[Mesa-dev] [PATCH] intel: Use a CPU map of the batch on LLC-sharing architectures.

[Kenneth Graunke]

On 01/25/2013 09:31 AM, Eric Anholt wrote:
> Kenneth Graunke <kenneth at whitecape.org> writes:
>
>> On 01/20/2013 02:59 PM, Eric Anholt wrote:
>>> Before, we were keeping a CPU-only buffer to accumulate the batchbuffer in,
>>> which was an improvement over mapping the batch through the GTT directly
>>> (since any readback or other failure to stream through write combining
>>> correctly would hurt).  However, on LLC-sharing architectures we can do better
>>> by mapping the batch directly, which reduces the cache footprint of the
>>> application since we no longer have this extra copy of a batchbuffer around.
>>>
>>> Improves performance of GLBenchmark 2.1 offscreen on IVB by 3.5% +/- 0.4%
>>> (n=21).  Improves Lightsmark performance by 1.1 +/- 0.1% (n=76).  Improves
>>> cairo-gl performance by 1.9% +/- 1.4% (n=57).
>>>
>>> No statistically significant difference in GLB2.1 on SNB (n=37).  Improves
>>> cairo-gl performance by 2.1% +/- 0.1% (n=278).
>>
>> Looks good to me.  Have you tested this on a non-LLC machine?
>
> Not in a long time.  It shouldn't affect performance, since they get the
> same behavior as before.

Okay.  I mashed has_llc to false and ran glxgears, which was proof 
enough that the non-LLC path still works.  I figured it did, but just 
wanted to check.