[Mesa-dev] [PATCH] radv: call constant folding before opt algebraic

Timothy Arceri tarceri at itsqueeze.com
Tue May 7 05:14:49 UTC 2019


ping!

On 2/5/19 1:38 pm, Timothy Arceri wrote:
> The pattern of calling opt algebraic first seems to have originated
> in i965. The order in OpenGL drivers generally doesn't matter
> because the GLSL IR optimisations do constant folding before
> opt algebraic.
> 
> However in Vulkan drivers calling opt algebraic first can result
> in missed constant folding opportunities.
> 
> vkpipeline-db results (VEGA64):
> 
> Totals from affected shaders:
> SGPRS: 3160 -> 3176 (0.51 %)
> VGPRS: 3588 -> 3580 (-0.22 %)
> Spilled SGPRs: 52 -> 44 (-15.38 %)
> Spilled VGPRs: 0 -> 0 (0.00 %)
> Private memory VGPRs: 0 -> 0 (0.00 %)
> Scratch size: 12 -> 12 (0.00 %) dwords per thread
> Code Size: 261812 -> 261036 (-0.30 %) bytes
> LDS: 7 -> 7 (0.00 %) blocks
> Max Waves: 346 -> 348 (0.58 %)
> Wait states: 0 -> 0 (0.00 %)
> ---
>   src/amd/vulkan/radv_shader.c | 2 +-
>   1 file changed, 1 insertion(+), 1 deletion(-)
> 
> diff --git a/src/amd/vulkan/radv_shader.c b/src/amd/vulkan/radv_shader.c
> index cd5a9f2afb4..ad7b2439735 100644
> --- a/src/amd/vulkan/radv_shader.c
> +++ b/src/amd/vulkan/radv_shader.c
> @@ -162,8 +162,8 @@ radv_optimize_nir(struct nir_shader *shader, bool optimize_conservatively,
>                   NIR_PASS(progress, shader, nir_opt_dead_cf);
>                   NIR_PASS(progress, shader, nir_opt_cse);
>                   NIR_PASS(progress, shader, nir_opt_peephole_select, 8, true, true);
> -                NIR_PASS(progress, shader, nir_opt_algebraic);
>                   NIR_PASS(progress, shader, nir_opt_constant_folding);
> +                NIR_PASS(progress, shader, nir_opt_algebraic);
>                   NIR_PASS(progress, shader, nir_opt_undef);
>                   NIR_PASS(progress, shader, nir_opt_conditional_discard);
>                   if (shader->options->max_unroll_iterations) {
> 


More information about the mesa-dev mailing list