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

aco: fix waiting for scalar stores before "writing back" data on GFX8-GFX9

Browse Source 
Seems required also on GFX8-GFX9 to achieve correct behaviour. This
is an undocumented behaviour but it makes real sense to me.

pipeline-db on GFX9:
Totals from affected shaders:
SGPRS: 1018 -> 1018 (0.00 %)
VGPRS: 516 -> 516 (0.00 %)
Code Size: 40516 -> 40636 (0.30 %) bytes
Max Waves: 280 -> 280 (0.00 %)

This fixes some sort of sun flickering with Assassins Creed Origins.

Closes: https://gitlab.freedesktop.org/mesa/mesa/issues/2488
Cc: <mesa-stable@lists.freedesktop.org>
Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-by: Daniel Schürmann <daniel@schuermann.dev>
Tested-by: Marge Bot <https://gitlab.freedesktop.org/mesa/mesa/merge_requests/3750>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/merge_requests/3750>
This commit is contained in:
Samuel Pitoiset
2020-02-07 16:33:35 +01:00
committed by Marge Bot
parent 7283c33b98
commit 34fd894e42
1 changed files with 9 additions and 6 deletions
Download Patch File Download Diff File Expand all files Collapse all files
src/amd/compiler/aco_insert_waitcnt.cpp
+9 -6
View File
@@ -374,13 +374,16 @@ wait_imm kill(Instruction* instr, wait_ctx& ctx)
imm.combine(parse_wait_instr(ctx, instr));
if (ctx.chip_class >= GFX10) {
/* Seems to be required on GFX10 to achieve correct behaviour.
* It shouldn't cost anything anyways since we're about to do s_endpgm.
*/
if (ctx.lgkm_cnt && instr->opcode == aco_opcode::s_dcache_wb)
imm.lgkm = 0;
/* It's required to wait for scalar stores before "writing back" data.
* It shouldn't cost anything anyways since we're about to do s_endpgm.
*/
if (ctx.lgkm_cnt && instr->opcode == aco_opcode::s_dcache_wb) {
assert(ctx.chip_class >= GFX8);
imm.lgkm = 0;
}
if (ctx.chip_class >= GFX10) {
/* GFX10: A store followed by a load at the same address causes a problem because
* the load doesn't load the correct values unless we wait for the store first.
* This is NOT mitigated by an s_nop.