very bad performance
Mihai Donțu
mihai.dontu at gmail.com
Wed Aug 8 15:16:33 PDT 2012
On Wed, 08 Aug 2012 19:35:12 +0100 Chris Wilson wrote:
> On Wed, 8 Aug 2012 16:42:36 +0300, Mihai Donțu wrote:
> > On Tue, 07 Aug 2012 17:28:40 +0100 Chris Wilson wrote:
> > > There have bugs where the cacheability flags of the pages in
> > > memory have been lost over suspend and resume, and where the GPU
> > > became stuck in a low power mode. So such degradation of
> > > performance over time is not unheard of. One of the key questions
> > > is if we have no errors reported how do the performance
> > > characteristics of the system vary over time? Running x11perf
> > > along with the benchmarks from intel-gpu-tools would be a good
> > > way to get some results that should help give some insight as to
> > > what changes. Along with comparing perf profiles over time.
> > >
> >
> > Are there any specific x11perf tests I should be looking at? Or
> > should I just script the run of all supported tests? There are
> > many. :-)
>
> "-copywinwin500 -putimage500 -aa10text" should give enough coverage to
> suggest the next step.
>
> > I'm looking into intel-gpu-tools too. Btw: how can I find out if the
> > GPU is stuck in a low power mode?
>
> cat /sys/kernel/debug/dri/0/i915_drpc_info, the CAGF field is the
> current frquency (presuming you have a GPU that supports dynamic
> clocking).
>
Looks like I'm missing that field:
$ cat /sys/kernel/debug/dri/0/i915_drpc_info
RC information accurate: yes
Video Turbo Mode: yes
HW control enabled: yes
SW control enabled: no
RC1e Enabled: no
RC6 Enabled: yes
Deep RC6 Enabled: no
Deepest RC6 Enabled: no
Current RC state: RC6
Core Power Down: no
RC6 "Locked to RPn" residency since boot: 0
RC6 residency since boot: 3599731174
RC6+ residency since boot: 0
RC6++ residency since boot: 0
xorg sees the device as: "Integrated Graphics Chipset: Intel(R)
Sandybridge Mobile (GT2+)"
while lspci says: "00:02.0 VGA compatible controller: Intel Corporation
2nd Generation Core Processor Family Integrated Graphics Controller
(rev 09)"
I suppose that this particular GPU does not support dynamic clocking. :|
--
Mihai Donțu
More information about the xorg
mailing list