[Mesa-dev] [PATCH] [RFC] r600g: improve flushed depth texture allocation

[Andy Furniss]

Vadim Girlin wrote:
> On Mon, 2012-06-25 at 12:19 +0200, Andreas Boll wrote:
>> 2012/6/24 Vadim Girlin <vadimgirlin at gmail.com>:
>>> On Sun, 2012-06-24 at 17:43 +0400, Vadim Girlin wrote:
>>>> Allocate flushed depth texture in the VRAM if we aren't going to access it by
>>>> CPU. If we need CPU access later, then it'll be reallocated in the GTT.
>>>> Currently it's not reallocated in the opposite direction (GTT->VRAM), though
>>>> probably we might want to do it too. Anyway, it helps the apps that don't need
>>>> to access flushed depth texture by CPU at all, e.g. Lightsmark.
>>>>
>>>> Improves performance for Lightsmark.
>>>>
>>>> Signed-off-by: Vadim Girlin <vadimgirlin at gmail.com>
>>>> ---
>>>
>>> You can also find this patch and some other patches that aren't
>>> committed yet (shader caching, dual export) in the r600-perf branch in
>>> my github mesa repo:
>>>
>>> git://github.com/VadimGirlin/mesa.git
>>>
>>> www: https://github.com/VadimGirlin/mesa/compare/r600-perf
>>>
>>> Could anybody test them for regressions on non-evergreen cards
>>> (R6xx/7xx/CAYMAN)?
>>>
>>> Vadim
>>>
>>> _______________________________________________
>>> mesa-dev mailing list
>>> mesa-dev at lists.freedesktop.org
>>> http://lists.freedesktop.org/mailman/listinfo/mesa-dev
>>
>> following piglit tests regresses on my rv770:
>>
>
> Thanks. I've updated the branch with a possible fix, could you test it
> again?

With or without the fix my limited testing has found one fail case vs 
master.

Using vdpau for mpeg2 decode is corrupt - you can still see what it's 
supposed to be it's just blocky.

mesa demos, nexuiz, etqw and vdpau display look OK.