[Mesa-dev] [PATCH v2 07/23] nir/lower_load_const_to_scalar: suppport doubles and multiple bit sizes

Samuel Iglesias Gonsálvez siglesias at igalia.com
Mon Apr 4 07:35:08 UTC 2016



On 01/04/16 21:06, Jason Ekstrand wrote:
> On Thu, Mar 31, 2016 at 2:59 AM, Samuel Iglesias Gonsálvez <
> siglesias at igalia.com> wrote:
> 
>> From: Iago Toral Quiroga <itoral at igalia.com>
>>
>> ---
>>  src/compiler/nir/nir_lower_load_const_to_scalar.c | 7 +++++--
>>  1 file changed, 5 insertions(+), 2 deletions(-)
>>
>> diff --git a/src/compiler/nir/nir_lower_load_const_to_scalar.c
>> b/src/compiler/nir/nir_lower_load_const_to_scalar.c
>> index d290c30..aae2f7c 100644
>> --- a/src/compiler/nir/nir_lower_load_const_to_scalar.c
>> +++ b/src/compiler/nir/nir_lower_load_const_to_scalar.c
>> @@ -49,8 +49,11 @@ lower_load_const_instr_scalar(nir_load_const_instr
>> *lower)
>>     nir_ssa_def *loads[4];
>>     for (unsigned i = 0; i < lower->def.num_components; i++) {
>>        nir_load_const_instr *load_comp =
>> -         nir_load_const_instr_create(b.shader, 1, 32);
>> -      load_comp->value.u32[0] = lower->value.u32[i];
>> +         nir_load_const_instr_create(b.shader, 1, lower->def.bit_size);
>> +      if (lower->def.bit_size == 64)
>> +         load_comp->value.f64[0] = lower->value.f64[i];
>> +      else
>> +         load_comp->value.u32[0] = lower->value.u32[i];
>>
> 
> Could we please throw an assert in here that it's only 32 or 64.  I want
> this to be easy to find when we start supporting 16-bit floats.
> 

OK, I'll add it.

Sam

> 
>>        nir_builder_instr_insert(&b, &load_comp->instr);
>>        loads[i] = &load_comp->def;
>>     }
>> --
>> 2.5.0
>>
>> _______________________________________________
>> mesa-dev mailing list
>> mesa-dev at lists.freedesktop.org
>> https://lists.freedesktop.org/mailman/listinfo/mesa-dev
>>
> 


More information about the mesa-dev mailing list