On Wed, Mar 21, 2018 at 8:18 AM, Jason Ekstrand <ja...@jlekstrand.net>
wrote:
> Let's handle 16-bit while we're at it.
>
In particular, why not just create a nir_imm_floatN_t helper similar to the
nir_imm_intN_t helper I added not too long ago. It would take a double and
automatically convert to float or float16 (via _mesa_half_to_float) as
needed. Then a bunch of these things would just involve calling that
helper and you get all three for free.
--Jason
> On Thu, Mar 8, 2018 at 8:07 AM, Neil Roberts <nrobe...@igalia.com> wrote:
>
>> The only change neccessary is to change the type of the constant used
>> to compare against.
>>
>> This has been tested against the arb_gpu_shader_fp64/execution/
>> fs-isinf-dvec tests using the ARB_gl_spirv branch.
>> ---
>> src/compiler/spirv/vtn_alu.c | 11 ++++++++---
>> 1 file changed, 8 insertions(+), 3 deletions(-)
>>
>> diff --git a/src/compiler/spirv/vtn_alu.c b/src/compiler/spirv/vtn_alu.c
>> index d0c9e316935..2b03a2e4d09 100644
>> --- a/src/compiler/spirv/vtn_alu.c
>> +++ b/src/compiler/spirv/vtn_alu.c
>> @@ -529,10 +529,15 @@ vtn_handle_alu(struct vtn_builder *b, SpvOp opcode,
>> val->ssa->def = nir_fne(&b->nb, src[0], src[0]);
>> break;
>>
>> - case SpvOpIsInf:
>> - val->ssa->def = nir_ieq(&b->nb, nir_fabs(&b->nb, src[0]),
>> - nir_imm_float(&b->nb, INFINITY));
>> + case SpvOpIsInf: {
>> + nir_ssa_def *inf;
>> + if (src[0]->bit_size == 64)
>> + inf = nir_imm_double(&b->nb, INFINITY);
>> + else
>> + inf = nir_imm_float(&b->nb, INFINITY);
>> + val->ssa->def = nir_ieq(&b->nb, nir_fabs(&b->nb, src[0]), inf);
>> break;
>> + }
>>
>> case SpvOpFUnordEqual:
>> case SpvOpFUnordNotEqual:
>> --
>> 2.14.3
>>
>> _______________________________________________
>> mesa-dev mailing list
>> mesa-dev@lists.freedesktop.org
>> https://lists.freedesktop.org/mailman/listinfo/mesa-dev
>>
>
>
_______________________________________________
mesa-dev mailing list
mesa-dev@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/mesa-dev
