fdo-mirrors/mesa
1
0
Fork 0
mirror of https://gitlab.freedesktop.org/mesa/mesa.git synced 2025-12-27 10:30:08 +01:00
Code Issues Projects Releases Packages Wiki Activity Actions 1

aco/optimizer: optimize add(mad_u32_u16(a, b, 0), c)

Browse source 
Foz-DB Navi48:
Totals from 104 (0.13% of 82419) affected shaders:
Instrs: 3554243 -> 3553555 (-0.02%); split: -0.02%, +0.00%
CodeSize: 18836004 -> 18830572 (-0.03%); split: -0.03%, +0.00%
Latency: 19288034 -> 19287208 (-0.00%); split: -0.01%, +0.00%
InvThroughput: 3527510 -> 3526925 (-0.02%); split: -0.02%, +0.00%
VClause: 89526 -> 89522 (-0.00%); split: -0.02%, +0.01%
SClause: 62484 -> 62492 (+0.01%); split: -0.00%, +0.01%
Copies: 266415 -> 266404 (-0.00%); split: -0.04%, +0.03%
Branches: 102123 -> 102125 (+0.00%)
VALU: 1987067 -> 1986531 (-0.03%); split: -0.03%, +0.00%
SALU: 471348 -> 471346 (-0.00%); split: -0.00%, +0.00%

Foz-DB Navi21:
Totals from 228 (0.28% of 82387) affected shaders:
Instrs: 3069693 -> 3068317 (-0.04%); split: -0.05%, +0.00%
CodeSize: 16582476 -> 16574920 (-0.05%); split: -0.05%, +0.00%
Latency: 20038755 -> 20030986 (-0.04%); split: -0.04%, +0.00%
InvThroughput: 4742546 -> 4738245 (-0.09%); split: -0.10%, +0.00%
VClause: 93157 -> 93135 (-0.02%); split: -0.03%, +0.01%
Copies: 265019 -> 264959 (-0.02%); split: -0.04%, +0.02%
VALU: 2025352 -> 2023897 (-0.07%); split: -0.07%, +0.00%
SALU: 447385 -> 447375 (-0.00%); split: -0.00%, +0.00%

Reviewed-by: Daniel Schürmann <daniel@schuermann.dev>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/38530>
This commit is contained in:
Georg Lehmann 2024-12-14 19:25:22 +01:00 • committed by Marge Bot
parent 0359c8a901
commit d9919c3e10
1 changed files with 1 additions and 0 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

1
src/amd/compiler/aco_optimizer.cpp
View file

@ -4849,6 +4849,7 @@ combine_instruction(opt_ctx& ctx, aco_ptr<Instruction>& instr)
} else if (info.opcode == aco_opcode::v_add_u32 && !info.clamp) {
assert(ctx.program->gfx_level >= GFX9);
add_opt(v_bcnt_u32_b32, v_bcnt_u32_b32, 0x3, "102", remove_const_cb<0>, true);
add_opt(v_mad_u32_u16, v_mad_u32_u16, 0x3, "1203", remove_const_cb<0>, true);
add_opt(v_mul_u32_u24, v_mad_u32_u24, 0x3, "120", nullptr, true);
add_opt(v_mul_i32_i24, v_mad_i32_i24, 0x3, "120", nullptr, true);
add_opt(v_xor_b32, v_xad_u32, 0x3, "120", nullptr, true);