Bug 553059 - KMS:RV770PRO:HD4850 1G VRAM performance regression from F11 to Rawhide(F13)
Summary: KMS:RV770PRO:HD4850 1G VRAM performance regression from F11 to Rawhide(F13)
Keywords:
Status: CLOSED WONTFIX
Alias: None
Product: Fedora
Classification: Fedora
Component: xorg-x11-drv-ati
Version: 14
Hardware: All
OS: Linux
medium
medium
Target Milestone: ---
Assignee: Jérôme Glisse
QA Contact: Fedora Extras Quality Assurance
URL:
Whiteboard: card_R700/M
Depends On:
Blocks:
TreeView+ depends on / blocked
 
Reported: 2010-01-06 23:45 UTC by Dennis Jacobfeuerborn
Modified: 2012-08-16 21:39 UTC (History)
3 users (show)

Fixed In Version:
Doc Type: Bug Fix
Doc Text:
Clone Of:
Environment:
Last Closed: 2012-08-16 21:39:22 UTC
Type: ---
Embargoed:


Attachments (Terms of Use)
dmesg log after clean runlevel 5 boot with "drm.debug=15 nouveau.modeset=0" (51.32 KB, text/plain)
2010-01-08 02:46 UTC, Dennis Jacobfeuerborn
no flags Details
Xorg.0.log after clean runlevel 5 boot with "drm.debug=15 nouveau.modeset=0" (74.11 KB, text/plain)
2010-01-08 02:47 UTC, Dennis Jacobfeuerborn
no flags Details
F13 alpha dmesg log after clean runlevel 5 boot with "drm.debug=15 nouveau.modeset=0" (85.71 KB, text/plain)
2010-03-11 00:54 UTC, Dennis Jacobfeuerborn
no flags Details
F13 alpha Xorg.0.log after clean runlevel 5 boot with "drm.debug=15 nouveau.modeset=0" (100.67 KB, text/plain)
2010-03-11 00:55 UTC, Dennis Jacobfeuerborn
no flags Details
performance comparison (1.96 KB, image/png)
2010-03-11 01:02 UTC, Dennis Jacobfeuerborn
no flags Details
Fedora 11 sysprof profile (70.27 KB, text/plain)
2010-08-12 21:29 UTC, Dennis Jacobfeuerborn
no flags Details
Fedora 14 alpha sysprof profile (1.12 MB, text/plain)
2010-08-12 21:30 UTC, Dennis Jacobfeuerborn
no flags Details

Description Dennis Jacobfeuerborn 2010-01-06 23:45:57 UTC
The performance of this card is considerably worse in Rawhide(F13) compared to F11. Playing a youtube video takes almost 100% (one core) of my dual core system but using F11 it only uses about 15%. Also dragging windows, scrolling and opening menus lags and the slow repainting is visible.

Smolt profile of the system:
http://www.smolts.org/client/show/pub_d3abe799-fc48-4b2e-9366-9cf035a6e7b4

(This bug was split of from bug 530784)

Comment 1 François Cami 2010-01-07 00:04:20 UTC
Dennis,

Thank you for opening the bug.

Could you please attach dmesg after successfully booting into runlevel 5 with drm.debug=15 as kernel parameter (and nouveau.modeset=0 of course)?
Oh, and Xorg.0.log as well.

Thanks in advance,

-- 
Fedora Bugzappers volunteer triage team
https://fedoraproject.org/wiki/BugZappers

Comment 2 Dennis Jacobfeuerborn 2010-01-08 02:46:27 UTC
Created attachment 382376 [details]
dmesg log after clean runlevel 5 boot with "drm.debug=15 nouveau.modeset=0"

Comment 3 Dennis Jacobfeuerborn 2010-01-08 02:47:07 UTC
Created attachment 382377 [details]
Xorg.0.log after clean runlevel 5 boot with "drm.debug=15 nouveau.modeset=0"

Comment 4 François Cami 2010-01-08 08:37:45 UTC
Thanks for the logs.

Setting Triaged keyword and keeping status as NEW as this is a rawhide bug, as per rawhide/F13 policy.

-- 
Fedora Bugzappers volunteer triage team
https://fedoraproject.org/wiki/BugZappers

Comment 5 Dennis Jacobfeuerborn 2010-03-10 16:35:03 UTC
Is there any progress on this? Since F11 will loose support shortly after the F13 release I'm wondering what to do at that point. Is there a way to get the F11 radeon driver compiled for F13 so I can upgrade and still have a usable system?

Comment 6 François Cami 2010-03-10 19:44:04 UTC
Dennis,

Judging from the xorg-x11-drv-ati changelog, there were many changes since early January. Could you please check F13 Alpha, at least the Live Media, available at the following url:

http://fedoraproject.org/get-prerelease?anF13a

and report? If it too slow still for you, please post the same kind of logs asked in comment #1 above again.

Thanks in advance,

-- 
Fedora Bugzappers volunteer triage team
https://fedoraproject.org/wiki/BugZappers

Comment 7 Dennis Jacobfeuerborn 2010-03-11 00:54:35 UTC
Created attachment 399213 [details]
F13 alpha dmesg log after clean runlevel 5 boot with "drm.debug=15 nouveau.modeset=0"

Comment 8 Dennis Jacobfeuerborn 2010-03-11 00:55:52 UTC
Created attachment 399214 [details]
F13 alpha Xorg.0.log after clean runlevel 5 boot with "drm.debug=15 nouveau.modeset=0"

Comment 9 Dennis Jacobfeuerborn 2010-03-11 01:02:50 UTC
Created attachment 399215 [details]
performance comparison

This is how the cpu usage looks in the system monitor applet when playing back the same 480p video in youtube on F11 and F13 alpha. Since this is a dual core cpu the F13 alpha playback uses more than one core which leads to occasional hickups. On the F11 system videos play fine even with other activity going on.
Scrolling in firefox is slow but can be improved by disabling smooth scrolling.

Comment 10 Bug Zapper 2010-03-15 13:46:31 UTC
This bug appears to have been reported against 'rawhide' during the Fedora 13 development cycle.
Changing version to '13'.

More information and reason for this action is here:
http://fedoraproject.org/wiki/BugZappers/HouseKeeping

Comment 11 Jérôme Glisse 2010-05-04 09:13:35 UTC
Profiling information would be helpfull here. Something like sysprof, you need start sysprof let a video play and then stop and save profile. Do the same under F11 & F13.

Comment 12 Dennis Jacobfeuerborn 2010-08-12 21:29:03 UTC
Created attachment 438534 [details]
Fedora 11 sysprof profile

Comment 13 Dennis Jacobfeuerborn 2010-08-12 21:30:06 UTC
Created attachment 438535 [details]
Fedora 14 alpha sysprof profile

Comment 14 Dennis Jacobfeuerborn 2010-08-12 21:41:02 UTC
What is immediately apparent in the two profiles above is that the F14alpha version contains a *lot* more samples than the f11 one. The raw_local_irq_restore symbol shows up prominently in the F14alpha profile.

I also want to point out that I now own a Dell notebook with a radeon chip:
http://www.smolts.org/client/show/pub_b7a4b8e3-0b3b-486d-89de-59d0e6873daf

On that system I can easily play 1080p video and don't see the same performance issues I see on my desktop system.

Comment 15 Hans de Goede 2010-09-09 20:04:52 UTC
Hi,

I just hit the exact same problem today, including the raw_local_irq_restore cpu usage standing out. It was pointed out to me this is likely caused by the debug options turned on in rawhide / test release kernels. So I did a local rebuild of the kernel with the debug options turned off and my problems disappeared.

I would have expected this to make a difference of maximum 10% or so, but in certain circumstances the difference is a lot larger, see bug 630955 for more on this.

Regards,

Hans

*** This bug has been marked as a duplicate of bug 630955 ***

Comment 16 Dennis Jacobfeuerborn 2010-09-09 21:55:18 UTC
Nope, unfortunately that doesn't seem to be the problem. I just downloaded the official F13 x86_64 live cd iso and I see the same performance issues as before.

Comment 17 Dennis Jacobfeuerborn 2010-10-10 00:17:41 UTC
Has anyone be able to take a close look at the profile?
I've now install the F14 Beta on a 10G partition to make testing easier.

One other thing I have now noticed is that doing "top" in a shell shows that top itself eats up 9% cpu. On the F11 install it uses 1%. 
During video playback Xorg uses 20% cpu on F14 Beta but only 5% on F11.

Does anyone have an idea what could cause such a behavior? Are there kernel options that could influence the things I'm seeing?

(I'm beginning to suspect that this bug might actually be a kernel problem rather than an issue with the radeon driver)

Comment 18 Fedora End Of Life 2012-08-16 21:39:25 UTC
This message is a notice that Fedora 14 is now at end of life. Fedora 
has stopped maintaining and issuing updates for Fedora 14. It is 
Fedora's policy to close all bug reports from releases that are no 
longer maintained.  At this time, all open bugs with a Fedora 'version'
of '14' have been closed as WONTFIX.

(Please note: Our normal process is to give advanced warning of this 
occurring, but we forgot to do that. A thousand apologies.)

Package Maintainer: If you wish for this bug to remain open because you
plan to fix it in a currently maintained version, feel free to reopen 
this bug and simply change the 'version' to a later Fedora version.

Bug Reporter: Thank you for reporting this issue and we are sorry that 
we were unable to fix it before Fedora 14 reached end of life. If you 
would still like to see this bug fixed and are able to reproduce it 
against a later version of Fedora, you are encouraged to click on 
"Clone This Bug" (top right of this page) and open it against that 
version of Fedora.

Although we aim to fix as many bugs as possible during every release's 
lifetime, sometimes those efforts are overtaken by events.  Often a 
more recent Fedora release includes newer upstream software that fixes 
bugs or makes them obsolete.

The process we are following is described here: 
http://fedoraproject.org/wiki/BugZappers/HouseKeeping


Note You need to log in before you can comment on or make changes to this bug.