Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]

Groups > linux.debian.user > #161370 > unrolled thread

X using ~ 35% of a CPU core

Started byCelejar <celejar@gmail.com>
First post2015-11-30 19:50 +0100
Last post2015-12-03 06:10 +0100
Articles 5 — 2 participants

Back to article view | Back to linux.debian.user

Contents

  X using ~ 35% of a CPU core Celejar <celejar@gmail.com> - 2015-11-30 19:50 +0100
    Re: X using ~ 35% of a CPU core Sven Arvidsson <sa@whiz.se> - 2015-11-30 21:50 +0100
      Re: X using ~ 35% of a CPU core Celejar <celejar@gmail.com> - 2015-12-02 00:20 +0100
        Re: X using ~ 35% of a CPU core Sven Arvidsson <sa@whiz.se> - 2015-12-02 16:00 +0100
          Re: X using ~ 35% of a CPU core Celejar <celejar@gmail.com> - 2015-12-03 06:10 +0100

#161370 — X using ~ 35% of a CPU core

FromCelejar <celejar@gmail.com>
Date2015-11-30 19:50 +0100
SubjectX using ~ 35% of a CPU core
Message-ID<qABGq-7gu-19@gated-at.bofh.it>
Hi,

On my ThinkPad T61 (Core 2 Duo T7300 @ 2.00GHZ) running Debian stable
(8.2) with Xfce, the X process (/usr/bin/X) has recently begun to
consume some 35% (seems to vary between 33-40 %) for several minutes at
a time, with the system becoming unpleasantly unresponsive during this
time. It seems to mostly happen after resuming from suspend-to-ram, and
usually clears up after a few minutes with the system performance
returning to normal. I run mostly straightforward stable, albeit with
self-compiled kernels from upstream (vanilla). I first noticed the
problem on 4.1.13 and am now seeing it under 4.2.6. I don't think I've
seen this behavior in the past, and I'm not sure what has triggered it.
I see it even when not running much of a load under X at all, e.g.,
just recently, when all I had running was Xfce Terminal, Sylpheed, and
one (simple, small) OpenOffice writer document. Any idea what this is,
or how I can fix / troubleshoot?

Thanks,
Celejar

[toc] | [next] | [standalone]

#161373

FromSven Arvidsson <sa@whiz.se>
Date2015-11-30 21:50 +0100
Message-ID<qADyy-8ry-21@gated-at.bofh.it>
In reply to#161370

[Multipart message — attachments visible in raw view] — view raw

On Mon, 2015-11-30 at 13:44 -0500, Celejar wrote:
> Hi,
> 
> On my ThinkPad T61 (Core 2 Duo T7300 @ 2.00GHZ) running Debian stable
> (8.2) with Xfce, the X process (/usr/bin/X) has recently begun to
> consume some 35% (seems to vary between 33-40 %) for several minutes
> at
> a time, with the system becoming unpleasantly unresponsive during
> this
> time. It seems to mostly happen after resuming from suspend-to-ram,
> and
> usually clears up after a few minutes with the system performance
> returning to normal. I run mostly straightforward stable, albeit with
> self-compiled kernels from upstream (vanilla). I first noticed the
> problem on 4.1.13 and am now seeing it under 4.2.6. I don't think
> I've
> seen this behavior in the past, and I'm not sure what has triggered
> it.
> I see it even when not running much of a load under X at all, e.g.,
> just recently, when all I had running was Xfce Terminal, Sylpheed,
> and
> one (simple, small) OpenOffice writer document. Any idea what this
> is,
> or how I can fix / troubleshoot?
> 
> Thanks,
> Celejar

Long shot, but maybe it's something similar to this? Probing for new
screens after resume and spamming a log before giving up?

https://bbs.archlinux.org/viewtopic.php?id=204098

Would be interesting to know where the X process spends all that time,
but I guess you would need to profile it to find out.

-- 
Cheers,
Sven Arvidsson
http://www.whiz.se
PGP Key ID 6FAB5CD5

[toc] | [prev] | [next] | [standalone]

#161442

FromCelejar <celejar@gmail.com>
Date2015-12-02 00:20 +0100
Message-ID<qB2nh-7GU-53@gated-at.bofh.it>
In reply to#161373
On Mon, 30 Nov 2015 21:47:43 +0100
Sven Arvidsson <sa@whiz.se> wrote:

> On Mon, 2015-11-30 at 13:44 -0500, Celejar wrote:
> > Hi,
> > 
> > On my ThinkPad T61 (Core 2 Duo T7300 @ 2.00GHZ) running Debian stable
> > (8.2) with Xfce, the X process (/usr/bin/X) has recently begun to
> > consume some 35% (seems to vary between 33-40 %) for several minutes
> > at
> > a time, with the system becoming unpleasantly unresponsive during
> > this

...

> Long shot, but maybe it's something similar to this? Probing for new
> screens after resume and spamming a log before giving up?
> 
> https://bbs.archlinux.org/viewtopic.php?id=204098
> 
> Would be interesting to know where the X process spends all that time,
> but I guess you would need to profile it to find out.

Thanks for the tip! My problem is actually at least somewhat similar. I
tried tailing the X log when I caught the problem occurring, and I
found zillions of these:

[ 82198.046] (II) intel(0): EDID vendor "LEN", prod id 16435
[ 82198.046] (II) intel(0): Printing DDC gathered Modelines:
[ 82198.046] (II) intel(0): Modeline "1440x900"x0.0   96.50  1440 1522 1576 1706  900 905 914 943 -hsync -vsync (56.6 kHz eP)
[ 82198.046] (II) intel(0): Modeline "1440x900"x0.0   80.44  1440 1522 1576 1706  900 905 914 943 -hsync -vsync (47.2 kHz e)
[ 82198.084] (II) intel(0): EDID vendor "LEN", prod id 16435
[ 82198.084] (II) intel(0): Printing DDC gathered Modelines:
[ 82198.084] (II) intel(0): Modeline "1440x900"x0.0   96.50  1440 1522 1576 1706  900 905 914 943 -hsync -vsync (56.6 kHz eP)
[ 82198.084] (II) intel(0): Modeline "1440x900"x0.0   80.44  1440 1522 1576 1706  900 905 914 943 -hsync -vsync (47.2 kHz e)
[ 82198.166] (II) intel(0): EDID vendor "LEN", prod id 16435
[ 82198.166] (II) intel(0): Printing DDC gathered Modelines:
[ 82198.166] (II) intel(0): Modeline "1440x900"x0.0   96.50  1440 1522 1576 1706  900 905 914 943 -hsync -vsync (56.6 kHz eP)
[ 82198.166] (II) intel(0): Modeline "1440x900"x0.0   80.44  1440 1522 1576 1706  900 905 914 943 -hsync -vsync (47.2 kHz e)

...

[ 82198.574] (II) intel(0): EDID vendor "LEN", prod id 16435
[ 82198.574] (II) intel(0): Printing DDC gathered Modelines:
[ 82198.574] (II) intel(0): Modeline "1440x900"x0.0   96.50  1440 1522 1576 1706  900 905 914 943 -hsync -vsync (56.6 kHz eP)
[ 82198.574] (II) intel(0): Modeline "1440x900"x0.0   80.44  1440 1522 1576 1706  900 905 914 943 -hsync -vsync (47.2 kHz e)
[ 82198.662] (II) intel(0): EDID vendor "LEN", prod id 16435
[ 82198.662] (II) intel(0): Printing DDC gathered Modelines:
[ 82198.662] (II) intel(0): Modeline "1440x900"x0.0   96.50  1440 1522 1576 1706  900 905 914 943 -hsync -vsync (56.6 kHz eP)
[ 82198.662] (II) intel(0): Modeline "1440x900"x0.0   80.44  1440 1522 1576 1706  900 905 914 943 -hsync -vsync (47.2 kHz e)
[ 82198.710] (II) intel(0): EDID vendor "LEN", prod id 16435
[ 82198.710] (II) intel(0): Printing DDC gathered Modelines:
[ 82198.710] (II) intel(0): Modeline "1440x900"x0.0   96.50  1440 1522 1576 1706  900 905 914 943 -hsync -vsync (56.6 kHz eP)
[ 82198.710] (II) intel(0): Modeline "1440x900"x0.0   80.44  1440 1522 1576 1706  900 905 914 943 -hsync -vsync (47.2 kHz e)


When the problem ceases (CPU usage back down to zero / normal), these
lines cease, too. Anyone know what this is, and / or how to fix it?

I found this:

https://forums.gentoo.org/viewtopic-t-716774-start-0.html

But I don't have anything like his xrandr script running, and my
problem is intermittent - it starts and stops at apparently random
times.

> Sven Arvidsson

Celejar

[toc] | [prev] | [next] | [standalone]

#161456

FromSven Arvidsson <sa@whiz.se>
Date2015-12-02 16:00 +0100
Message-ID<qBh2X-8un-37@gated-at.bofh.it>
In reply to#161442

[Multipart message — attachments visible in raw view] — view raw

On Tue, 2015-12-01 at 18:14 -0500, Celejar wrote:
> tried tailing the X log when I caught the problem occurring, and I
> found zillions of these:
> 
[snipped log]
> 
> When the problem ceases (CPU usage back down to zero / normal), these
> lines cease, too. Anyone know what this is, and / or how to fix it?

I think a bug report would be in order. 

As for a work-around, maybe it's possible to turn off logging for X, or
at least make it less verbose?

It might be the case that the bug is actually fixed in a later version
of the intel driver, but that might necessitate an upgrade to
testing... 


-- 
Cheers,
Sven Arvidsson
http://www.whiz.se
PGP Key ID 6FAB5CD5

[toc] | [prev] | [next] | [standalone]

#161496

FromCelejar <celejar@gmail.com>
Date2015-12-03 06:10 +0100
Message-ID<qBujv-y6-5@gated-at.bofh.it>
In reply to#161456
On Wed, 02 Dec 2015 15:49:40 +0100
Sven Arvidsson <sa@whiz.se> wrote:

> On Tue, 2015-12-01 at 18:14 -0500, Celejar wrote:
> > tried tailing the X log when I caught the problem occurring, and I
> > found zillions of these:
> > 
> [snipped log]
> > 
> > When the problem ceases (CPU usage back down to zero / normal), these
> > lines cease, too. Anyone know what this is, and / or how to fix it?
> 
> I think a bug report would be in order. 

http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=806929

> As for a work-around, maybe it's possible to turn off logging for X, or
> at least make it less verbose?
> 
> It might be the case that the bug is actually fixed in a later version
> of the intel driver, but that might necessitate an upgrade to
> testing... 

Thanks for the help and suggestions. I'll see if I get any response
from the maintainers.

Celejar

[toc] | [prev] | [standalone]

Back to top | Article view | linux.debian.user

csiph-web