[Mesa-dev] [PATCH 1/4] nir: add option to lower slt/sge/seq/sne

[Rob Clark]

On Tue, Mar 31, 2015 at 8:54 PM, Kenneth Graunke <kenneth at whitecape.org> wrote:
> On Tuesday, March 31, 2015 06:57:30 PM Rob Clark wrote:
>> From: Rob Clark <robclark at freedesktop.org>
>>
>> In freedreno these get implemented as the matching f* instruction plus a
>> u2f to convert the result to float 1.0/0.0.  But less lines of code to
>> just let nir_opt_algebraic handle this for us, plus opens up some small
>> window for other opt passes to improve (ie. if some shader ended up with
>> both a flt and slt with same src args, for example).
>>
>> v2: use b2f rather than u2f
>>
>> Signed-off-by: Rob Clark <robclark at freedesktop.org>
>> ---
>>  src/glsl/nir/nir.h                | 3 +++
>>  src/glsl/nir/nir_opt_algebraic.py | 4 ++++
>>  2 files changed, 7 insertions(+)
>>
>> diff --git a/src/glsl/nir/nir.h b/src/glsl/nir/nir.h
>> index e980635..74927e5 100644
>> --- a/src/glsl/nir/nir.h
>> +++ b/src/glsl/nir/nir.h
>> @@ -1372,6 +1372,9 @@ typedef struct nir_shader_compiler_options {
>>     /** lowers fneg and ineg to fsub and isub. */
>>     bool lower_negate;
>>
>
>                                               should be b2f
>                                                         |||
>                                                         vvv
>> +   /* lower {slt,sge,seq,sne} to {flt,fge,feq,fne} + u2f: */
>> +   bool lower_scmp;
>> +

thanks.. fixed locally..

BR,
-R

>
> Patches 1-2 are:
> Reviewed-by: Kenneth Graunke <kenneth at whitecape.org>
>
>>     /**
>>      * Does the driver support real 32-bit integers?  (Otherwise, integers
>>      * are simulated by floats.)
>> diff --git a/src/glsl/nir/nir_opt_algebraic.py b/src/glsl/nir/nir_opt_algebraic.py
>> index 20ec4d3..506d4df 100644
>> --- a/src/glsl/nir/nir_opt_algebraic.py
>> +++ b/src/glsl/nir/nir_opt_algebraic.py
>> @@ -96,6 +96,10 @@ optimizations = [
>>     (('fsat', a), ('fmin', ('fmax', a, 0.0), 1.0), 'options->lower_fsat'),
>>     (('fsat', ('fsat', a)), ('fsat', a)),
>>     (('fmin', ('fmax', ('fmin', ('fmax', a, 0.0), 1.0), 0.0), 1.0), ('fmin', ('fmax', a, 0.0), 1.0)),
>> +   (('slt', a, b), ('b2f', ('flt', a, b)), 'options->lower_scmp'),
>> +   (('sge', a, b), ('b2f', ('fge', a, b)), 'options->lower_scmp'),
>> +   (('seq', a, b), ('b2f', ('feq', a, b)), 'options->lower_scmp'),
>> +   (('sne', a, b), ('b2f', ('fne', a, b)), 'options->lower_scmp'),
>>     # Emulating booleans
>>     (('fmul', ('b2f', a), ('b2f', b)), ('b2f', ('iand', a, b))),
>>     (('fsat', ('fadd', ('b2f', a), ('b2f', b))), ('b2f', ('ior', a, b))),
>>