-:
Signature · 7 lines
Cheers, Gene AMD K6-III@500mhz 320M Athlon1600XP@1400mhz 512M 99.27% setiathome rank, not too shabby for a WV hillbilly Yahoo.com attornies please note, additions to this message by Gene Heskett are: Copyright 2003 by Maurice Eugene Heskett, all rights reserved.
Quoted 10 lines · Gene Heskett
Quoted 5 lines · Nick Piggin
I think you might have confused Andrew a bit more ;)
To start with, you are talking about IO schedulers, while the thread is about CPU interactivity.
I wasn't aware that such performance issues were divorced. In my admitted limited view, a laggy mouse is a laggy mouse, and its due to irq latency in achieving the context switch to service the mouses data. To me, its sorta like 2=2. :)
I guess its mostly due to scheduling latency. I think IRQ latency is generally very good these days.
Scheduling latency caused not by long critical sections in the kernel, but having things scheduled in front of you (X) for a long time.
Quoted 6 lines
Quoted 6 lines
The problem here looks like something that is caused by something in mm3, not in mm2, not linus.patch, and not context-switch-accounting-fix.patch.
Quoted 7 lines
Quoted 4 lines
Off topic: it would be good if you could try the as disk scheduler in mm3. I recall you had some problems with it earlier, but they should be fixed in mm3. Thanks.
Ok Nick. I'll reboot tomorrow without the elevator argument. Right now, amanda is fixin to be fired off in about 25 minutes and I want to see how badly its estimate phase hogs the machine using the cfq scheduler. With the -mm2 as, it was almost psychedelic to watch the mouse move.
OK thanks. I think this problem you were seeing _was_ interrupt latency due to AS doing millions of WARNs. It should be fixed in mm3.
Quoted 3 lines
Also off topic re mouse performance, and I expect this is an X issue, but when its been blanked because I'm typing, it takes about a full seconds worth of hand waving before it becomes visible again. This is an X issue and I should go away, right?
Sounds like it.