[Mesa-dev] Let's identify the R9 285 bottleneck in respect to R7 370

[Mike Lothian]

Ah it's not a regression, in that case it might be possible to profile the
running benchmark to see where some bottlenecks lie, but from what ive read
on here before, this is difficult to set up and the results aren't always
useful.

As you're unaffected by this I doubt you'll be able to help further, I take
it your hardware doesn't have these issues?

On Mon, 20 Jun 2016, 3:03 p.m. ⚛, <0xe2.0x9a.0x9b at gmail.com> wrote:

> On Mon, Jun 20, 2016 at 3:36 PM, Nicolai Hähnle <nhaehnle at gmail.com>
> wrote:
> > On 20.06.2016 11:59, wrote:
> >>
> >> Hello.
> >>
> >> The result file at
> >> http://openbenchmarking.org/result/1606180-PTS-MIDJUNEA78 is showing
> >> multiple issues, some of them are:
> >>
> >> [Minor issue] Kernel 4.7.0-999-generic (x86_64) 20160616 has a
> >> performance regression on R9 290
> >>
> >> [Major issue] R7 370 yields 133 FPS in Tesseract, while R9 285
> >> unexpectedly yields 111.81 FPS in Tesseract
> >
> > It would be most helpful if people with the affected hardware could first
> > identify the responsible component and then bisect.
>
> About the R9 285 issue:
>
> I don't understand. How might a person bisect something that doesn't exist?
>
> Or, if I am in error and you are right, and bisecting is possible then
> I would like to see a previous Tesseract v2014-05-12 result with >=
> 140 FPS on R9 285.
>
> There was a point in time when Mesa was able to run Tesseract at >=
> 140 FPS on R9 285?
>
> > Cheers,
> > Nicolai
> >
> >> R7 370: 1996.8 GFLOPS
> >> R9 285: 3290 GLOPS
> >>
> >> I would like this forum thread to identify the cause of the major
> >> issue. So the question is:
> >>
> >> *** What might be causing R9 285 to perform worse than R7 370? ***
> >>
> >> (I don't have access to R7 370 nor R9 285. My current answer to the
> >> question is: I don't know.)
> >> _______________________________________________
> >> mesa-dev mailing list
> >> mesa-dev at lists.freedesktop.org
> >> https://lists.freedesktop.org/mailman/listinfo/mesa-dev
> >>
> >
> _______________________________________________
> mesa-dev mailing list
> mesa-dev at lists.freedesktop.org
> https://lists.freedesktop.org/mailman/listinfo/mesa-dev
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.freedesktop.org/archives/mesa-dev/attachments/20160620/c622b8e1/attachment.html>