[Mesa-dev] [PATCH 4/4] i965/vec4: Generate better code for ir_triop_csel.

[Ian Romanick]

On 10/15/2014 08:51 PM, Kenneth Graunke wrote:
> Previously, we generated an extra CMP instruction:
> 
>    cmp.ge.f0(8)    g6<1>D          g1<0,4,1>F      0F
>    cmp.nz.f0(8)    null            g6<4,4,1>D      0D
>    (+f0) sel(8)    g5<1>F          g1.4<0,4,1>F    g2<0,4,1>F
> 
> The first operand is always a boolean, and we want to predicate the SEL
> on that.  Rather than producing a boolean value and comparing it against
> zero, we can just produce a condition code in the flag register.
> 
> Now we generate:
> 
>    cmp.ge.f0(8)    null            g1<0,4,1>F      0F
>    (+f0) sel(8)    g5<1>F          g1.4<0,4,1>F    g2<0,4,1>F
> 
> No difference in shader-db.

Do any piglit tests generate ir_triop_csel for VS or GS?  The change is
nearly identical to the FS change, but it would be nice to know any of
it was exercised.

> Signed-off-by: Kenneth Graunke <kenneth at whitecape.org>
> ---
>  src/mesa/drivers/dri/i965/brw_vec4_visitor.cpp | 14 ++++++++++++++
>  1 file changed, 14 insertions(+)
> 
> diff --git a/src/mesa/drivers/dri/i965/brw_vec4_visitor.cpp b/src/mesa/drivers/dri/i965/brw_vec4_visitor.cpp
> index 93aa533..4c517af 100644
> --- a/src/mesa/drivers/dri/i965/brw_vec4_visitor.cpp
> +++ b/src/mesa/drivers/dri/i965/brw_vec4_visitor.cpp
> @@ -1277,6 +1277,20 @@ vec4_visitor::visit(ir_expression *ir)
>     dst_reg result_dst(this, ir->type);
>     src_reg result_src(result_dst);
>  
> +   if (ir->operation == ir_triop_csel) {
> +      ir->operands[1]->accept(this);
> +      op[1] = this->result;
> +      ir->operands[2]->accept(this);
> +      op[2] = this->result;
> +
> +      enum brw_predicate predicate;
> +      emit_bool_to_cond_code(ir->operands[0], &predicate);
> +      inst = emit(BRW_OPCODE_SEL, result_dst, op[1], op[2]);
> +      inst->predicate = predicate;
> +      this->result = result_src;
> +      return;
> +   }
> +
>     for (operand = 0; operand < ir->get_num_operands(); operand++) {
>        this->result.file = BAD_FILE;
>        ir->operands[operand]->accept(this);
>