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

nir,radeonsi: move ffma fusing to late optimizations for better codegen

Browse Source 
The freedreno trace changes were suggested by Rob Clark.

ALU performance is higher, because ffma is used more often, but so is
register usage, because trinary opcodes (such as ffma) usually need
at least 3 live registers.

54793 shaders in 33659 tests
Totals:
SGPRS: 2639746 -> 2642938 (0.12 %)
VGPRS: 1534120 -> 1536392 (0.15 %)
Spilled SGPRs: 3541 -> 3618 (2.17 %)
Spilled VGPRs: 33 -> 44 (33.33 %)
Scratch size: 292 -> 312 (6.85 %) dwords per thread
Code Size: 55639836 -> 55620116 (-0.04 %) bytes
Max Waves: 964785 -> 963977 (-0.08 %)

Totals from affected shaders:
SGPRS: 1105800 -> 1108992 (0.29 %)
VGPRS: 635292 -> 637564 (0.36 %)
Spilled SGPRs: 3193 -> 3270 (2.41 %)
Spilled VGPRs: 33 -> 44 (33.33 %)
Scratch size: 36 -> 56 (55.56 %) dwords per thread
Code Size: 31568708 -> 31548988 (-0.06 %) bytes
Max Waves: 319991 -> 319183 (-0.25 %)

Reviewed-by: Connor Abbott <cwabbott0@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/6596>
This commit is contained in:
Marek Olšák
2020-09-04 05:55:25 -04:00
committed by Marge Bot
parent a3512ddfdf
commit 57bf4c2028
4 changed files with 30 additions and 22 deletions
Download Patch File Download Diff File Expand all files Collapse all files

4
src/compiler/nir/nir_opt_algebraic.py
View File

@@ -194,7 +194,8 @@ optimizations.extend([
(('ffract', a), ('fsub', a, ('ffloor', a)), 'options->lower_ffract'),
(('fceil', a), ('fneg', ('ffloor', ('fneg', a))), 'options->lower_fceil'),
(('ffma', a, b, c), ('fadd', ('fmul', a, b), c), 'options->lower_ffma'),
(('~fadd', ('fmul', a, b), c), ('ffma', a, b, c), 'options->fuse_ffma'),
# Always lower inexact ffma, because it will be fused back by late optimizations (nir_opt_algebraic_late).
(('~ffma', a, b, c), ('fadd', ('fmul', a, b), c), 'options->fuse_ffma'),
(('~fmul', ('fadd', ('iand', ('ineg', ('b2i', 'a@bool')), ('fmul', b, c)), '#d'), '#e'),
('bcsel', a, ('fmul', ('fadd', ('fmul', b, c), d), e), ('fmul', d, e))),
@@ -2027,6 +2028,7 @@ late_optimizations = [
(('fneg', a), ('fsub', 0.0, a), 'options->lower_negate'),
(('ineg', a), ('isub', 0, a), 'options->lower_negate'),
(('iabs', a), ('imax', a, ('ineg', a)), 'options->lower_iabs'),
(('~fadd', ('fmul', a, b), c), ('ffma', a, b, c), 'options->fuse_ffma'),
# These are duplicated from the main optimizations table. The late
# patterns that rearrange expressions like x - .5 < 0 to x < .5 can create