On Mon, Nov 9, 2015 at 6:26 PM, Jason Ekstrand wrote:
> On Mon, Nov 9, 2015 at 3:52 PM, Ian Romanick wrote:
>> On 11/04/2015 03:55 PM, Jason Ekstrand wrote:
>>> Here's some shader-db numbers:
>>>
>>> total instructions in shared programs: 6236146 -> 6255385 (0.31%)
>>> instructions in affected p
On Mon, Nov 9, 2015 at 3:52 PM, Ian Romanick wrote:
> On 11/04/2015 03:55 PM, Jason Ekstrand wrote:
>> Here's some shader-db numbers:
>>
>> total instructions in shared programs: 6236146 -> 6255385 (0.31%)
>> instructions in affected programs: 203629 -> 222868 (9.45%)
>> helped:
On 11/04/2015 03:55 PM, Jason Ekstrand wrote:
> Here's some shader-db numbers:
>
> total instructions in shared programs: 6236146 -> 6255385 (0.31%)
> instructions in affected programs: 203629 -> 222868 (9.45%)
> helped:118
> HURT:
Here's some shader-db numbers:
total instructions in shared programs: 6236146 -> 6255385 (0.31%)
instructions in affected programs: 203629 -> 222868 (9.45%)
helped:118
HURT: 1989
GAINED:18
LOST:
This little series fixes a problem where we lower
v[i] = s
to
v = vector_insert(v, s, i)
turning a component write into a full vector write. This used to be
ok, since none of this was visible to other threads. Now with SSBOs
(and compute shader SLM and tesselation outputs) this rewrite is
