* 2.6.9-rc2-mm2 vs glxgears
@ 2004-09-23 4:12 Gene Heskett
0 siblings, 0 replies; 6+ messages in thread
From: Gene Heskett @ 2004-09-23 4:12 UTC (permalink / raw)
To: linux-kernel
Greetings;
One of the things I noted about the last few kernels is that glxgears
seems to get tangled up in a bowl of very stiff molasses, for
instance from 2.6.9-rc1-mm5:
[root@coyote root]# glxgears
1622 frames in 5.0 seconds = 324.400 FPS
932 frames in 5.0 seconds = 186.400 FPS
800 frames in 5.0 seconds = 160.000 FPS
712 frames in 5.0 seconds = 142.400 FPS
637 frames in 5.0 seconds = 127.400 FPS
624 frames in 5.0 seconds = 124.800 FPS
564 frames in 5.0 seconds = 112.800 FPS
554 frames in 5.0 seconds = 110.800 FPS
553 frames in 5.0 seconds = 110.600 FPS
554 frames in 5.0 seconds = 110.800 FPS
554 frames in 5.0 seconds = 110.800 FPS
525 frames in 5.0 seconds = 105.000 FPS
499 frames in 5.0 seconds = 99.800 FPS
498 frames in 5.0 seconds = 99.600 FPS
499 frames in 5.0 seconds = 99.800 FPS
499 frames in 5.0 seconds = 99.800 FPS
498 frames in 5.0 seconds = 99.600 FPS
499 frames in 5.0 seconds = 99.800 FPS
499 frames in 5.0 seconds = 99.800 FPS
498 frames in 5.0 seconds = 99.600 FPS
498 frames in 5.0 seconds = 99.600 FPS
499 frames in 5.0 seconds = 99.800 FPS
499 frames in 5.0 seconds = 99.800 FPS
498 frames in 5.0 seconds = 99.600 FPS
Broken pipe
But, imagine by surprise when I rebooted to rc2-mm2 and found this:
[root@coyote root]# glxgears
49 frames in 5.0 seconds = 9.800 FPS
46 frames in 5.0 seconds = 9.200 FPS
46 frames in 5.0 seconds = 9.200 FPS
31 frames in 5.0 seconds = 6.200 FPS <-I started kmail here
37 frames in 5.0 seconds = 7.400 FPS
49 frames in 5.0 seconds = 9.800 FPS
47 frames in 5.0 seconds = 9.400 FPS
49 frames in 5.0 seconds = 9.800 FPS
50 frames in 5.0 seconds = 10.000 FPS
50 frames in 5.0 seconds = 10.000 FPS
50 frames in 5.0 seconds = 10.000 FPS
50 frames in 5.0 seconds = 10.000 FPS
Broken pipe
Now, just for grins, I going to rebuild 2.6.9-rc1-mm5 without cachefs
to see if that makes any difference.
Not enough to bother calling mother over:
[root@coyote root]# glxgears
1516 frames in 5.0 seconds = 303.200 FPS
983 frames in 5.0 seconds = 196.600 FPS
832 frames in 5.0 seconds = 166.400 FPS
703 frames in 5.0 seconds = 140.600 FPS
663 frames in 5.0 seconds = 132.600 FPS
623 frames in 5.0 seconds = 124.600 FPS
585 frames in 5.0 seconds = 117.000 FPS
553 frames in 5.0 seconds = 110.600 FPS
554 frames in 5.0 seconds = 110.800 FPS
554 frames in 5.0 seconds = 110.800 FPS
554 frames in 5.0 seconds = 110.800 FPS
554 frames in 5.0 seconds = 110.800 FPS
514 frames in 5.0 seconds = 102.800 FPS
499 frames in 5.0 seconds = 99.800 FPS
499 frames in 5.0 seconds = 99.800 FPS
499 frames in 5.0 seconds = 99.800 FPS
499 frames in 5.0 seconds = 99.800 FPS
498 frames in 5.0 seconds = 99.600 FPS
499 frames in 5.0 seconds = 99.800 FPS
496 frames in 5.0 seconds = 99.200 FPS
Broken pipe
Its not using any cpu unless you count once or twice I saw it
report .25%.
Definitely weirdsville incorporated though.
--
Cheers, Gene
"There are four boxes to be used in defense of liberty:
soap, ballot, jury, and ammo. Please use in that order."
-Ed Howdershelt (Author)
99.26% setiathome rank, not too shabby for a WV hillbilly
Yahoo.com attorneys please note, additions to this message
by Gene Heskett are:
Copyright 2004 by Maurice Eugene Heskett, all rights reserved.
^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: 2.6.9-rc2-mm2 vs glxgears
2004-09-23 14:57 ` Gene Heskett
@ 2004-09-23 18:28 ` Dave Jones
0 siblings, 0 replies; 6+ messages in thread
From: Dave Jones @ 2004-09-23 18:28 UTC (permalink / raw)
To: Gene Heskett; +Cc: linux-kernel, Dave Airlie, Frank Phillips
On Thu, Sep 23, 2004 at 10:57:15AM -0400, Gene Heskett wrote:
> X-tacy version of an ATI Radeon 9200SE, 128 megs of ram. And these
> lines from messages at about the time I did the startx I've not seen
> before:
> Sep 23 03:19:55 coyote kernel: agpgart: Found an AGP 3.0 compliant
> device at 0000:00:00.0.
> Sep 23 03:19:55 coyote kernel: agpgart: Putting AGP V3 device at
> 0000:00:00.0 into 4x mode
> Sep 23 03:19:55 coyote kernel: agpgart: Putting AGP V3 device at
> 0000:02:00.0 into 4x mode
> Sep 23 03:19:55 coyote kernel: [drm] Loading R200 Microcode
>
> So this is something new with rc2-mm2 (new to me anyway). The card,
> and its mobo socket are supposedly 8X, so why the setting to 4X?
Because X still doesn't support an AGPMode "x8" option yet.
Dave
^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: 2.6.9-rc2-mm2 vs glxgears
2004-09-23 8:19 ` Dave Airlie
@ 2004-09-23 14:57 ` Gene Heskett
2004-09-23 18:28 ` Dave Jones
0 siblings, 1 reply; 6+ messages in thread
From: Gene Heskett @ 2004-09-23 14:57 UTC (permalink / raw)
To: linux-kernel, Dave Airlie; +Cc: Frank Phillips
On Thursday 23 September 2004 04:19, Dave Airlie wrote:
>> Which even I have to agree is pretty pathetic.
>
>What do you get on a Linus kernel, I'm not tracking -mm as much as I
>should, the missing pci_enable_device might have caused some
> issues...
I hadn't even built 2.6.9-rc2 due to the widespread and constant hang
reports.
>On my 2.6.8.1 at the moment glxgears stays constant enough, I'll
> boot into a 2.6.9 later on and check it out...
>
I'm not sure if I have a 2.6.8.1 in my grub.conf still. This brings
up a minor question: How many entries can one actually have in the
grub.conf before something overflows? I'd set a rather abitrary
limit of 16 here, but I have more room in the /boot partition than
that, and the 2.6.8.1 kernel still exists I believe.
>What graphics card you have?
>Dave.
X-tacy version of an ATI Radeon 9200SE, 128 megs of ram. And these
lines from messages at about the time I did the startx I've not seen
before:
Sep 23 03:19:55 coyote kernel: agpgart: Found an AGP 3.0 compliant
device at 0000:00:00.0.
Sep 23 03:19:55 coyote kernel: agpgart: Putting AGP V3 device at
0000:00:00.0 into 4x mode
Sep 23 03:19:55 coyote kernel: agpgart: Putting AGP V3 device at
0000:02:00.0 into 4x mode
Sep 23 03:19:55 coyote kernel: [drm] Loading R200 Microcode
So this is something new with rc2-mm2 (new to me anyway). The card,
and its mobo socket are supposedly 8X, so why the setting to 4X? Not
that this has very much to do with this problem since 10fps can be
done on AGP 0.05X :)
[...]
--
Cheers, Gene
"There are four boxes to be used in defense of liberty:
soap, ballot, jury, and ammo. Please use in that order."
-Ed Howdershelt (Author)
99.26% setiathome rank, not too shabby for a WV hillbilly
Yahoo.com attorneys please note, additions to this message
by Gene Heskett are:
Copyright 2004 by Maurice Eugene Heskett, all rights reserved.
^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: 2.6.9-rc2-mm2 vs glxgears
2004-09-23 7:27 ` Gene Heskett
@ 2004-09-23 8:19 ` Dave Airlie
2004-09-23 14:57 ` Gene Heskett
0 siblings, 1 reply; 6+ messages in thread
From: Dave Airlie @ 2004-09-23 8:19 UTC (permalink / raw)
To: gene.heskett; +Cc: linux-kernel, Frank Phillips
>
> Which even I have to agree is pretty pathetic.
What do you get on a Linus kernel, I'm not tracking -mm as much as I
should, the missing pci_enable_device might have caused some issues...
On my 2.6.8.1 at the moment glxgears stays constant enough, I'll boot
into a 2.6.9 later on and check it out...
What graphics card you have?
Dave.
>
> >courtesy Jon Smirl. See this thread:
> > http://marc.theaimsgroup.com/?t=109530394200002&r=1&w=2
> >
> >With this I get consistent 350s on 2.6.9-rc2-mm1-VP-S1.
> >
> >Frank
>
> Other than the glxgears being slow, it seems to be working, so I'm
> gonna go sleep in it.
>
>
>
> --
> Cheers, Gene
> "There are four boxes to be used in defense of liberty:
> soap, ballot, jury, and ammo. Please use in that order."
> -Ed Howdershelt (Author)
> 99.26% setiathome rank, not too shabby for a WV hillbilly
> Yahoo.com attorneys please note, additions to this message
> by Gene Heskett are:
> Copyright 2004 by Maurice Eugene Heskett, all rights reserved.
> -
> To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at http://vger.kernel.org/majordomo-info.html
> Please read the FAQ at http://www.tux.org/lkml/
>
^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: 2.6.9-rc2-mm2 vs glxgears
2004-09-23 5:23 Frank Phillips
@ 2004-09-23 7:27 ` Gene Heskett
2004-09-23 8:19 ` Dave Airlie
0 siblings, 1 reply; 6+ messages in thread
From: Gene Heskett @ 2004-09-23 7:27 UTC (permalink / raw)
To: linux-kernel; +Cc: Frank Phillips
On Thursday 23 September 2004 01:23, Frank Phillips wrote:
>Hello,
>
>I don't know why your FPS would be decreasing like that, but as for
>the 9FPS - radeon, right? Look for this line in Xorg.0.log:
>
>(EE) RADEON(0): [pci] Out of memory (-1007)
On 2.6.9-rc1-mm5 ATM, and the above line doesn't exist in the
Xorg.0.log
>this is an easy fix:
>
>===== linux/drm_scatter.h 1.6 vs edited =====
>--- 1.6/linux/drm_scatter.h Sun Sep 5 21:22:06 2004
>+++ edited/linux/drm_scatter.h Thu Sep 16 01:11:13 2004
>@@ -73,7 +73,7 @@
>
> DRM_DEBUG( "%s\n", __FUNCTION__ );
>
>- if (drm_core_check_feature(dev, DRIVER_SG))
>+ if (!drm_core_check_feature(dev, DRIVER_SG))
> return -EINVAL;
>
> if ( dev->sg )
And this '!' is already in 2.6.9-rc2-mm2.
I'm building a 2.6.9-rc1-mm5 with the exclamation mark now & we'll see
what effect that has on glxgears once amanda is done.
This patch seems to be a no-op here. Booted to it now, with no really
significant difference in the slowdown:
1463 frames in 5.0 seconds = 292.600 FPS
985 frames in 5.0 seconds = 197.000 FPS
831 frames in 5.0 seconds = 166.200 FPS
727 frames in 5.0 seconds = 145.400 FPS
636 frames in 5.0 seconds = 127.200 FPS
682 frames in 5.0 seconds = 136.400 FPS
622 frames in 5.0 seconds = 124.400 FPS
622 frames in 5.0 seconds = 124.400 FPS
554 frames in 5.0 seconds = 110.800 FPS
552 frames in 5.0 seconds = 110.400 FPS
552 frames in 5.0 seconds = 110.400 FPS
464 frames in 5.0 seconds = 92.800 FPS
310 frames in 5.0 seconds = 62.000 FPS
424 frames in 5.0 seconds = 84.800 FPS
404 frames in 5.0 seconds = 80.800 FPS
422 frames in 5.0 seconds = 84.400 FPS
461 frames in 5.0 seconds = 92.200 FPS
437 frames in 5.0 seconds = 87.400 FPS
287 frames in 5.0 seconds = 57.400 FPS
So while the patch may be correct, I'm apparently not hitting that
exact piece of code here.
Then, rebooted to 2.6.9-rc2-mm2, I'm back to this, also without any
errors in the Xorg.0.log:
[root@coyote root]# glxgears
60 frames in 5.0 seconds = 12.000 FPS
49 frames in 5.0 seconds = 9.800 FPS
49 frames in 5.0 seconds = 9.800 FPS
50 frames in 5.0 seconds = 10.000 FPS
49 frames in 5.0 seconds = 9.800 FPS
50 frames in 5.0 seconds = 10.000 FPS
49 frames in 5.0 seconds = 9.800 FPS
49 frames in 5.0 seconds = 9.800 FPS
48 frames in 5.0 seconds = 9.600 FPS
48 frames in 5.0 seconds = 9.600 FPS
50 frames in 5.0 seconds = 10.000 FPS
49 frames in 5.0 seconds = 9.800 FPS
43 frames in 5.0 seconds = 8.600 FPS
47 frames in 5.0 seconds = 9.400 FPS
49 frames in 5.0 seconds = 9.800 FPS
44 frames in 5.0 seconds = 8.800 FPS
31 frames in 5.0 seconds = 6.200 FPS
45 frames in 5.0 seconds = 9.000 FPS
Which even I have to agree is pretty pathetic.
>courtesy Jon Smirl. See this thread:
> http://marc.theaimsgroup.com/?t=109530394200002&r=1&w=2
>
>With this I get consistent 350s on 2.6.9-rc2-mm1-VP-S1.
>
>Frank
Other than the glxgears being slow, it seems to be working, so I'm
gonna go sleep in it.
--
Cheers, Gene
"There are four boxes to be used in defense of liberty:
soap, ballot, jury, and ammo. Please use in that order."
-Ed Howdershelt (Author)
99.26% setiathome rank, not too shabby for a WV hillbilly
Yahoo.com attorneys please note, additions to this message
by Gene Heskett are:
Copyright 2004 by Maurice Eugene Heskett, all rights reserved.
^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: 2.6.9-rc2-mm2 vs glxgears
@ 2004-09-23 5:23 Frank Phillips
2004-09-23 7:27 ` Gene Heskett
0 siblings, 1 reply; 6+ messages in thread
From: Frank Phillips @ 2004-09-23 5:23 UTC (permalink / raw)
To: gene.heskett; +Cc: linux-kernel
Hello,
I don't know why your FPS would be decreasing like that, but as for
the 9FPS - radeon, right? Look for this line in Xorg.0.log:
(EE) RADEON(0): [pci] Out of memory (-1007)
this is an easy fix:
===== linux/drm_scatter.h 1.6 vs edited =====
--- 1.6/linux/drm_scatter.h Sun Sep 5 21:22:06 2004
+++ edited/linux/drm_scatter.h Thu Sep 16 01:11:13 2004
@@ -73,7 +73,7 @@
DRM_DEBUG( "%s\n", __FUNCTION__ );
- if (drm_core_check_feature(dev, DRIVER_SG))
+ if (!drm_core_check_feature(dev, DRIVER_SG))
return -EINVAL;
if ( dev->sg )
courtesy Jon Smirl. See this thread: http://marc.theaimsgroup.com/?t=109530394200002&r=1&w=2
With this I get consistent 350s on 2.6.9-rc2-mm1-VP-S1.
Frank
--
______________________________________________
Check out the latest SMS services @ http://www.linuxmail.org
This allows you to send and receive SMS through your mailbox.
Powered by Outblaze
^ permalink raw reply [flat|nested] 6+ messages in thread
end of thread, other threads:[~2004-09-23 18:28 UTC | newest]
Thread overview: 6+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2004-09-23 4:12 2.6.9-rc2-mm2 vs glxgears Gene Heskett
2004-09-23 5:23 Frank Phillips
2004-09-23 7:27 ` Gene Heskett
2004-09-23 8:19 ` Dave Airlie
2004-09-23 14:57 ` Gene Heskett
2004-09-23 18:28 ` Dave Jones
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.