On Tue, Mar 17, 2015 at 10:56 PM, Matt Turner wrote:
> Thanks. Looked through stats and at some of the regressions.
I disabled the pow(x, y) optimization in NIR, turned off the SEL
peephole in the fs backend, and added neg/abs to
block_check_for_allowed_instrs() (will test and send a patch) to ma
On Wed, Mar 18, 2015 at 9:52 AM, Matt Turner wrote:
> On Wed, Mar 18, 2015 at 8:27 AM, Jason Ekstrand wrote:
>> On Wed, Mar 18, 2015 at 7:22 AM, Connor Abbott wrote:
>>> On Wed, Mar 18, 2015 at 1:56 AM, Matt Turner wrote:
Thanks. Looked through stats and at some of the regressions.
>>
On Wed, Mar 18, 2015 at 8:27 AM, Jason Ekstrand wrote:
> On Wed, Mar 18, 2015 at 7:22 AM, Connor Abbott wrote:
>> On Wed, Mar 18, 2015 at 1:56 AM, Matt Turner wrote:
>>> Thanks. Looked through stats and at some of the regressions.
>>>
>>> Some of the areas I noticed we were doing worse:
>>>
>>>
On Wed, Mar 18, 2015 at 11:27 AM, Jason Ekstrand wrote:
> On Wed, Mar 18, 2015 at 7:22 AM, Connor Abbott wrote:
>> On Wed, Mar 18, 2015 at 1:56 AM, Matt Turner wrote:
>>> Thanks. Looked through stats and at some of the regressions.
>>>
>>> Some of the areas I noticed we were doing worse:
>>>
>>>
On Wed, Mar 18, 2015 at 11:27 AM, Jason Ekstrand wrote:
> On Wed, Mar 18, 2015 at 7:22 AM, Connor Abbott wrote:
>> On Wed, Mar 18, 2015 at 1:56 AM, Matt Turner wrote:
>>> Thanks. Looked through stats and at some of the regressions.
>>>
>>> Some of the areas I noticed we were doing worse:
>>>
>>>
On Wed, Mar 18, 2015 at 7:22 AM, Connor Abbott wrote:
> On Wed, Mar 18, 2015 at 1:56 AM, Matt Turner wrote:
>> Thanks. Looked through stats and at some of the regressions.
>>
>> Some of the areas I noticed we were doing worse:
>>
>> We generate two CMPs for discard_if; only one without NIR. I thi
On Wed, Mar 18, 2015 at 7:22 AM, Connor Abbott wrote:
> On Wed, Mar 18, 2015 at 1:56 AM, Matt Turner wrote:
>> Thanks. Looked through stats and at some of the regressions.
>>
>> Some of the areas I noticed we were doing worse:
>>
>> We generate two CMPs for discard_if; only one without NIR. I thi
On Wed, Mar 18, 2015 at 1:56 AM, Matt Turner wrote:
> Thanks. Looked through stats and at some of the regressions.
>
> Some of the areas I noticed we were doing worse:
>
> We generate two CMPs for discard_if; only one without NIR. I think you
> had an idea about solving this.
>
> SEL peephole inte
Thanks. Looked through stats and at some of the regressions.
Some of the areas I noticed we were doing worse:
We generate two CMPs for discard_if; only one without NIR. I think you
had an idea about solving this.
SEL peephole interference -- we knew about this in general, but I
found an interest
I did this for SNB+ only because we did have some shader-db
regressions. However, with the patch 8.5 I just sent out, we're down
to about +0.2% on ILK-. Arguably, that's probably close enough for
the old platforms especially if it lets us move forward.
The majority of hurt shaders there are plan
GLSL IR vs. NIR FS instructions on snb:
total instructions in shared programs: 4976784 -> 4973072 (-0.07%)
instructions in affected programs: 3309521 -> 3305809 (-0.11%)
helped:7161
HURT: 9839
GAINED:
11 matches
Mail list logo