5:32 p.m.
On Tue, Jun 2, 2009 at 2:44 PM, Kevin DeKorte wrote:
On 06/02/2009 03:37 PM, David L wrote:
> My f11 system seems extremely slow when running
> some 2D gtk/cairo apps. Is there a benchmark
> suite that is yum installable for testing X performance?
<snip>
>
Personally I find that gtkperf is useful although not an exact science.
I like to run it with the following options
gtkperf -c 500 -a
Thanks Kevin,
That helped confirm my suspicions. Here's my output with
the intel driver on my Intel 82865G:
GtkPerf 0.40 - Starting testing: Tue Jun 2 15:06:29 2009
GtkEntry - time: 0.47
GtkComboBox - time: 6.78
GtkComboBoxEntry - time: 3.31
GtkSpinButton - time: 2.38
GtkProgressBar - time: 1.16
GtkToggleButton - time: 1.50
GtkCheckButton - time: 0.45
GtkRadioButton - time: 1.21
GtkTextView - Add text - time: 3.39
GtkTextView - Scroll - time: 3.55
GtkDrawingArea - Lines - time: 1.76
GtkDrawingArea - Circles - time: 44.60
GtkDrawingArea - Text - time: 15.80
GtkDrawingArea - Pixbufs - time: 1.38
---
Total time: 87.74
Same computer using the vesa driver:
GtkPerf 0.40 - Starting testing: Tue Jun 2 15:19:09 2009
GtkEntry - time: 0.14
GtkComboBox - time: 1.77
GtkComboBoxEntry - time: 1.29
GtkSpinButton - time: 0.30
GtkProgressBar - time: 0.20
GtkToggleButton - time: 0.33
GtkCheckButton - time: 0.12
GtkRadioButton - time: 0.28
GtkTextView - Add text - time: 1.01
GtkTextView - Scroll - time: 0.53
GtkDrawingArea - Lines - time: 0.61
GtkDrawingArea - Circles - time: 1.43
GtkDrawingArea - Text - time: 1.31
GtkDrawingArea - Pixbufs - time: 0.10
---
Total time: 9.42
I was running at a higher resolution using the intel driver,
but it's nearly 10x slower. Is this expected from this
driver? Until very recently it was hanging my system,
so maybe it's still not mature?
Thanks,
David
5:38 p.m.
On Tue, 2009-06-02 at 15:32 -0700, David L wrote:
On Tue, Jun 2, 2009 at 2:44 PM, Kevin DeKorte wrote:
> On 06/02/2009 03:37 PM, David L wrote:
>> My f11 system seems extremely slow when running
>> some 2D gtk/cairo apps. Is there a benchmark
>> suite that is yum installable for testing X performance?
<snip>
>>
>
> Personally I find that gtkperf is useful although not an exact science.
> I like to run it with the following options
>
> gtkperf -c 500 -a
>
Thanks Kevin,
That helped confirm my suspicions. Here's my output with
the intel driver on my Intel 82865G:
[snip various results]
GtkDrawingArea - Circles - time: 44.60
GtkDrawingArea - Text - time: 15.80
[snip]
Same computer using the vesa driver:
GtkPerf 0.40 - Starting testing: Tue Jun 2 15:19:09 2009
[snip]
GtkDrawingArea - Circles - time: 1.43
GtkDrawingArea - Text - time: 1.31
[snip]
Looks like the circles/text tests are the most obvious differences,
though most tests show marked differences.
I was running at a higher resolution using the intel driver,
but it's nearly 10x slower. Is this expected from this
You may want to try
"sysprof"; it's an excellent tool for figuring out
where slowdowns are occurring across the whole system (e.g. of the
combined gtk app/X server system, as in this case). See:
http://live.gnome.org/Sysprof
driver? Until very recently it was hanging my system,
so maybe it's still not mature?
(I have no idea about this)
Hope this is helpful
Dave
9:46 a.m.
On Tue, 2009-06-02 at 18:38 -0400, David Malcolm wrote:
On Tue, 2009-06-02 at 15:32 -0700, David L wrote:
> On Tue, Jun 2, 2009 at 2:44 PM, Kevin DeKorte wrote:
> > On 06/02/2009 03:37 PM, David L wrote:
> >> My f11 system seems extremely slow when running
> >> some 2D gtk/cairo apps. Is there a benchmark
> >> suite that is yum installable for testing X performance?
> <snip>
> >>
> >
> > Personally I find that gtkperf is useful although not an exact science.
> > I like to run it with the following options
> >
> > gtkperf -c 500 -a
> >
> Thanks Kevin,
>
> That helped confirm my suspicions. Here's my output with
> the intel driver on my Intel 82865G:
KMS or UMS?
[snip various results]
> GtkDrawingArea - Circles - time: 44.60
> GtkDrawingArea - Text - time: 15.80
[snip]
> Same computer using the vesa driver:
> GtkPerf 0.40 - Starting testing: Tue Jun 2 15:19:09 2009
[snip]
>GtkDrawingArea - Circles - time: 1.43
> GtkDrawingArea - Text - time: 1.31
[snip]
Looks like the circles/text tests are the most obvious differences,
though most tests show marked differences.
As with everything, it helps to know what you're measuring.
Circles is the wide arc rendering path in the X server. It's
essentially unused by gtk apps in general, but gtkperf does it anyway.
The arcs specified by the X protocol are insanely ugly (which is why
nobody uses them in real apps) and also not a hardware-accelerated
primitive. We could break them down to spans inside the X server and
accelerate filling those spans, but we don't.
So they happen in software. Which is also true for the vesa driver, but
there's a catch. The vesa driver uses a trick called 'shadowfb', where
the whole screen is rendered in (cached) host memory and then the
updated regions are uploaded to the actual scanout memory. This is
adequately fast, because it minimizes the number of memory cycles (read
cycles in particular) that you do to the framebuffer, which is typically
uncached.
In the intel driver, it's a different story. We don't keep a shadow, so
the software fallback happens either cached or uncached, depending how
we map the framebuffer. If it's cached, you have to do a big cache
flush when you finish rendering so the bits actually make it out from
the CPU's cache to the framebuffer. If it's uncached, you're hitting
main memory on every cycle, which is also not great.
I don't remember offhand whether the text test is using Render or the
old core font path. If the latter, then the same scenario applies; it's
not accelerated (because it's actually rather hard to accelerate well),
and the software path can't help but suck.
For the other stuff, 865 and 855 appear to have a chipset bug where the
command buffer doesn't always flush to the GPU reliably, so we have to
flush the entire CPU cache on every acceleration command:
http://cvs.fedoraproject.org/viewvc/rpms/kernel/F-11/drm-intel-big-hammer...
We do try to mitigate this by batching up big sequences of commands
rather than lots of little ones, but it still hurts.
So, two things. Compare gtkperf results with different resolutions on
the intel driver; tests which are hitting the software fallback path
will be faster at smaller resolutions, because there's less framebuffer
to clflush out. Also, figure out whether you're using KMS or not, and
try the other way.
- ajax
10:18 a.m.
On Wed, Jun 3, 2009 at 7:46 AM, Adam Jackson wrote:
On Tue, 2009-06-02 at 18:38 -0400, David Malcolm wrote:
> On Tue, 2009-06-02 at 15:32 -0700, David L wrote:
> > On Tue, Jun 2, 2009 at 2:44 PM, Kevin DeKorte wrote:
> > > On 06/02/2009 03:37 PM, David L wrote:
> > >> My f11 system seems extremely slow when running
> > >> some 2D gtk/cairo apps. Is there a benchmark
> > >> suite that is yum installable for testing X performance?
> > <snip>
> > >>
> > >
> > > Personally I find that gtkperf is useful although not an exact science.
> > > I like to run it with the following options
> > >
> > > gtkperf -c 500 -a
I forgot to say that I didn't run with the options that Kevin suggested.
I just ran gtkperf with no options and clicked start. Using Kevin's
options, I got this with the intel driver with KMS:
GtkEntry - time: 2.91
GtkComboBox - time: 32.25
GtkComboBoxEntry - time: 15.41
GtkSpinButton - time: 12.25
GtkProgressBar - time: 5.25
GtkToggleButton - time: 7.11
GtkCheckButton - time: 1.85
GtkRadioButton - time: 5.79
GtkTextView - Add text - time: 33.95
GtkTextView - Scroll - time: 19.05
GtkDrawingArea - Lines - time: 9.57
GtkDrawingArea - Circles - time: 223.77
GtkDrawingArea - Text - time: 77.81
GtkDrawingArea - Pixbufs - time: 3.74
---
Total time: 450.72
KMS or UMS?
With no options to gtkperf I got these results:
intel driver KMS 1680x1050
GtkEntry - time: 0.47
GtkComboBox - time: 6.78
GtkComboBoxEntry - time: 3.31
GtkSpinButton - time: 2.38
GtkProgressBar - time: 1.16
GtkToggleButton - time: 1.50
GtkCheckButton - time: 0.45
GtkRadioButton - time: 1.21
GtkTextView - Add text - time: 3.39
GtkTextView - Scroll - time: 3.55
GtkDrawingArea - Lines - time: 1.76
GtkDrawingArea - Circles - time: 44.60
GtkDrawingArea - Text - time: 15.80
GtkDrawingArea - Pixbufs - time: 1.38
---
Total time: 87.74
intel driver KMS 1280x1024:
GtkEntry - time: 0.49
GtkComboBox - time: 6.90
GtkComboBoxEntry - time: 4.83
GtkSpinButton - time: 2.43
GtkProgressBar - time: 1.16
GtkToggleButton - time: 2.04
GtkCheckButton - time: 0.46
GtkRadioButton - time: 1.17
GtkTextView - Add text - time: 3.42
GtkTextView - Scroll - time: 3.85
GtkDrawingArea - Lines - time: 1.65
GtkDrawingArea - Circles - time: 45.98
GtkDrawingArea - Text - time: 16.29
GtkDrawingArea - Pixbufs - time: 1.05
---
Total time: 91.73
intel driver nomodeset 1680x1050 :
GtkEntry - time: 0.36
GtkComboBox - time: 3.87
GtkComboBoxEntry - time: 2.58
GtkSpinButton - time: 1.33
GtkProgressBar - time: 0.66
GtkToggleButton - time: 1.02
GtkCheckButton - time: 0.29
GtkRadioButton - time: 0.88
GtkTextView - Add text - time: 2.55
GtkTextView - Scroll - time: 2.25
GtkDrawingArea - Lines - time: 1.51
GtkDrawingArea - Circles - time: 14.21
GtkDrawingArea - Text - time: 11.44
GtkDrawingArea - Pixbufs - time: 1.02
---
Total time
43.97
vesa driver KMS (1280x1024??):
GtkEntry - time: 0.14
GtkComboBox - time: 1.77
GtkComboBoxEntry - time: 1.29
GtkSpinButton - time: 0.30
GtkProgressBar - time: 0.20
GtkToggleButton - time: 0.33
GtkCheckButton - time: 0.12
GtkRadioButton - time: 0.28
GtkTextView - Add text - time: 1.01
GtkTextView - Scroll - time: 0.53
GtkDrawingArea - Lines - time: 0.61
GtkDrawingArea - Circles - time: 1.43
GtkDrawingArea - Text - time: 1.31
GtkDrawingArea - Pixbufs - time: 0.10
---
Total time: 9.42
I didn't try the vesa driver with nomodeset.
Cheers,
David
5:59 p.m.
New subject: intel graphics driver slow (was: slightly OT - X performance benchmark)
-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1
On 06/02/2009 04:32 PM, David L wrote:
On Tue, Jun 2, 2009 at 2:44 PM, Kevin DeKorte wrote:
> On 06/02/2009 03:37 PM, David L wrote:
>> My f11 system seems extremely slow when running
>> some 2D gtk/cairo apps. Is there a benchmark
>> suite that is yum installable for testing X performance?
<snip>
>>
>
> Personally I find that gtkperf is useful although not an exact science.
> I like to run it with the following options
>
> gtkperf -c 500 -a
>
Thanks Kevin,
That helped confirm my suspicions. Here's my output with
the intel driver on my Intel 82865G:
GtkPerf 0.40 - Starting testing: Tue Jun 2 15:06:29 2009
GtkEntry - time: 0.47
GtkComboBox - time: 6.78
GtkComboBoxEntry - time: 3.31
GtkSpinButton - time: 2.38
GtkProgressBar - time: 1.16
GtkToggleButton - time: 1.50
GtkCheckButton - time: 0.45
GtkRadioButton - time: 1.21
GtkTextView - Add text - time: 3.39
GtkTextView - Scroll - time: 3.55
GtkDrawingArea - Lines - time: 1.76
GtkDrawingArea - Circles - time: 44.60
GtkDrawingArea - Text - time: 15.80
GtkDrawingArea - Pixbufs - time: 1.38
---
Total time: 87.74
Those numbers are actually pretty acceptable... when my radeon card
didn't have dri running I was seeing numbers about 150 and it was still
quite usable.
It is interesting that the vesa driver is much faster. Perhaps there is
other magic happening to make it appear faster.
Kevin
- --
Get my public GnuPG key from
http://pgp.mit.edu:11371/pks/lookup?op=get&search=0x7D0BD5D1
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.9 (GNU/Linux)
Comment: Using GnuPG with Fedora - http://enigmail.mozdev.org/
iEYEARECAAYFAkolru4ACgkQ6w2kMH0L1dF6XQCfWjSLuMFIwBKXXifetV+/qfHU
w3kAnjpUgbLbacDWOMFeofa7idGvHkHd
=XyIW
-----END PGP SIGNATURE-----
10:01 a.m.
<snip>
> Personally I find that gtkperf is useful although not an exact
science.
Here's my gtkperf output with the intel driver on my Intel 82865G:
<snip>
Total time: 87.74
Same computer using the vesa driver:
<snip>
Total time: 9.42
<snip>
FYI, I filed this bug report since this system is at least 10x
slower than it should be wrt graphics performance:
https://bugzilla.redhat.com/show_bug.cgi?id=504048
Thanks,
David
5493
days inactive
5495
days old
5 comments
4 participants
participants (4)
-
Adam Jackson -
David L -
David Malcolm -
Kevin DeKorte