[Mesa-dev] [PATCH 2/5] i965/vec4: adding vec4_cmod_propagation optimization

Matt Turner mattst88 at gmail.com
Thu Oct 15 08:57:11 PDT 2015


On Wed, Oct 14, 2015 at 12:51 AM, Alejandro PiƱeiro
<apinheiro at igalia.com> wrote:
> On 13/10/15 23:36, Matt Turner wrote:
>> The good news is that, unless I've done something wrong, this doesn't
>> affect any shaders in shader-db on ILK or HSW (I only tested those
>> two, but I expect the results are the same everywhere). Apparently
>> this is a pretty rare case.
>
> Are you sure? I have made a run adding your condition, and now comparing
> master vs having the optimization I get this:
> total instructions in shared programs: 6240631 -> 6240471 (-0.00%)
> instructions in affected programs:     18965 -> 18805 (-0.84%)
> helped:                                160
> HURT:                                  0
>
> That is a really small gain. Or put in other way, if we compare the
> conditions I have on the original patches vs adding the condition you
> are proposing, I get this:
>
> total instructions in shared programs: 6223900 -> 6240471 (0.27%)
> instructions in affected programs:     477537 -> 494108 (3.47%)
> helped:                                0
> HURT:                                  3047

Strange, I must have made a mistake in my shader-db results. Your
results certainly make more sense.


More information about the mesa-dev mailing list