[Mesa-dev] [PATCH 031/133] nir: Add fine and coarse derivative opcodes

[Connor Abbott]

On Wed, Dec 17, 2014 at 6:52 AM, Jason Ekstrand <jason at jlekstrand.net> wrote:
>
>
> On Tue, Dec 16, 2014 at 10:31 PM, Connor Abbott <cwabbott0 at gmail.com> wrote:
>>
>> As a future cleanup, would it be worth getting rid of fddx and fddy
>> entirely and just generating the correct variant based on the
>> glHint/drirc configuration?
>
>
> Nope.  https://www.opengl.org/registry/specs/ARB/derivative_control.txt

I just meant doing the

if (fs_key->high_quality_derivatives) { emit ddx_fine } else { emit
ddx_coarse } part in NIR so other drivers don't have to handle it, I
still don't understand why that's not possible.

>
>>
>>
>> On Tue, Dec 16, 2014 at 1:04 AM, Jason Ekstrand <jason at jlekstrand.net>
>> wrote:
>> > ---
>> >  src/glsl/nir/nir_opcodes.h | 4 ++++
>> >  1 file changed, 4 insertions(+)
>> >
>> > diff --git a/src/glsl/nir/nir_opcodes.h b/src/glsl/nir/nir_opcodes.h
>> > index bee4bd1..9a7a2ab 100644
>> > --- a/src/glsl/nir/nir_opcodes.h
>> > +++ b/src/glsl/nir/nir_opcodes.h
>> > @@ -111,6 +111,10 @@ UNOP(fcos_reduced, nir_type_float)
>> >  /*@{*/
>> >  UNOP(fddx, nir_type_float)
>> >  UNOP(fddy, nir_type_float)
>> > +UNOP(fddx_fine, nir_type_float)
>> > +UNOP(fddy_fine, nir_type_float)
>> > +UNOP(fddx_coarse, nir_type_float)
>> > +UNOP(fddy_coarse, nir_type_float)
>> >  /*@}*/
>> >
>> >  /**
>> > --
>> > 2.2.0
>> >
>> > _______________________________________________
>> > mesa-dev mailing list
>> > mesa-dev at lists.freedesktop.org
>> > http://lists.freedesktop.org/mailman/listinfo/mesa-dev