NeroReflex/mesa
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity

aco: give spiller more room to assign spilled SGPRs to VGPRs

Browse source 
On chordal graphs, a greedy coloring can be done in a way that never uses
more colors than are required for the largest clique. However, since we
have vector values and force phi resources into the same spill slots, the
interference graphs are not chordal, and thus, this assumption doesn't hold.

Use twice as many spill slots as upper bound.

Totals from 10 (0.01% of 79242) affected shaders: (GFX11)
MaxWaves: 52 -> 54 (+3.85%)
Instrs: 271386 -> 271779 (+0.14%)
CodeSize: 1362544 -> 1365432 (+0.21%)
VGPRs: 2536 -> 2532 (-0.16%)
SpillVGPRs: 778 -> 818 (+5.14%)
Scratch: 73472 -> 76800 (+4.53%)
Latency: 3331718 -> 3328798 (-0.09%); split: -0.14%, +0.05%
InvThroughput: 1665860 -> 1643350 (-1.35%); split: -1.40%, +0.05%
VClause: 3292 -> 3329 (+1.12%); split: -0.06%, +1.18%
Copies: 46082 -> 46257 (+0.38%)

Cc: mesa-stable
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/27011>
(cherry picked from commit e3098bb23284b598ec6d5030ceaa1c3d5bd9d428)
This commit is contained in:
Daniel Schürmann 2024-01-11 10:38:40 +01:00 • committed by Eric Engestrom
parent 0392e4bf5c
commit e1d20b69c4
2 changed files with 3 additions and 3 deletions
Download patch file Download diff file Expand all files Collapse all files

2
.pick_status.json
View file

@ -714,7 +714,7 @@
"description": "aco: give spiller more room to assign spilled SGPRs to VGPRs",
"nominated": true,
"nomination_type": 0,
"resolution": 0,
"resolution": 1,
"main_sha": null,
"because_sha": null,
"notes": null

4
src/amd/compiler/aco_spill.cpp
View file

@ -1938,7 +1938,7 @@ spill(Program* program, live& live_vars)
/* calculate extra VGPRs required for spilling SGPRs */
if (demand.sgpr > sgpr_limit) {
unsigned sgpr_spills = demand.sgpr - sgpr_limit;
extra_vgprs = DIV_ROUND_UP(sgpr_spills, program->wave_size) + 1;
extra_vgprs = DIV_ROUND_UP(sgpr_spills * 2, program->wave_size) + 1;
}
/* add extra SGPRs required for spilling VGPRs */
if (demand.vgpr + extra_vgprs > vgpr_limit) {
@ -1949,7 +1949,7 @@ spill(Program* program, live& live_vars)
if (demand.sgpr + extra_sgprs > sgpr_limit) {
/* re-calculate in case something has changed */
unsigned sgpr_spills = demand.sgpr + extra_sgprs - sgpr_limit;
extra_vgprs = DIV_ROUND_UP(sgpr_spills, program->wave_size) + 1;
extra_vgprs = DIV_ROUND_UP(sgpr_spills * 2, program->wave_size) + 1;
}
}
/* the spiller has to target the following register demand */