radv: Vectorize load_global_constant

Totals from 325 (0.38% of 84845) affected shaders:
Instrs: 2883767 -> 2880939 (-0.10%); split: -0.10%, +0.00%
CodeSize: 15388396 -> 15364976 (-0.15%); split: -0.16%, +0.00%
Latency: 20212583 -> 20197719 (-0.07%); split: -0.20%, +0.13%
InvThroughput: 5489738 -> 5489551 (-0.00%); split: -0.14%, +0.14%
VClause: 104225 -> 103944 (-0.27%); split: -0.30%, +0.03%
SClause: 71616 -> 71612 (-0.01%); split: -0.01%, +0.01%
Copies: 252143 -> 252284 (+0.06%); split: -0.10%, +0.15%
Branches: 117826 -> 117821 (-0.00%); split: -0.01%, +0.00%
PreVGPRs: 23502 -> 23506 (+0.02%)

Reviewed-by: Georg Lehmann <dadschoorse@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/27084>
This commit is contained in:
Konstantin Seurer
2024-01-14 13:46:46 +01:00
committed by Marge Bot
parent 220c912080
commit e83ef27ba6

View File

@@ -315,6 +315,7 @@ radv_mem_vectorize_callback(unsigned align_mul, unsigned align_offset, unsigned
switch (low->intrinsic) {
case nir_intrinsic_load_global:
case nir_intrinsic_load_global_constant:
case nir_intrinsic_store_global:
case nir_intrinsic_store_ssbo:
case nir_intrinsic_load_ssbo: