AlexIndustrial/mesa
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

aco: disallow various v_add_u32 opts if modifiers are used

Browse Source 
Check for clamp, SDWA or DPP. The optimization isn't possible with SDWA
and DPP, so it would have been skipped anyway. Doing any of these with a
clamp modifier present would be incorrect.

No fossil-db changes.

Signed-off-by: Rhys Perry <pendingchaos02@gmail.com>
Reviewed-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Cc: mesa-stable
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/7045>
This commit is contained in:
Rhys Perry
2020-10-07 11:45:30 +01:00
committed by Marge Bot
parent 91ffeed88a
commit 966732e8ca
2 changed files with 29 additions and 1 deletions
Download Patch File Download Diff File Expand all files Collapse all files
src/amd/compiler/aco_optimizer.cpp
+1 -1
View File
@@ -2907,7 +2907,7 @@ void combine_instruction(opt_ctx &ctx, Block& block, aco_ptr<Instruction>& instr
} else if (instr->opcode == aco_opcode::v_add_u32) {
if (combine_add_sub_b2i(ctx, instr, aco_opcode::v_addc_co_u32, 1 | 2)) ;
else if (combine_add_bcnt(ctx, instr)) ;
else if (ctx.program->chip_class >= GFX9) {
else if (ctx.program->chip_class >= GFX9 && !instr->usesModifiers()) {
if (combine_three_valu_op(ctx, instr, aco_opcode::s_xor_b32, aco_opcode::v_xad_u32, "120", 1 | 2)) ;
else if (combine_three_valu_op(ctx, instr, aco_opcode::v_xor_b32, aco_opcode::v_xad_u32, "120", 1 | 2)) ;
else if (combine_three_valu_op(ctx, instr, aco_opcode::s_add_i32, aco_opcode::v_add3_u32, "012", 1 | 2)) ;