fdo-mirrors/mesa
1
0
Fork 0
mirror of https://gitlab.freedesktop.org/mesa/mesa.git synced 2025-12-23 06:50:11 +01:00
Code Issues Projects Releases Packages Wiki Activity Actions 1
128297 commits 64 branches 914 tags 1 GiB
Commit graph

904 commits

Author SHA1 Message Date
Timur Kristóf
26299c87f8 aco: Add base argument to emit_mbcnt. 
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/6699>
 2020-09-14 12:19:24 +00:00
Timur Kristóf
f3780e7b8c aco: Clean up emit_mbcnt. 
Make it less error-prone and more consistent with other helpers.
Pass the masks as a single argument rather than two.
In wave64 mode, split the argument into low and high halves in
emit_mbcnt rather than where it is called.

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/6699>
 2020-09-14 12:19:24 +00:00
Timur Kristóf
efa1c760d1 aco: Fix emit_boolean_exclusive_scan in wave32 mode. 
Use the lane mask instead of s2 for the register class.

Cc: mesa-stable@lists.freedesktop.org
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/6699>
 2020-09-14 12:19:24 +00:00
Rhys Perry
e394ff9b5f aco: fix validation of sub-dword parallel-copies 
Only the operands with a sub-dword definition need to be checked.

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/6662>
 2020-09-09 15:00:45 +00:00
Rhys Perry
834b449a46 aco: fix value numbering of reductions 
Non-ssa definitions caused an assertion in value numbering.

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/6662>
 2020-09-09 15:00:45 +00:00
Rhys Perry
5a8447cbd4 aco: don't apply constant to SDWA on GFX8 
Signed-off-by: Rhys Perry <pendingchaos02@gmail.com>
Reviewed-by: Daniel Schürmann <daniel@schuermann.dev>
Cc: 20.2 <mesa-stable>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/6662>
 2020-09-09 15:00:45 +00:00
Rhys Perry
29cb3cba9f aco: workaround disassembler bug of v_writelane_b32 with literal 
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/6662>
 2020-09-09 15:00:45 +00:00
Rhys Perry
36e58a14cc aco: fix v_writelane_b32 with two sgprs 
v_writelane_b32 can take two sgprs but only if one is m0.

Signed-off-by: Rhys Perry <pendingchaos02@gmail.com>
Reviewed-by: Daniel Schürmann <daniel@schuermann.dev>
Cc: 20.2 <mesa-stable>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/6662>
 2020-09-09 15:00:45 +00:00
Tony Wasserka
fefeaeef06 aco/isel: Compile all helper functions with static linkage 
Reviewed-by: Daniel Schürmann <daniel@schuermann.dev>
Reviewed-by: Timur Kristóf <timur.kristof@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/6504>
 2020-09-08 20:13:51 +00:00
Tony Wasserka
793dc668ea aco/isel: Move add_startpgm to aco_instruction_selection.cpp 
Reviewed-by: Daniel Schürmann <daniel@schuermann.dev>
Reviewed-by: Timur Kristóf <timur.kristof@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/6504>
 2020-09-08 20:13:51 +00:00
Tony Wasserka
47de553283 aco/isel: Move context initialization code to a dedicated file 
aco_instruction_selection_setup.cpp (previously used as a header) has
been split into a header and an implementation file. The latter "only"
implements init_context and setup_isel_context, but since these files
carry a long trail of helper functions, this cleans up the isel header
a lot.

Reduces library size by 3.1% due to more functions being compiled with
static linkage. Makes aco_instruction_selection.cpp compile 3% faster.

Reviewed-by: Daniel Schürmann <daniel@schuermann.dev>
Reviewed-by: Timur Kristóf <timur.kristof@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/6504>
 2020-09-08 20:13:51 +00:00
Tony Wasserka
1eac0b52e3 aco/isel: Remove unused definitions 
Reviewed-by: Daniel Schürmann <daniel@schuermann.dev>
Reviewed-by: Timur Kristóf <timur.kristof@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/6504>
 2020-09-08 20:13:51 +00:00
Tony Wasserka
150de6358d aco/isel: Consistently use references for input parameters in emit_load 
Reviewed-by: Daniel Schürmann <daniel@schuermann.dev>
Reviewed-by: Timur Kristóf <timur.kristof@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/6504>
 2020-09-08 20:13:51 +00:00
Tony Wasserka
dab0af0616 aco/isel: Simplify nested branching code 
Reviewed-by: Daniel Schürmann <daniel@schuermann.dev>
Reviewed-by: Timur Kristóf <timur.kristof@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/6504>
 2020-09-08 20:13:51 +00:00
Tony Wasserka
757de68a43 aco/isel: Turn the function template emit_load into a proper function 
Statically known values were encoded using template parameters previously,
causing specializations for each of the 5 sets of template arguments to be
generated. Since emit_load is not performance critical (the inner loop
never runs more often than twice), it's better for build time to use
runtime arguments everywhere.

Reduces build time of this file by 9% (17.3s -> 15.7s on my machine) and
reduces libaco's size by 2.6%.

Reviewed-by: Daniel Schürmann <daniel@schuermann.dev>
Reviewed-by: Timur Kristóf <timur.kristof@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/6504>
 2020-09-08 20:13:51 +00:00
Daniel Schürmann
0b6448bbe7 aco/isel: refactor emit_vop3a_instruction() to handle 2 operand instructions 
Only AC:O has been affected.

Totals from 4 (0.00% of 136546) affected shaders (RAVEN):
CodeSize: 16428 -> 16420 (-0.05%)
Instrs: 3294 -> 3292 (-0.06%)
Cycles: 14208 -> 14200 (-0.06%)
VMEM: 936 -> 978 (+4.49%)
VClause: 80 -> 77 (-3.75%)
Copies: 211 -> 209 (-0.95%)
PreVGPRs: 127 -> 126 (-0.79%)

Reviewed-by: Rhys Perry <pendingchaos02@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/6635>
 2020-09-08 16:20:44 +00:00
Daniel Schürmann
5b31056257 aco/isel: refactor code and remove unnecessary v_mov 
Changes mainly due to avoided v_movs for fmin/fmax/fadd/fmul.

Totals from 12783 (9.36% of 136546) affected shaders (RAVEN):
SGPRs: 1097752 -> 1098264 (+0.05%); split: -0.09%, +0.14%
VGPRs: 856920 -> 850800 (-0.71%); split: -0.82%, +0.11%
SpillSGPRs: 49494 -> 49496 (+0.00%); split: -0.00%, +0.01%
CodeSize: 99997916 -> 99989948 (-0.01%); split: -0.04%, +0.03%
MaxWaves: 53895 -> 54448 (+1.03%)
Instrs: 19634960 -> 19632626 (-0.01%); split: -0.05%, +0.04%
Cycles: 1620601696 -> 1620900712 (+0.02%); split: -0.02%, +0.04%
VMEM: 3334181 -> 3299626 (-1.04%); split: +1.62%, -2.66%
SMEM: 865573 -> 865876 (+0.04%); split: +0.84%, -0.81%
VClause: 337100 -> 335224 (-0.56%); split: -0.88%, +0.32%
SClause: 696813 -> 697267 (+0.07%); split: -0.14%, +0.21%
Copies: 1549897 -> 1548023 (-0.12%); split: -0.52%, +0.40%
Branches: 682118 -> 682108 (-0.00%); split: -0.01%, +0.00%
PreSGPRs: 893524 -> 895129 (+0.18%); split: -0.00%, +0.18%
PreVGPRs: 790180 -> 783036 (-0.90%)

Reviewed-by: Rhys Perry <pendingchaos02@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/6635>
 2020-09-08 16:20:44 +00:00
Rhys Perry
6049dc1a9d aco: improve fsign selection 
Idea from https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/6284

fossil-db (Navi):
Totals from 4053 (2.95% of 137413) affected shaders:
SGPRs: 305810 -> 305906 (+0.03%); split: -0.01%, +0.04%
VGPRs: 249000 -> 249144 (+0.06%); split: -0.01%, +0.07%
CodeSize: 29967092 -> 29885768 (-0.27%); split: -0.27%, +0.00%
Instrs: 5749494 -> 5737971 (-0.20%); split: -0.20%, +0.00%
Cycles: 255028584 -> 254955444 (-0.03%); split: -0.04%, +0.01%

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/6583>
 2020-09-08 12:17:43 +00:00
Rhys Perry
0e9425a1b7 aco: fix one-off error in Operand(uint16_t) 
Signed-off-by: Rhys Perry <pendingchaos02@gmail.com>
Reviewed-by: Daniel Schürmann <daniel@schuermann.dev>
Fixes: 3d6f67950d ('aco: improve 8/16-bit constants')
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/6583>
 2020-09-08 12:17:43 +00:00
Samuel Pitoiset
73eb24ab31 aco: handle unaligned loads on GFX10.3 
Same as GFX10.

Cc: 20.2 mesa-stable
Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-by: Rhys Perry <pendingchaos02@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/6594>
 2020-09-04 13:19:45 +00:00
Rhys Perry
8faf85f687 aco: fix byte_align_scalar for 3 dword vectors 
Signed-off-by: Rhys Perry <pendingchaos02@gmail.com>
Reviewed-by: Daniel Schürmann <daniel@schuermann.dev>
Fixes: fe08f0ccf9
   ('aco: add byte_align_scalar() & trim_subdword_vector() helper functions')

Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/4710>
 2020-09-04 13:03:50 +00:00
Samuel Pitoiset
ebf2576862 radv,aco: disable opts if VK_PIPELINE_CREATE_DISABLE_OPTIMIZATION_BIT 
Sounds useful to determine if ACO breaks a specific pipeline
because of various optimizations.

Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-by: Daniel Schürmann <daniel@schuermann.dev>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/6487>
 2020-09-04 06:59:45 +00:00
Tony Wasserka
2182bbf84f aco: Fix integer overflows when emitting parallel copies during RA 
32-bit shifts were accidentally used before this change despite the intended
output being 64 bits.

This was observed when compiling Dolphin's ubershaders.

Cc: mesa-stable
Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>
Reviewed-by: Daniel Schürmann <daniel@schuermann.dev>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/6568>
 2020-09-03 20:20:24 +00:00
Rhys Perry
a99ae1943d aco: remove omod_success/clamp_success 
This simplifies the optimizer and should make SDWA optimizations easier.

No fossil-db changes on Navi.

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/6293>
 2020-09-03 19:06:52 +00:00
Rhys Perry
dc79152be9 aco: fix mad splitting after applying output modifiers 
Previously, this wasn't done because the mad label wasn't passed to the
new definition.

fossil-db (Navi):
Totals from 5770 (4.24% of 135946) affected shaders:
SGPRs: 391920 -> 391872 (-0.01%)
VGPRs: 349084 -> 348424 (-0.19%); split: -0.20%, +0.01%
CodeSize: 34639636 -> 34637496 (-0.01%); split: -0.02%, +0.01%
MaxWaves: 58828 -> 58862 (+0.06%)
Instrs: 6723436 -> 6723297 (-0.00%); split: -0.02%, +0.02%
Cycles: 197594168 -> 197591968 (-0.00%); split: -0.02%, +0.02%

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/6293>
 2020-09-03 19:06:52 +00:00
Karol Herbst
70cbddc4a7 nir: use enum operator helper for nir_variable_mode and nir_metadata 
those are used quite a bit

Signed-off-by: Karol Herbst <kherbst@redhat.com>
Reviewed-by: Jason Ekstrand <jason@jlekstrand.net>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/6520>
 2020-09-01 17:45:08 +00:00
Timur Kristóf
68957a8256 aco: Add README which explains about what ACO is and how it works. 
This is based on an early summary written by Daniel, but updated to
reflect the current state of ACO and reworded to better fit the format.
Also added is a table which details what SW stages correspond to what
HW stages on each HW generation.

Signed-off-by: Timur Kristóf <timur.kristof@gmail.com>
Reviewed-by: Tony Wasserka <tony.wasserka@gmx.de>
Reviewed-by: Rhys Perry <pendingchaos02@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/4164>
 2020-09-01 10:41:34 +00:00
Timur Kristóf
b4d4c548eb aco: Fixup markdown formatting of the README-ISA. 
Signed-off-by: Timur Kristóf <timur.kristof@gmail.com>
Reviewed-by: Tony Wasserka <tony.wasserka@gmx.de>
Reviewed-by: Rhys Perry <pendingchaos02@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/4164>
 2020-09-01 10:41:34 +00:00
Timur Kristóf
086fafc4e0 aco: Move README to README-ISA 
The old "readme" is not really a readme but rather just
a bunch of notes with our findings about the GCN/RDNA ISA.

Signed-off-by: Timur Kristóf <timur.kristof@gmail.com>
Reviewed-by: Tony Wasserka <tony.wasserka@gmx.de>
Reviewed-by: Rhys Perry <pendingchaos02@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/4164>
 2020-09-01 10:41:34 +00:00
Samuel Pitoiset
8076c7596d aco: fix wrong source position for constant with nir_op_cube_face_coord 
Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-by: Rhys Perry <pendingchaos02@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/6480>
 2020-08-28 08:03:55 +02:00
Samuel Pitoiset
502b9daa7a aco: add ACO_DEBUG=novn,noopt,nosched for debugging purposes 
To disable value numbering, optimizations and scheduling.

Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-by: Rhys Perry <pendingchaos02@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/6470>
 2020-08-27 10:23:51 +00:00
Rhys Perry
d2cf6a8399 aco: sink get_alu_src() in bfe lowering 
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/6424>
 2020-08-26 13:46:23 +00:00
Rhys Perry
14d748eb28 aco: fix sgpr ubfe/ibfe if the offset is too large 
If the offset is large enough, it could affect the width. I'm also not
sure if the hardware masks the offset by 0x1f.

Found by inspection. No fossil-db changes.

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/6424>
 2020-08-26 13:46:23 +00:00
Rhys Perry
454bc595d1 aco: remove 64-bit SGPR ubfe/ibfe 
ubfe/ibfe is always 32-bit.

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/6424>
 2020-08-26 13:46:23 +00:00
Rhys Perry
eb3c16e1f8 aco/tests: add tests for long jumps 
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/6212>
 2020-08-26 13:26:58 +00:00
Rhys Perry
192b9f4303 aco: shorten disassembly for repeated instructions 
Future tests will do this.

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/6212>
 2020-08-26 13:26:58 +00:00
Rhys Perry
ae6330d955 aco/tests: add test for GFX10 0x3f bug 
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/6212>
 2020-08-26 13:26:58 +00:00
Rhys Perry
fe2dc41258 aco: create long jumps 
When the branch offset can't be encoded, we have to use s_setpc_b64.

Fixes hang in RPCS3 vertex ubershader.

Signed-off-by: Rhys Perry <pendingchaos02@gmail.com>
Reviewed-by: Daniel Schürmann <daniel@schuermann.dev>
Closes: https://gitlab.freedesktop.org/mesa/mesa/-/issues/3231
Cc: 20.2 <mesa-stable>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/6212>
 2020-08-26 13:26:58 +00:00
Rhys Perry
156fd58cda aco: reserve 2 sgprs for each branch 
We'll need two sgprs for the possibility of a long jump.

fossil-db (Navi):
Totals from 10197 (7.50% of 135946) affected shaders:
SGPRs: 946268 -> 946468 (+0.02%)
VGPRs: 705884 -> 707956 (+0.29%); split: -0.00%, +0.30%
SpillSGPRs: 31485 -> 36212 (+15.01%); split: -0.04%, +15.05%
CodeSize: 88296484 -> 88384604 (+0.10%); split: -0.01%, +0.11%
MaxWaves: 81379 -> 81171 (-0.26%)
Instrs: 17219111 -> 17231682 (+0.07%); split: -0.03%, +0.10%
Cycles: 1594875900 -> 1596450136 (+0.10%); split: -0.05%, +0.15%
VMEM: 1687263 -> 1689080 (+0.11%); split: +0.14%, -0.03%
SMEM: 657726 -> 660262 (+0.39%); split: +0.61%, -0.22%
VClause: 294806 -> 294638 (-0.06%); split: -0.08%, +0.02%
SClause: 556702 -> 556210 (-0.09%); split: -0.12%, +0.03%
Copies: 1466323 -> 1469349 (+0.21%); split: -0.57%, +0.78%
Branches: 619793 -> 618556 (-0.20%); split: -0.28%, +0.08%
PreSGPRs: 806364 -> 811477 (+0.63%); split: -0.14%, +0.77%
PreVGPRs: 655845 -> 657174 (+0.20%)

Signed-off-by: Rhys Perry <pendingchaos02@gmail.com>
Reviewed-by: Daniel Schürmann <daniel@schuermann.dev>
Cc: 20.2 <mesa-stable>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/6212>
 2020-08-26 13:26:58 +00:00
Rhys Perry
e8ac14527a aco: keep loop live-through variables spilled 
fossil-db (Navi):
Totals from 3149 (2.32% of 135946) affected shaders:
VGPRs: 280928 -> 280932 (+0.00%)
SpillSGPRs: 51133 -> 30042 (-41.25%)
CodeSize: 43063076 -> 41377252 (-3.91%); split: -3.92%, +0.00%
Instrs: 8278435 -> 8037133 (-2.91%); split: -2.92%, +0.00%
Cycles: 709575456 -> 683366172 (-3.69%); split: -3.69%, +0.00%
VMEM: 542887 -> 542937 (+0.01%); split: +0.05%, -0.04%
SMEM: 210255 -> 206368 (-1.85%); split: +0.12%, -1.97%
SClause: 258847 -> 258019 (-0.32%); split: -0.52%, +0.20%
Copies: 731836 -> 684784 (-6.43%); split: -6.44%, +0.01%
Branches: 305422 -> 292844 (-4.12%); split: -4.12%, +0.00%
PreSGPRs: 333103 -> 332701 (-0.12%)
PreVGPRs: 280086 -> 280089 (+0.00%)

Helps mostly Detroit: Become Human and the single spilling Doom Eternal
shader.

Signed-off-by: Rhys Perry <pendingchaos02@gmail.com>
Reviewed-by: Daniel Schürmann <daniel@schuermann.dev>
Cc: 20.2 <mesa-stable>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/6212>
 2020-08-26 13:26:58 +00:00
Rhys Perry
75d6c30572 aco: fix spills_entry heuristic for branch blocks in init_live_in_vars() 
fossil-db (Navi):
Totals from 222 (0.16% of 135946) affected shaders:
SpillSGPRs: 9121 -> 9117 (-0.04%)
SpillVGPRs: 2820 -> 1821 (-35.43%)
CodeSize: 5134264 -> 5053336 (-1.58%); split: -1.63%, +0.05%
Instrs: 953435 -> 938761 (-1.54%); split: -1.59%, +0.05%
Cycles: 100567688 -> 97252432 (-3.30%); split: -3.34%, +0.04%
VMEM: 40752 -> 39219 (-3.76%); split: +0.04%, -3.80%
SMEM: 15416 -> 15509 (+0.60%); split: +0.64%, -0.03%
VClause: 20120 -> 19091 (-5.11%)
SClause: 23540 -> 23544 (+0.02%); split: -0.11%, +0.12%
Copies: 125912 -> 122017 (-3.09%); split: -3.36%, +0.26%
Branches: 31131 -> 30009 (-3.60%)

Mostly affects parallel-rdp ubershaders.

Signed-off-by: Rhys Perry <pendingchaos02@gmail.com>
Reviewed-by: Daniel Schürmann <daniel@schuermann.dev>
Cc: 20.2 <mesa-stable>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/6212>
 2020-08-26 13:26:58 +00:00
Rhys Perry
fc9f502a5b aco: fix regclass checks when fixing to vcc/exec with Builder 
Signed-off-by: Rhys Perry <pendingchaos02@gmail.com>
Reviewed-by: Daniel Schürmann <daniel@schuermann.dev>
Cc: 20.2 <mesa-stable>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/6212>
 2020-08-26 13:26:58 +00:00
Rhys Perry
a537c9e73f aco: don't fix break condition for break+discard to exec 
This would move the old exec mask back into exec. This also fixes the
live_out_exec.

Issue found in dEQP-VK.graphicsfuzz.cosh-return-inf-unused

Signed-off-by: Rhys Perry <pendingchaos02@gmail.com>
Reviewed-by: Daniel Schürmann <daniel@schuermann.dev>
Cc: 20.2 <mesa-stable>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/6212>
 2020-08-26 13:26:58 +00:00
Rhys Perry
1a5444b900 aco: don't consider the first partial spill if it's the wrong type 
Signed-off-by: Rhys Perry <pendingchaos02@gmail.com>
Reviewed-by: Daniel Schürmann <daniel@schuermann.dev>
Cc: 20.2 <mesa-stable>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/6212>
 2020-08-26 13:26:58 +00:00
Rhys Perry
8f6a900d5e aco: consider branch definitions in spiller 
Signed-off-by: Rhys Perry <pendingchaos02@gmail.com>
Reviewed-by: Daniel Schürmann <daniel@schuermann.dev>
Cc: 20.2 <mesa-stable>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/6212>
 2020-08-26 13:26:58 +00:00
Daniel Schürmann
a79dad950b nir,amd: remove trinary_minmax opcodes 
These consist of the variations nir_op_{i|u|f}{min|max|med}3 which are either
lowered in the backend (LLVM) anyway or can be recombined by the backend (ACO).

Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/6421>
 2020-08-24 20:56:11 +00:00
Timur Kristóf
f820dde201 aco: Fix convert_to_SDWA when instruction has 3 operands. 
Previously, when the instruction had 3 operands, this would cause
possible corruption because of writing to sdwa->sel[2].
This was noticed thanks to GCC 10's -Wstringop-overflow warning.

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/6436>
 2020-08-24 15:55:14 +02:00
Timur Kristóf
0d194a70c6 aco: Fix unused variable warning by adding ASSERTED. 
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/6436>
 2020-08-24 15:55:06 +02:00
Samuel Pitoiset
8fd2f5c16d radv: add a small interface for creating the trap handler shader 
Similar to the GS copy shader except that NIR is unused because
the shader is written directly using ACO IR.

Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/6384>
 2020-08-24 11:08:24 +00:00
Samuel Pitoiset
a0814a873d aco: skip unnecessary compiler pass for the trap handler program 
The shader is written by hands with assigned registers, so most of
the pass are unnecessary.

Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-by: Daniel Schürmann <daniel@schuermann.dev>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/6384>
 2020-08-24 11:08:24 +00:00