fdo-mirrors/mesa
1
0
Fork 0
mirror of https://gitlab.freedesktop.org/mesa/mesa.git synced 2026-05-18 11:38:06 +02:00
Code Issues Projects Releases Packages Wiki Activity Actions 2
158156 commits 66 branches 935 tags 943 MiB
Commit graph

27 commits

Author SHA1 Message Date
Rhys Perry
cfb306b4e4 aco: test branch opcode if removing it in try_optimize_branching_sequence 
We shouldn't remove a p_cbranch_nz branch in this situation.

Signed-off-by: Rhys Perry <pendingchaos02@gmail.com>
Reviewed-by: Georg Lehmann <dadschoorse@gmail.com>
Reviewed-by: Timur Kristóf <timur.kristof@gmail.com>
Fixes: b731be2e96 ("aco: Remove branch instruction when exec is constant non-zero.")
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18077>
 2022-08-16 17:31:33 +00:00
Rhys Perry
21cb002b60 aco: fix re-write of uses of exec_val's lo/hi half 
The isConstant() check isn't useful. If it's a constant, then the
physReg() check will fail.

Signed-off-by: Rhys Perry <pendingchaos02@gmail.com>
Reviewed-by: Georg Lehmann <dadschoorse@gmail.com>
Reviewed-by: Timur Kristóf <timur.kristof@gmail.com>
Fixes: baab6f18c9 ("aco: Optimize branching sequence during SSA elimination.")
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18077>
 2022-08-16 17:31:33 +00:00
Rhys Perry
e493aab3c3 aco: fix consecutive exec writes when finding exec_copy instruction 
This can happen with transitions to exact:
 s2: %0:exec = p_parallelcopy %622:s[0-1]
 s2: %625:s[0-1],  s1: %624:scc,  s2: %0:exec = s_and_saveexec_b64 %141:vcc, %0:exec
 s2: %626:s[12-13] = p_cbranch_z %0:exec BB2, BB1

Signed-off-by: Rhys Perry <pendingchaos02@gmail.com>
Reviewed-by: Georg Lehmann <dadschoorse@gmail.com>
Reviewed-by: Timur Kristóf <timur.kristof@gmail.com>
Fixes: 410eff4d2f ("aco: Fix optimizing branching sequence with s_and_saveexec.")
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18077>
 2022-08-16 17:31:33 +00:00
Georg Lehmann
7b9d3ebe42 aco: Use v_cmpx pre GFX10. 
Foz-DB Vega10:
Totals from 29508 (21.85% of 135041) affected shaders:
CodeSize: 184345656 -> 184345820 (+0.00%)
Instrs: 35906154 -> 35906195 (+0.00%)
Latency: 581696114 -> 581530021 (-0.03%); split: -0.03%, +0.00%
InvThroughput: 245625572 -> 245561351 (-0.03%); split: -0.03%, +0.00%
Copies: 3134925 -> 3278672 (+4.59%)

Signed-off-by: Georg Lehmann <dadschoorse@gmail.com>
Reviewed-by: Timur Kristóf <timur.kristof@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18049>
 2022-08-15 13:25:38 +00:00
Georg Lehmann
410eff4d2f aco: Fix optimizing branching sequence with s_and_saveexec. 
This optimization was broken for two reasons:
- s_and_saveexec has two operands, the copy value and exec
- s_and_saveexec has an exec read, so exec_write_used will always be true
  before we find branch_exec_val_idx

Foz-DB Navi21:
Totals from 31453 (23.31% of 134913) affected shaders:
CodeSize: 204831260 -> 204831156 (-0.00%)
Instrs: 38157117 -> 38157091 (-0.00%)
Latency: 533708882 -> 531211721 (-0.47%); split: -0.47%, +0.00%
InvThroughput: 107088408 -> 106719188 (-0.34%); split: -0.35%, +0.00%
Copies: 2326179 -> 2502490 (+7.58%)

Signed-off-by: Georg Lehmann <dadschoorse@gmail.com>
Reviewed-by: Timur Kristóf <timur.kristof@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18049>
 2022-08-15 13:25:38 +00:00
Georg Lehmann
9e5f311efe aco: Check that we don't override exec_val operands during branching sequence optimization. 
Fixes: baab6f18c9 ("aco: Optimize branching sequence during SSA elimination.")
Signed-off-by: Georg Lehmann <dadschoorse@gmail.com>
Reviewed-by: Timur Kristóf <timur.kristof@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18049>
 2022-08-15 13:25:38 +00:00
Timur Kristóf
baf314e2c0 aco: Check for instructions that inhibit the branching sequence optimization. 
Fixes: baab6f18c9 ("aco: Optimize branching sequence during SSA elimination.")
Signed-off-by: Timur Kristóf <timur.kristof@gmail.com>
Reviewed-by: Georg Lehmann <dadschoorse@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18049>
 2022-08-15 13:25:38 +00:00
Timur Kristóf
d88b2e4ab5 aco: Fix invalidated reference in branching sequence optimization. 
Inserting in the instructions vector may invalidate the exec_val reference,
so do that last.

Fixes: baab6f18c9 ("aco: Optimize branching sequence during SSA elimination.")
Signed-off-by: Timur Kristóf <timur.kristof@gmail.com>
Reviewed-by: Georg Lehmann <dadschoorse@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18049>
 2022-08-15 13:25:38 +00:00
Timur Kristóf
b731be2e96 aco: Remove branch instruction when exec is constant non-zero. 
This mainly helps the "if (elect())" that is used in
NGG culling shaders, effectively removing a useless branch
from every culling shader.

Totals from 58346 (45.35% of 128653) affected shaders:
CodeSize: 153238668 -> 153005284 (-0.15%)
Instrs: 29066198 -> 29007852 (-0.20%)
Latency: 133626003 -> 133598182 (-0.02%); split: -0.02%, +0.00%
InvThroughput: 20208765 -> 20208689 (-0.00%)
Branches: 1190209 -> 1131863 (-4.90%)

Signed-off-by: Timur Kristóf <timur.kristof@gmail.com>
Reviewed-by: Daniel Schürmann <daniel@schuermann.dev>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/13383>
 2022-08-05 23:58:57 +00:00
Timur Kristóf
baab6f18c9 aco: Optimize branching sequence during SSA elimination. 
Totals from 63245 (49.16% of 128653) affected shaders:
CodeSize: 166703680 -> 166436164 (-0.16%)
Instrs: 31618383 -> 31551504 (-0.21%)
Latency: 149404811 -> 149337729 (-0.04%); split: -0.05%, +0.00%
InvThroughput: 23996388 -> 23994734 (-0.01%); split: -0.01%, +0.00%
Copies: 2794107 -> 2727228 (-2.39%)

Signed-off-by: Timur Kristóf <timur.kristof@gmail.com>
Reviewed-by: Georg Lehmann <dadschoorse@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/13383>
 2022-08-05 23:58:57 +00:00
Daniel Schürmann
007cb02db9 aco: use branch definition as scratch register for SSA lowering 
Reviewed-by: Rhys Perry <pendingchaos02@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/15505>
 2022-03-28 07:36:46 +00:00
Daniel Schürmann
1e2639026f aco: Format. 
Manually adjusted some comments for more intuitive line breaks.

Reviewed-by: Tony Wasserka <tony.wasserka@gmx.de>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/11258>
 2021-07-12 21:27:31 +00:00
Daniel Schürmann
59fdaa1985 aco: reorder and cleanup #includes 
Reviewed-by: Tony Wasserka <tony.wasserka@gmx.de>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/11271>
 2021-07-12 12:09:31 +00:00
Timur Kristóf
c783293e47 aco: Don't eliminate exec write when it's used by a copy later. 
Fixes: bc13049747
Signed-off-by: Timur Kristóf <timur.kristof@gmail.com>
Reviewed-by: Daniel Schürmann <daniel@schuermann.dev>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/10920>
 2021-05-25 13:50:43 +00:00
Timur Kristóf
bc13049747 aco: Eliminate useless exec writes in jump threading. 
Eliminate exec writes which are unused by subsequent instructions.

Fossil DB results on Sienna Cichlid:

Totals from 80960 (54.03% of 149839) affected shaders:
CodeSize: 162953748 -> 161749372 (-0.74%)
Instrs: 31462273 -> 31161179 (-0.96%)
Copies: 2171239 -> 1942293 (-10.54%)
Branches: 807771 -> 807747 (-0.00%)

Signed-off-by: Timur Kristóf <timur.kristof@gmail.com>
Reviewed-by: Tony Wasserka <tony.wasserka@gmx.de>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/10691>
 2021-05-18 11:48:22 +00:00
Timur Kristóf
e230dcc30b aco: Refactor SSA elimination phi info to use vector instead of map. 
No Fossil DB changes.

Signed-off-by: Timur Kristóf <timur.kristof@gmail.com>
Reviewed-by: Tony Wasserka <tony.wasserka@gmx.de>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/10691>
 2021-05-18 11:48:22 +00:00
Daniel Schürmann
5d7b3bf1a7 aco: handle non-temp phi definitions and operands 
This will be necessary as we make exec non-temp.

No fossil-db changes.

Reviewed-by: Rhys Perry <pendingchaos02@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/8870>
 2021-02-12 22:41:31 +00:00
Rhys Perry
e115b01948 aco: return references in instruction cast methods 
Signed-off-by: Rhys Perry <pendingchaos02@gmail.com>
Reviewed-by: Daniel Schürmann <daniel@schuermann.dev>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/8595>
 2021-01-22 14:12:33 +00:00
Rhys Perry
1d245cd18b aco: use format-check methods 
Signed-off-by: Rhys Perry <pendingchaos02@gmail.com>
Reviewed-by: Daniel Schürmann <daniel@schuermann.dev>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/8595>
 2021-01-22 14:12:32 +00:00
Rhys Perry
70dbcfa1c9 aco: use instruction cast methods 
Signed-off-by: Rhys Perry <pendingchaos02@gmail.com>
Reviewed-by: Daniel Schürmann <daniel@schuermann.dev>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/8595>
 2021-01-22 14:12:32 +00:00
Tony Wasserka
6ba83d820c aco: Annotate switch fallthroughs 
Reviewed-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-by: Rhys Perry <pendingchaos02@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/7847>
 2020-12-15 18:22:46 +00:00
Tony Wasserka
2bb8874320 aco: Fix -Wshadow warnings 
Reviewed-by: Rhys Perry <pendingchaos02@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/7430>
 2020-11-20 09:29:19 +00:00
Rhys Perry
29bfe18abd aco: fix fall-through test in try_remove_simple_block() with back-edges 
3bca0af2 enhanced empty block determination which exposed this bug and
created an infinite loop in a Guild Wars 2 shader.

Signed-off-by: Rhys Perry <pendingchaos02@gmail.com>
Fixes: 3bca0af25d
     ('aco: ignore parallelcopies to the same register on jump threading')

Closes: https://gitlab.freedesktop.org/mesa/mesa/issues/2364
Tested-by: Marge Bot <https://gitlab.freedesktop.org/mesa/mesa/merge_requests/3452>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/merge_requests/3452>
 2020-01-20 11:51:45 +00:00
Daniel Schürmann
3bca0af25d aco: ignore parallelcopies to the same register on jump threading 
The more conservative lowering to CSSA inserts unnecessary parallelcopies
which might get coalesced and can be ignored on jump threading.

v2: outline is_empty_block() check.

Reviewed-by: Rhys Perry <pendingchaos02@gmail.com>
Tested-by: Marge Bot <https://gitlab.freedesktop.org/mesa/mesa/merge_requests/3385>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/merge_requests/3385>
 2020-01-16 16:01:59 +01:00
Rhys Perry
01ccd7839c aco: improve jump threading with wave32 
Totals from affected shaders:
SGPRS: 748746 -> 748746 (0.00 %)
VGPRS: 636984 -> 636984 (0.00 %)
Spilled SGPRs: 387 -> 387 (0.00 %)
Spilled VGPRs: 15 -> 15 (0.00 %)
Code Size: 61138824 -> 60928620 (-0.34 %) bytes
Max Waves: 48602 -> 48602 (0.00 %)
Instructions: 11967660 -> 11915084 (-0.44 %)

Signed-off-by: Rhys Perry <pendingchaos02@gmail.com>
Reviewed-by: Daniel Schürmann <daniel@schuermann.dev>
 2019-12-21 12:38:42 +01:00
Timur Kristóf
e0bcefc3a0 aco/wave32: Use lane mask regclass for exec/vcc. 
Currently all usages of exec and vcc are hardcoded to use s2 regclass.
This commit makes it possible to use s1 in wave32 mode and
s2 in wave64 mode.

Signed-off-by: Timur Kristóf <timur.kristof@gmail.com>
Reviewed-by: Daniel Schürmann <daniel@schuermann.dev>
 2019-12-04 10:36:01 +00:00
Daniel Schürmann
93c8ebfa78 aco: Initial commit of independent AMD compiler 
ACO (short for AMD Compiler) is a new compiler backend with the goal to replace
LLVM for Radeon hardware for the RADV driver.

ACO currently supports only VS, PS and CS on VI and Vega.
There are some optimizations missing because of unmerged NIR changes
which may decrease performance.

Full commit history can be found at
https://github.com/daniel-schuermann/mesa/commits/backend

Co-authored-by: Daniel Schürmann <daniel@schuermann.dev>
Co-authored-by: Rhys Perry <pendingchaos02@gmail.com>
Co-authored-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>
Co-authored-by: Connor Abbott <cwabbott0@gmail.com>
Co-authored-by: Michael Schellenberger Costa <mschellenbergercosta@googlemail.com>
Co-authored-by: Timur Kristóf <timur.kristof@gmail.com>

Acked-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Acked-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>
 2019-09-19 12:10:00 +02:00