[i845] xorg intel driver 2.7.99 36% slower than 2.7.0 on i845

Latest results from karmic 2.6.30.10.10:

I'm not sure what KMS is for on intel video graphics?

i830 Thinkpad R31 boots to blackscreen. nomodeset in kernel line turns off KMS then boots O.K. but of course KMS is turned off. That's another launchpad bug.

i845 ThinkCentre 2.0 gHz Celeron works O.K. but Noticeably slower than generic video driver vesa, for example GtkPerf from Synaptic:

06/24/09 karmic 2.6.30.10.10
karmic.i845_________ vesa__KMS=1__ KMS_slower
Entry_________________ 0.22 0.26 18%
ComboBox____________ 5.55 9.25 67%
ComboBoxEntry_______ 3.84 5.52 44%
SpinButton____________ 0.85 2.5 194%
ProgressBar___________ 0.82 3.97 384%
ToggleButton__________ 0.95 2.47 160%
CheckButton___________ 0.60 1.10 83%
RadioButton____________ 0.87 1.35 55%
TextView-Add-text______ 1.93 1.83 -5%
TextView-Scroll_________ 1.29 1.97 53%
DrawingArea-Lines______ 2.11 1.83 -13%
DrawingArea-Circles____ 4.15 9.81 136%
DrawingArea-Text_______ 4.77 5.53 16%
DrawingArea-Pixbufs____ 0.26 0.51 96%

Total_time_____________ 28.24 47.92 70%

Progress bar in particular visibly drags across the screen....

? Thanks, Jerry

Using Ubuntu Jaunty - Ubuntu 9.04, kernel 2.6.30-020630
intel video graphics xserver-xorg-video-intel-2.7.1
freezes running googleearth 5.0.11733.9347
Graphic Card Intel Corporation 82845G/GL[Brookdale-G]/GE

glxgears gives me
1402 frames in 5.0 seconds = 280.367 FPS
1438 frames in 5.0 seconds = 287.416 FPS
1956 frames in 5.0 seconds = 390.953 FPS

moving x-windows really slow.

using as office desktop (open-office and printing) is OK.

thanx, Marcelo

Hi jerrylamos,

I've forwarded your bug upstream to https://bugs.freedesktop.org/show_bug.cgi?id=22759 - I've subscribed you to it in case upstream needs further info.

Two more things I'd ask you to do:

First, please attach to the upstream bug report an example Xorg.0.log from one of the 2.7.99 cases, as it may have info upstream needs for debugging the problem.

Second, could you re-run your tests against current updated Karmic with the 2.6.31 kernel? There were a *ton* of -intel driver fixes in .31 that fixed regressions seen in .30, so it is important to verify whether this issue is still present in .31.

Bryce,

Results with 2.6.31-2. Formatting columnar data is a bit of a bear on launchpad so I'll attach the spreadsheet. I also have results on IBM Thinkpad R31 i830 and IBM Thinkpad T40 Radeon Mobility 7400 which I hope to post tomorrow.

Jerry

1st results column: jauntyCD.i845.2.6.3.-0ubuntu9
2nd results column: karmic.2.26.31-2.i845.vesa
3rd results column: karmic.i845.2.27.99.901+git20090702.74227141-0ubuntu1

GtkPerf 0.40Startingtesting: Mon Jul 13 12:12:44 AM 2009
jauntyCD.i845.2.6.3.-0ubuntu9 jaunty Karmic vesa difference 2.79901 vs. jaunty vs. vesa
GtkEntrytime: 0.28 0.25 --------- 0.27 -4% 8%
GtkComboBoxtime: 6.66 6.22 --------- 7.21 8% 16%
GtkComboBoxEntrytim: 4.12 4.56 --------- 4.68 14% 3%
GtkSpinButtontime: 1.05 0.93 --------- 1.98 89% 113%
GtkProgressBartime: 1.98 0.85 --------- 2.08 5% 145%
GtkToggleButtontime: 1.1 1.04 --------- 1.55 41% 49%
GtkCheckButtontime: 0.85 0.64 --------- 1.22 44% 91%
GtkRadioButtontime: 1.6 0.96 --------- 1.12 -30% 17%
GtkTextViewAddtextti: 1.96 2.04 --------- 1.85 -6% -9%
GtkTextViewScrolltime: 1.57 1.5 --------- 1.61 3% 7%
GtkDrawingAreaLines: 2.85 3.07 --------- 3.05 7% -1%
GtkDrawingAreaCircles: 2.82 5.25 --------- 11.90 322% 127%
GtkDrawingAreaTextt: 6.09 5.65 --------- 6.44 6% 14%
GtkDrawingAreaPixbuf: 1.07 0.28 --------- 0.52 -51% 86%
 --- ---------
Totaltime: 34.02 33.26 --------- 45.49 34% 37%

Jerry, I've packaged the patch that upstream says should fix this issue and posted it to this PPA for you:

https://edge.launchpad.net/~bryceharrington/+archive/blue

Please test and see whether that fixes it for you, and provide feedback on the upstream bug.

If it does fix it, we possibly may be able to include it for karmic, however anholt's comment in the patch makes it sound like it can cause a hang on i865, which sounds sub-optimal, so if that's the case I'd want to see that issue resolved before incorporating this patch.

Bryce, thanks for the PPA. I can try to put it on however Alpha 3 i845 doesn't boot with intel driver 2.8.0 yet so I can't measure performance. Unless the PPA fixes the boot problem about all I know how to do for measurements is to install an Alpha 2 and put the PPA on that.
Thanks, Jerry

Bryce, I'm stuck.

I installed a karmic A2 which will still boot. It's got intel driver 2.7.99.901+git....

Re-measured GtkPerf.

Added the ppa lines to sources.list, accepted the key, added third party to sources.

Now it says: sudo apt-get update

to add the ppa. That will also add a ton of stuff including intel driver 2.8.0 and I won't be able to boot any more, that's bug #406460.

How do I update just what you have in the PPA?

Any ideas? Thanks, Jerry

GtkPerf0.40Startingtesting:TueAug0414:01:112009
Well, the PPA works, in KMS mode, and dpkg says it is
xserver-xorg-video-intel 2:2.8.0-0ubuntu2~bug382017~1

however it isn't any faster than 2:2.799 for example compare to vesa nomodeset on the same karmic and same i845 hardware:

karmicA2.noKMS.i845.vesa vesa PPA difference
GtkEntrytime:_____________ 0.26 0.46 77%
GtkComboBoxtime:________ 6.03 7.95 32%
GtkComboBoxEntrytime:____ 4.23 4.67 10%
GtkSpinButtontime:________ 0.94 2.3 145%
GtkProgressBartime:_______ 0.88 2.12 141%
GtkToggleButtontime:______ 1.06 1.57 48%
GtkCheckButtontime:_______ 0.67 0.95 42%
GtkRadioButtontime:_______ 0.98 1.11 13%
GtkTextViewAddtexttime:___ 2.05 1.89 -8%
GtkTextViewScrolltime:_____ 1.53 1.62 6%
GtkDrawingAreaLinestime:___ 3.3 2.8 -15%
GtkDrawingAreaCirclestime:_ 6.21 12.58 103%
GtkDrawingAreaTexttime:___ 5.46 6.35 16%
GtkDrawingAreaPixbufstime:_ 0.27 0.55 104%
 ---
Total time:________________ 33.9 46.96 39%

That's PPA 39% slower than vesa.

vesa is about the same speed as jaunty default xorg.

I did an A2 install and put the PPA on. It is a 2.8.0, and It does boot on A2, whereas 2.8.0 does not boot on A3 on i830 or i845.

The PPA didn't boot on A3 either. A bit of a sweat for me at my level of experience to install a PPA from recovery mode root prompt since A3 gnome desktop hangs bug #406460.

I tried copying intel driver 2.7.99 on A3 however this was no help. Maybe I need to install it if I knew how.

Thanks for the PPA, Jerry

Hi Jerry, thanks for all your hard work on this, looks like upstream came up with a patch we can test out in karmic. I'll upload it shortly.

This bug was fixed in the package xserver-xorg-video-intel - 2:2.8.1-1ubuntu1

---------------
xserver-xorg-video-intel (2:2.8.1-1ubuntu1) karmic; urgency=low

  * Merge with Debian. Remaining Ubuntu changes:
    + control: Add lpia architecture
    + Fixes bad chroma upsampling on videos (LP: #371605)
  * Add 100_8xx_perf_pict_a8.patch: Improve performance on i845 by adding
    limited support for a8 dests. Cherrypick from upstream master tree.
    (LP: #382017)
  * Add 101_reload_cursors.patch: Reload cursors as needed when setting
    new modes. Fixes issue where cursor doesn't rotate when the screen
    is rotated. Cherrypick from upstream master tree.
    (LP: #410255)

 -- Bryce Harrington <email address hidden> Wed, 26 Aug 2009 23:52:33 -0700

Update installed 2:2.8.1-1ubuntu1 karmic I think.

Hangs on boot with brown screen and ubuntu logo. Seems to me very similar to bug #406460 with different artwork.

Will attach dmesg, Xorg.0.log, dpkg -l, .xsession-errors which were copied from failed image by doing a dual boot.

Rebooted with vesa to attach the files. Note, the screens while booting are in 48 line mode like kms.

Any suggestions?

Thanks, Jerry

dmesg at hang

Xorg from hang with 2.8.1-1ubuntu1

xsession-errors

On Fri, Aug 28, 2009 at 02:05:22AM -0000, jerrylamos wrote:
> Update installed 2:2.8.1-1ubuntu1 karmic I think.
>
> Hangs on boot with brown screen and ubuntu logo.
> Any suggestions?

Please keep different issues to their own bug reports. This one is
about a performance regression, so the hang should be done as a new bug
report. For the hang check http://wiki.ubuntu.com/X/Troubleshooting
for debugging tips.

  "Launchpad Janitor wrote on 2009-08-27:
This bug was fixed in the package xserver-xorg-video-intel - 2:2.8.1-1ubuntu1"

There's no way to evaluate the performance regression with this package since it doesn't boot on i845, not even in nomodeset.

Jerry

Jerry, think Bryce realises that, he's saying can we have a hang bug reported seperately, get that fixed, and then come back to this bug. Does that make sense?

Alternate theory meanwhile: kernel 31-RC8 hangs on boot for me, 6 and 7 were fine. So if you've moved forward to a later RC since your original report, it could be the kernel piece that's causing the problem.

Zack, I think it was the kernel. Karmic RC is running now with kernel 2.6.31.14-generic, GtkPerf results with 2.9.0 some 28% slower than vesa with video GtkPerf from Synaptic. Bigget problem is Drawing Circles:

karmicRC.i845 2.0 gHz Celeron IBM ThinkCentre
GtkPerf0.40-Nov.2.20:38:14.2009
kernel-2.6.31-14-generic intel.driver-2.9.0 i845 vesa 2.9.0 slower than vesa
GtkEntry-time:____________ 0.4 0.22 82%
GtkComboBox-time:________ 6.19 5.69 9%
GtkComboBoxEntry-time:___ 3.29 2.74 20%
GtkSpinButton-time:________ 0.9 0.71 27%
GtkProgressBar-time:_______ 1.67 0.99 69%
GtkToggleButton-time:______ 1.03 0.94 10%
GtkCheckButton-time:______ 0.79 0.58 36%
GtkRadioButton-time:_______ 1.53 1.16 32%
GtkTextView-Add-text-time:__ 2.17 2.48 -13%
GtkTextView-Scroll-time:____ 1.25 1.38 -9%
GtkDrawingArea-Lines-time:_ 2.41 3.05 -21% a bit faster
GtkDrawingArea-Circles-time: 11.27 4.79 135% much slower
GtkDrawingArea-Text-time:__ 5.75 5.51 4%
GtkDrawingArea-Pixbufs-time: 0.52 0.29 79%
 ---
Total-time: 39.2 30.58 28% Intel 2.9.0 slower

Jerry

In previous comment, launchpad wasn't designed for columnar data. 1st column is 2.90, second column Vesa, 3rd is 2.90 slower than Vesa by X%.

Sorry 'bout that.

Jerry