summaryrefslogtreecommitdiffstats
diff options
context:
space:
mode:
authorTimothy Arceri <[email protected]>2019-03-14 11:02:11 +1100
committerTimothy Arceri <[email protected]>2019-03-18 09:29:40 +1100
commitde8ec6e1179d0cef559e19cac8918d47cebf1957 (patch)
tree5744e6bb3f55b899ec5e56889c99861d4b678e70
parent29179f58c6ba8099859ea25900214dbbd3814a92 (diff)
radeonsi/nir: call some more var optimisation passes
shader-db results (VEGA64): Totals from affected shaders: SGPRS: 5328912 -> 5329680 (0.01 %) VGPRS: 2969308 -> 2969164 (-0.00 %) Spilled SGPRs: 37921 -> 37917 (-0.01 %) Spilled VGPRs: 32882 -> 29024 (-11.73 %) Private memory VGPRs: 0 -> 0 (0.00 %) Scratch size: 1400 -> 1200 (-14.29 %) dwords per thread Code Size: 121126000 -> 121282784 (0.13 %) bytes LDS: 1501 -> 1501 (0.00 %) blocks Max Waves: 933188 -> 933229 (0.00 %) Wait states: 0 -> 0 (0.00 %) Reviewed-by: Bas Nieuwenhuizen <[email protected]>
-rw-r--r--src/gallium/drivers/radeonsi/si_shader_nir.c11
1 files changed, 8 insertions, 3 deletions
diff --git a/src/gallium/drivers/radeonsi/si_shader_nir.c b/src/gallium/drivers/radeonsi/si_shader_nir.c
index 25429fdeab7..5ac18e2ebc8 100644
--- a/src/gallium/drivers/radeonsi/si_shader_nir.c
+++ b/src/gallium/drivers/radeonsi/si_shader_nir.c
@@ -841,9 +841,6 @@ si_lower_nir(struct si_shader_selector* sel)
* - ensure constant offsets for texture instructions are folded
* and copy-propagated
*/
- NIR_PASS_V(sel->nir, nir_lower_vars_to_ssa);
- NIR_PASS_V(sel->nir, nir_lower_alu_to_scalar);
- NIR_PASS_V(sel->nir, nir_lower_phis_to_scalar);
static const struct nir_lower_tex_options lower_tex_options = {
.lower_txp = ~0u,
@@ -866,6 +863,14 @@ si_lower_nir(struct si_shader_selector* sel)
do {
progress = false;
+ NIR_PASS_V(sel->nir, nir_lower_vars_to_ssa);
+
+ NIR_PASS(progress, sel->nir, nir_opt_copy_prop_vars);
+ NIR_PASS(progress, sel->nir, nir_opt_dead_write_vars);
+
+ NIR_PASS_V(sel->nir, nir_lower_alu_to_scalar);
+ NIR_PASS_V(sel->nir, nir_lower_phis_to_scalar);
+
/* (Constant) copy propagation is needed for txf with offsets. */
NIR_PASS(progress, sel->nir, nir_copy_prop);
NIR_PASS(progress, sel->nir, nir_opt_remove_phis);