2018-04-05 16:35 GMT+02:00 Rob Clark :
> I'd vote for PIPE_CAP until there is any place inside nir passes were
> we'd want to do something differently. Since this is just a decision
> in mesa/st I think a cap makes sense..
>
Makes sense.. will go that route - thanks for your comments!
--
greets
I'd vote for PIPE_CAP until there is any place inside nir passes were
we'd want to do something differently. Since this is just a decision
in mesa/st I think a cap makes sense..
On Thu, Apr 5, 2018 at 10:25 AM, Jason Ekstrand wrote:
> I'm not sure if this should be a NIR compiler option or a PIP
I'm not sure if this should be a NIR compiler option or a PIPE_CAP. On
Intel hardware, we have both and we handle it ourselves.
On April 4, 2018 23:04:36 Christian Gmeiner
wrote:
As not every (upcoming) backend compiler is happy with
nir_lower_xxx_to_scalar lowerings do them only if the ba
As not every (upcoming) backend compiler is happy with
nir_lower_xxx_to_scalar lowerings do them only if the backend
is scalar (and not vec4) based.
Signed-off-by: Christian Gmeiner
---
src/amd/vulkan/radv_shader.c| 1 +
src/broadcom/compiler/nir_to_vir.c | 1 +
src/comp