fdo-mirrors/mesa
1
0
Fork 0
mirror of https://gitlab.freedesktop.org/mesa/mesa.git synced 2026-05-06 11:38:05 +02:00
Code Issues Projects Releases Packages Wiki Activity Actions 1

aco: optimize s_and(s_cselect, exec)

Browse source 
fossil-db (navi21):
Totals from 62 (0.08% of 79825) affected shaders:
Instrs: 178887 -> 178745 (-0.08%)
CodeSize: 942980 -> 942328 (-0.07%)
Latency: 1274513 -> 1273653 (-0.07%)
InvThroughput: 213862 -> 213774 (-0.04%); split: -0.04%, +0.00%
SALU: 26446 -> 26301 (-0.55%)

Signed-off-by: Rhys Perry <pendingchaos02@gmail.com>
Reviewed-by: Georg Lehmann <dadschoorse@gmail.com>
Reviewed-by: Daniel Schürmann <daniel@schuermann.dev>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/36005>
This commit is contained in:
Rhys Perry 2025-07-07 17:57:20 +01:00 • committed by Marge Bot
parent 8fd5266b69
commit d81add3dda
1 changed files with 18 additions and 2 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

20
src/amd/compiler/aco_optimizer.cpp
View file

@ -1164,7 +1164,7 @@ can_eliminate_fcanonicalize(opt_ctx& ctx, aco_ptr<Instruction>& instr, Temp tmp,
}
bool
can_eliminate_and_exec(opt_ctx& ctx, Temp tmp, unsigned pass_flags)
can_eliminate_and_exec(opt_ctx& ctx, Temp tmp, unsigned pass_flags, bool allow_cselect = false)
{
Instruction* instr = ctx.info[tmp.id()].parent_instr;
/* Remove superfluous s_and when the VOPC instruction uses the same exec and thus
@ -1172,6 +1172,12 @@ can_eliminate_and_exec(opt_ctx& ctx, Temp tmp, unsigned pass_flags)
if (instr->isVOPC())
return instr->pass_flags == pass_flags;
if (allow_cselect && instr->pass_flags == pass_flags &&
(instr->opcode == aco_opcode::s_cselect_b32 || instr->opcode == aco_opcode::s_cselect_b64)) {
return (instr->operands[0].constantEquals(0) && instr->operands[1].constantEquals(-1)) ||
(instr->operands[1].constantEquals(0) && instr->operands[0].constantEquals(-1));
}
if (instr->operands.size() != 2 || instr->pass_flags != pass_flags)
return false;
if (!(instr->operands[0].isTemp() && instr->operands[1].isTemp()))
@ -4550,9 +4556,19 @@ select_instruction(opt_ctx& ctx, aco_ptr<Instruction>& instr)
if (instr->operands[0].isTemp() && fixed_to_exec(instr->operands[1]) &&
ctx.uses[instr->operands[0].tempId()] == 1 &&
ctx.uses[instr->definitions[1].tempId()] == 0 &&
can_eliminate_and_exec(ctx, instr->operands[0].getTemp(), instr->pass_flags)) {
can_eliminate_and_exec(ctx, instr->operands[0].getTemp(), instr->pass_flags, true)) {
ctx.uses[instr->operands[0].tempId()]--;
Instruction* op_instr = ctx.info[instr->operands[0].tempId()].parent_instr;
if (op_instr->opcode == aco_opcode::s_cselect_b32 ||
op_instr->opcode == aco_opcode::s_cselect_b64) {
for (unsigned i = 0; i < 2; i++) {
if (op_instr->operands[i].constantEquals(-1))
op_instr->operands[i] = instr->operands[1];
}
ctx.info[op_instr->definitions[0].tempId()].label &= label_uniform_bool;
}
op_instr->definitions[0].setTemp(instr->definitions[0].getTemp());
ctx.info[op_instr->definitions[0].tempId()].parent_instr = op_instr;
instr.reset();