OpenHarmony/third_party_mesa3d
2
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

aco: don't transform v_interp_p2_f32 with constant into fma

Browse Source 
Since v_interp_p2_f32 with constant operands only happens on GFX11.5, this
should actually be fine in all cases where this is currently possible
(GFX11.5+ allows DPP with scalar src1). However, it does fail validation
because we haven't updated that yet.

Signed-off-by: Rhys Perry <pendingchaos02@gmail.com>
Reviewed-by: Georg Lehmann <dadschoorse@gmail.com>
Fixes: bee487df48 ("aco/gfx11.5+: use vinterp for fddx/fddy")
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/30477>
This commit is contained in:
Rhys Perry
2024-08-01 15:02:45 +01:00
committed by Marge Bot
parent 911fdce0b6
commit 8f3d0fbad7
1 changed files with 18 additions and 4 deletions
Download Patch File Download Diff File Expand all files Collapse all files

22
src/amd/compiler/aco_optimizer.cpp
View File

@@ -2990,6 +2990,21 @@ apply_sgprs(opt_ctx& ctx, aco_ptr<Instruction>& instr)
}
}
bool
interp_can_become_fma(opt_ctx& ctx, aco_ptr<Instruction>& instr)
{
if (instr->opcode != aco_opcode::v_interp_p2_f32_inreg)
return false;
instr->opcode = aco_opcode::v_fma_f32;
instr->format = Format::VOP3;
bool dpp_allowed = can_use_DPP(ctx.program->gfx_level, instr, false);
instr->opcode = aco_opcode::v_interp_p2_f32_inreg;
instr->format = Format::VINTERP_INREG;
return dpp_allowed;
}
void
interp_p2_f32_inreg_to_fma_dpp(aco_ptr<Instruction>& instr)
{
@@ -3020,9 +3035,8 @@ apply_omod_clamp(opt_ctx& ctx, aco_ptr<Instruction>& instr)
return false;
/* SDWA omod is GFX9+. */
bool can_use_omod =
(can_vop3 || ctx.program->gfx_level >= GFX9) && !instr->isVOP3P() &&
(!instr->isVINTERP_INREG() || instr->opcode == aco_opcode::v_interp_p2_f32_inreg);
bool can_use_omod = (can_vop3 || ctx.program->gfx_level >= GFX9) && !instr->isVOP3P() &&
(!instr->isVINTERP_INREG() || interp_can_become_fma(ctx, instr));
ssa_info& def_info = ctx.info[instr->definitions[0].tempId()];
@@ -3581,7 +3595,7 @@ combine_output_conversion(opt_ctx& ctx, aco_ptr<Instruction>& instr)
if (conv->usesModifiers())
return false;
if (instr->opcode == aco_opcode::v_interp_p2_f32_inreg)
if (interp_can_become_fma(ctx, instr))
interp_p2_f32_inreg_to_fma_dpp(instr);
if (!can_use_mad_mix(ctx, instr))