fdo-mirrors/mesa
1
0
Fork 0
mirror of https://gitlab.freedesktop.org/mesa/mesa.git synced 2026-05-22 08:48:07 +02:00
Code Issues Projects Releases Packages Wiki Activity Actions 2
151456 commits 66 branches 936 tags 939 MiB
Commit graph

1835 commits

Author SHA1 Message Date
Rhys Perry
22050fcf9d radv,aco: lower vulkan_resource_index in NIR 
fossil-db (Sienna Cichlid):
Totals from 31338 (19.31% of 162293) affected shaders:
MaxWaves: 758634 -> 758616 (-0.00%)
Instrs: 26398289 -> 26378282 (-0.08%); split: -0.09%, +0.01%
CodeSize: 141048208 -> 140971060 (-0.05%); split: -0.07%, +0.01%
VGPRs: 1373656 -> 1373736 (+0.01%)
SpillSGPRs: 9944 -> 9924 (-0.20%); split: -0.24%, +0.04%
SpillVGPRs: 1892 -> 1898 (+0.32%); split: -0.95%, +1.27%
Latency: 308570144 -> 308528462 (-0.01%); split: -0.03%, +0.02%
InvThroughput: 57698072 -> 57684901 (-0.02%); split: -0.07%, +0.04%
VClause: 440357 -> 440602 (+0.06%); split: -0.02%, +0.08%
SClause: 974724 -> 973315 (-0.14%); split: -0.18%, +0.04%
Copies: 1944925 -> 1945103 (+0.01%); split: -0.11%, +0.12%
Branches: 799444 -> 799461 (+0.00%); split: -0.00%, +0.00%
PreSGPRs: 1619860 -> 1619233 (-0.04%); split: -0.05%, +0.02%
PreVGPRs: 1252813 -> 1252863 (+0.00%); split: -0.00%, +0.00%

Signed-off-by: Rhys Perry <pendingchaos02@gmail.com>
Reviewed-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-by: Timur Kristóf <timur.kristof@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12773>
 2022-03-22 16:33:27 +00:00
Rhys Perry
e6672f6fd1 radv: move radv_declare_shader_args() out of shader_variant_compile() 
Declaring them earlier will allow us to access them in NIR.

Signed-off-by: Rhys Perry <pendingchaos02@gmail.com>
Reviewed-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-by: Timur Kristóf <timur.kristof@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12773>
 2022-03-22 16:33:27 +00:00
Rhys Perry
53996cf979 aco: implement load_{scalar,vector}_arg_amd and load_smem_amd 
Signed-off-by: Rhys Perry <pendingchaos02@gmail.com>
Reviewed-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-by: Timur Kristóf <timur.kristof@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12773>
 2022-03-22 16:33:27 +00:00
Rhys Perry
177b54ebe9 aco/tests: add v_fma_mix tests 
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/14769>
 2022-03-17 19:04:17 +00:00
Rhys Perry
1092f37805 aco: use v_fma_mix to combine mul/add/fma output conversions 
fossil-db (Sienna Cichlid):
Totals from 42 (0.03% of 134913) affected shaders:
CodeSize: 596904 -> 596332 (-0.10%); split: -0.10%, +0.00%
Instrs: 110194 -> 109902 (-0.26%)
Latency: 1205239 -> 1204915 (-0.03%); split: -0.03%, +0.00%
InvThroughput: 189697 -> 189375 (-0.17%)
VClause: 1365 -> 1366 (+0.07%)
Copies: 5429 -> 5414 (-0.28%); split: -0.33%, +0.06%
Branches: 4034 -> 4026 (-0.20%)

fossil-db (Navi):
Totals from 42 (0.03% of 134913) affected shaders:
CodeSize: 596044 -> 595488 (-0.09%); split: -0.10%, +0.00%
Instrs: 110845 -> 110540 (-0.28%)
Latency: 1206131 -> 1205747 (-0.03%)
InvThroughput: 190178 -> 189809 (-0.19%)
VClause: 1372 -> 1370 (-0.15%); split: -0.29%, +0.15%
Copies: 5671 -> 5641 (-0.53%); split: -0.56%, +0.04%
Branches: 4033 -> 4025 (-0.20%)

fossil-db (Vega):
Totals from 42 (0.03% of 135048) affected shaders:
CodeSize: 605824 -> 605352 (-0.08%); split: -0.08%, +0.00%
Instrs: 115975 -> 115706 (-0.23%)
Latency: 1399845 -> 1398912 (-0.07%)
InvThroughput: 489901 -> 489442 (-0.09%)
VClause: 1314 -> 1311 (-0.23%); split: -0.38%, +0.15%
Copies: 9673 -> 9666 (-0.07%); split: -0.12%, +0.05%
Branches: 4025 -> 4024 (-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/14769>
 2022-03-17 19:04:17 +00:00
Rhys Perry
21304b772c aco: apply clamp to v_fma_mix 
fossil-db (Sienna Cichlid):
Totals from 2536 (1.88% of 134913) affected shaders:
CodeSize: 17314568 -> 17282960 (-0.18%)
Instrs: 3191438 -> 3187487 (-0.12%)
Latency: 59465090 -> 59407885 (-0.10%)
InvThroughput: 10271466 -> 10260512 (-0.11%)

fossil-db (Navi):
Totals from 2512 (1.86% of 134913) affected shaders:
CodeSize: 17194700 -> 17173396 (-0.12%)
Instrs: 3215093 -> 3212430 (-0.08%)
Latency: 60174315 -> 60142593 (-0.05%)
InvThroughput: 9491103 -> 9483979 (-0.08%)

fossil-db (Vega):
Totals from 2512 (1.86% of 135048) affected shaders:
CodeSize: 17186776 -> 17165472 (-0.12%)
Instrs: 3311166 -> 3308503 (-0.08%)
Latency: 65737409 -> 65716096 (-0.03%)
InvThroughput: 21735857 -> 21719792 (-0.07%)

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/14769>
 2022-03-17 19:04:17 +00:00
Rhys Perry
35196b6d89 aco: combine add/mul as v_fma_mix into fma 
fossil-db (Sienna Cichlid):
Totals from 7345 (5.44% of 134913) affected shaders:
CodeSize: 73840060 -> 73768936 (-0.10%); split: -0.10%, +0.00%
Instrs: 13701603 -> 13684183 (-0.13%); split: -0.13%, +0.00%
Latency: 185389373 -> 185306538 (-0.04%); split: -0.04%, +0.00%
InvThroughput: 33785020 -> 33757593 (-0.08%); split: -0.08%, +0.00%
VClause: 237337 -> 237338 (+0.00%)
SClause: 485728 -> 485720 (-0.00%)
Copies: 935900 -> 935279 (-0.07%); split: -0.07%, +0.00%
Branches: 480721 -> 480722 (+0.00%)

fossil-db (Navi):
Totals from 10649 (7.89% of 134913) affected shaders:
VGPRs: 756624 -> 756516 (-0.01%); split: -0.02%, +0.01%
CodeSize: 92156580 -> 91707900 (-0.49%); split: -0.49%, +0.00%
MaxWaves: 159402 -> 159476 (+0.05%); split: +0.07%, -0.02%
Instrs: 17155827 -> 17070449 (-0.50%); split: -0.50%, +0.00%
Latency: 246296456 -> 245487120 (-0.33%); split: -0.33%, +0.00%
InvThroughput: 41438159 -> 41117424 (-0.77%); split: -0.77%, +0.00%
VClause: 323790 -> 323867 (+0.02%); split: -0.00%, +0.03%
SClause: 612077 -> 612034 (-0.01%); split: -0.01%, +0.00%
Copies: 1103012 -> 1102775 (-0.02%); split: -0.03%, +0.01%
Branches: 555893 -> 555896 (+0.00%); split: -0.00%, +0.00%
PreSGPRs: 824372 -> 824378 (+0.00%)
PreVGPRs: 740390 -> 740363 (-0.00%); split: -0.01%, +0.01%

fossil-db (Vega):
Totals from 10950 (8.11% of 135048) affected shaders:
SGPRs: 1034528 -> 1034560 (+0.00%)
VGPRs: 794092 -> 794104 (+0.00%); split: -0.01%, +0.01%
CodeSize: 94409768 -> 93955568 (-0.48%); split: -0.48%, +0.00%
MaxWaves: 38950 -> 38939 (-0.03%); split: +0.00%, -0.03%
Instrs: 18162637 -> 18070934 (-0.50%); split: -0.51%, +0.00%
Latency: 291718455 -> 290772451 (-0.32%); split: -0.32%, +0.00%
InvThroughput: 109114674 -> 108489767 (-0.57%); split: -0.57%, +0.00%
VClause: 334498 -> 334579 (+0.02%); split: -0.01%, +0.03%
SClause: 628871 -> 628825 (-0.01%); split: -0.01%, +0.00%
Copies: 1674477 -> 1674850 (+0.02%); split: -0.02%, +0.04%
PreSGPRs: 834800 -> 834802 (+0.00%)
PreVGPRs: 750460 -> 750415 (-0.01%); split: -0.01%, +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/14769>
 2022-03-17 19:04:17 +00:00
Rhys Perry
9934c86761 aco: use v_fma_mix to combine mul/add/fma input conversions 
fossil-db (Sienna Cichlid):
Totals from 11558 (8.57% of 134913) affected shaders:
VGPRs: 829392 -> 825200 (-0.51%); split: -0.52%, +0.02%
SpillSGPRs: 7845 -> 8399 (+7.06%)
CodeSize: 101822704 -> 101677172 (-0.14%); split: -0.25%, +0.11%
MaxWaves: 172216 -> 173182 (+0.56%); split: +0.59%, -0.03%
Instrs: 19061343 -> 18883450 (-0.93%); split: -0.93%, +0.00%
Latency: 256011590 -> 255177378 (-0.33%); split: -0.39%, +0.06%
InvThroughput: 46104438 -> 45604059 (-1.09%); split: -1.12%, +0.04%
VClause: 352211 -> 351948 (-0.07%); split: -0.21%, +0.13%
SClause: 676506 -> 676961 (+0.07%); split: -0.04%, +0.11%
Copies: 1246571 -> 1237745 (-0.71%); split: -0.97%, +0.26%
Branches: 626229 -> 626241 (+0.00%); split: -0.02%, +0.03%
PreSGPRs: 882176 -> 888853 (+0.76%); split: -0.00%, +0.76%
PreVGPRs: 796705 -> 792304 (-0.55%); split: -0.56%, +0.00%

fossil-db (Navi):
Totals from 11558 (8.57% of 134913) affected shaders:
VGPRs: 803900 -> 798660 (-0.65%); split: -0.73%, +0.08%
SpillSGPRs: 7894 -> 8492 (+7.58%); split: -0.10%, +7.68%
CodeSize: 96892596 -> 97134716 (+0.25%); split: -0.05%, +0.29%
MaxWaves: 181454 -> 183014 (+0.86%); split: +0.94%, -0.08%
Instrs: 18186813 -> 18093994 (-0.51%); split: -0.56%, +0.05%
Latency: 253385909 -> 253325528 (-0.02%); split: -0.15%, +0.12%
InvThroughput: 43315355 -> 42805541 (-1.18%); split: -1.33%, +0.15%
VClause: 338755 -> 338535 (-0.06%); split: -0.16%, +0.10%
SClause: 656561 -> 656829 (+0.04%); split: -0.07%, +0.11%
Copies: 1162235 -> 1153558 (-0.75%); split: -1.07%, +0.32%
Branches: 588536 -> 588542 (+0.00%); split: -0.03%, +0.03%
PreSGPRs: 854849 -> 861640 (+0.79%); split: -0.00%, +0.80%
PreVGPRs: 783401 -> 779031 (-0.56%); split: -0.56%, +0.00%

fossil-db (Vega):
Totals from 11516 (8.53% of 135048) affected shaders:
SGPRs: 1072128 -> 1076288 (+0.39%); split: -0.01%, +0.40%
VGPRs: 821312 -> 818124 (-0.39%); split: -0.43%, +0.04%
SpillSGPRs: 11952 -> 12677 (+6.07%)
CodeSize: 96378496 -> 96707596 (+0.34%); split: -0.04%, +0.38%
MaxWaves: 42614 -> 42883 (+0.63%); split: +0.68%, -0.04%
Instrs: 18672844 -> 18600274 (-0.39%); split: -0.44%, +0.05%
Latency: 296658786 -> 296338296 (-0.11%); split: -0.21%, +0.10%
InvThroughput: 111665547 -> 111283559 (-0.34%); split: -0.40%, +0.06%
VClause: 343001 -> 342826 (-0.05%); split: -0.14%, +0.09%
SClause: 646684 -> 646657 (-0.00%); split: -0.05%, +0.04%
Copies: 1715316 -> 1712895 (-0.14%); split: -0.53%, +0.39%
PreSGPRs: 850737 -> 856543 (+0.68%); split: -0.04%, +0.72%
PreVGPRs: 775293 -> 772215 (-0.40%); split: -0.41%, +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/14769>
 2022-03-17 19:04:17 +00:00
Rhys Perry
eeef1bbe65 aco: refactor selection of mad/fma 
In the future, whether we need to use fma will depend on which
multiplication is chosen.

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/14769>
 2022-03-17 19:04:17 +00:00
Rhys Perry
e12bee3cb7 aco: improve support for v_fma_mix 
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/14769>
 2022-03-17 19:04:17 +00:00
Rhys Perry
79c8740c6e aco: fix fp16 opcode definitions 
The v_fma_mix optimizations assume v_cvt_f16_f32 and v_mul_f16 use a v2b
definition.

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/14769>
 2022-03-17 19:04:17 +00:00
Rhys Perry
c4cf92cad7 radv,aco,ac/llvm: fix indirect dispatches on the compute queue on GFX7-10 
Since neither PKT3_LOAD_SH_REG_INDEX nor PKT3_COPY_DATA work with compute
queues on GFX7-10, we have to load the dispatch size from memory in the
shader.

Signed-off-by: Rhys Perry <pendingchaos02@gmail.com>
Reviewed-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/15064>
 2022-03-14 19:54:36 +00:00
Rhys Perry
973967c49d aco: split and recombine unaligned sgpr inputs 
An example is the num_work_groups argument. Fixes invalid assembly with
func.compute.num-workgroups.basic.q0

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/15064>
 2022-03-14 19:54:36 +00:00
Daniel Schürmann
70aea6b41a aco/ra: refactor collect_vars() to return a sorted vector 
The vector of IDs is sorted with decreasing sizes,
and by increasing assigned registers.
This decouples register assingment from ssa IDs.

Totals from 12694 (9.41% of 134913) affected shaders: (GFX10.3)
VGPRs: 757864 -> 757848 (-0.00%); split: -0.00%, +0.00%
CodeSize: 72350540 -> 72348688 (-0.00%); split: -0.02%, +0.02%
MaxWaves: 237018 -> 237020 (+0.00%); split: +0.00%, -0.00%
Instrs: 13545494 -> 13544699 (-0.01%); split: -0.03%, +0.02%
Latency: 148539203 -> 148533292 (-0.00%); split: -0.01%, +0.00%
InvThroughput: 30319086 -> 30320382 (+0.00%); split: -0.01%, +0.01%
VClause: 326875 -> 327028 (+0.05%); split: -0.05%, +0.09%
SClause: 479833 -> 479837 (+0.00%); split: -0.00%, +0.00%
Copies: 862152 -> 860914 (-0.14%); split: -0.43%, +0.28%
Branches: 317775 -> 317777 (+0.00%)

Reviewed-by: Rhys Perry <pendingchaos02@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/11526>
 2022-03-14 08:32:10 +00:00
Daniel Schürmann
61c36b6dc0 aco/ra: refactor find_vars() to return a vector 
instead of std::set<>

No fossil-db changes.

Reviewed-by: Rhys Perry <pendingchaos02@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/11526>
 2022-03-14 08:32:10 +00:00
Daniel Schürmann
9476986e6f aco/ra: special-case get_reg_for_create_vector_copy() 
This function implements separate handling for
p_create_vector during get_regs_for_copies().
This simplifies some code and lets more precisely select
swap instructions if possible.

Totals from 876 (0.65% of 134913) affected shaders: (GFX10.3)
VGPRs: 53312 -> 53336 (+0.05%)
CodeSize: 3792936 -> 3788160 (-0.13%); split: -0.15%, +0.03%
MaxWaves: 16084 -> 16078 (-0.04%)
Instrs: 707449 -> 706385 (-0.15%); split: -0.19%, +0.04%
Latency: 6288293 -> 6286677 (-0.03%); split: -0.03%, +0.01%
InvThroughput: 4264450 -> 4263671 (-0.02%); split: -0.02%, +0.00%
VClause: 18655 -> 18679 (+0.13%); split: -0.20%, +0.33%
Copies: 55397 -> 54353 (-1.88%); split: -2.45%, +0.57%
Branches: 12426 -> 12415 (-0.09%)

Reviewed-by: Rhys Perry <pendingchaos02@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/11526>
 2022-03-14 08:32:10 +00:00
Daniel Schürmann
9181e8ceba aco/ra: count constant moves in get_reg_create_vector() 
Also implements a correct_pos_mask to keep track which
operands are already at the right target position.
To mitigate some regressions, call get_reg_impl() less often.

Totals from 1229 (0.91% of 134913) affected shaders: (GFX10.3)
VGPRs: 60216 -> 59848 (-0.61%)
CodeSize: 3716496 -> 3711268 (-0.14%); split: -0.19%, +0.05%
MaxWaves: 27952 -> 28004 (+0.19%)
Instrs: 685983 -> 685035 (-0.14%); split: -0.20%, +0.06%
Latency: 6727587 -> 6725340 (-0.03%); split: -0.06%, +0.02%
InvThroughput: 9289043 -> 9289866 (+0.01%); split: -0.02%, +0.03%
VClause: 17730 -> 17740 (+0.06%); split: -0.25%, +0.30%
Copies: 54352 -> 53420 (-1.71%); split: -2.46%, +0.75%
Branches: 12122 -> 12121 (-0.01%)

Reviewed-by: Rhys Perry <pendingchaos02@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/11526>
 2022-03-14 08:32:10 +00:00
Samuel Pitoiset
6c1c9067d9 aco: always emit vk_cvt_pkrtz_f16_f32 for nir_op_pack_half_2x16_split 
From the VK_KHR_shader_float_controls extension:

    "5) Do any of the “Pack” GLSL.std.450 instructions count as
     conversion instructions and have the rounding mode applied?"

    "RESOLVED: No, only instructions listed in “section 3.32.11.
     Conversion Instructions” of the SPIR-V specification count as
     conversion instructions."

This is also the same logic as the LLVM backend.

No fossils-db changes on Sienna Cichlid.

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/15301>
 2022-03-09 16:24:20 +00:00
Samuel Pitoiset
342e6f8332 radv,aco,llvm: lower post shuffle vertex in NIR 
fossils-db (Sienna Cichlid):
Totals from 774 (0.57% of 134913) affected shaders:
VGPRs: 26496 -> 26312 (-0.69%)
CodeSize: 1825936 -> 1828812 (+0.16%); split: -0.04%, +0.20%
MaxWaves: 22046 -> 22062 (+0.07%)
Instrs: 347634 -> 347975 (+0.10%); split: -0.05%, +0.15%
Latency: 1363949 -> 1356426 (-0.55%); split: -0.59%, +0.04%
InvThroughput: 221529 -> 221380 (-0.07%); split: -0.10%, +0.04%
VClause: 5682 -> 5676 (-0.11%); split: -1.46%, +1.36%
SClause: 7485 -> 7411 (-0.99%); split: -1.48%, +0.49%
Copies: 30481 -> 30420 (-0.20%); split: -0.51%, +0.31%
PreVGPRs: 19717 -> 19656 (-0.31%)

fossil-db (Polaris10):
Totals from 896 (0.66% of 135960) affected shaders:
SGPRs: 49824 -> 49648 (-0.35%); split: -0.39%, +0.03%
VGPRs: 31040 -> 29948 (-3.52%); split: -3.62%, +0.10%
CodeSize: 875960 -> 875920 (-0.00%); split: -0.06%, +0.05%
MaxWaves: 6380 -> 6429 (+0.77%)
Instrs: 171522 -> 171482 (-0.02%); split: -0.07%, +0.05%
Latency: 1356082 -> 1334386 (-1.60%); split: -1.61%, +0.01%
InvThroughput: 553389 -> 552957 (-0.08%); split: -0.08%, +0.00%
VClause: 4317 -> 4244 (-1.69%); split: -2.41%, +0.72%
SClause: 6157 -> 6139 (-0.29%); split: -0.45%, +0.16%
Copies: 9340 -> 9235 (-1.12%); split: -1.24%, +0.12%
PreVGPRs: 22366 -> 22116 (-1.12%)

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/15113>
 2022-03-08 19:18:01 +00:00
Rhys Perry
d068eb53e8 aco/insert_exec_mask: optimize top-level transition to exact before demote 
fossil-db (Sienna Cichlid):
Totals from 5767 (3.55% of 162293) affected shaders:
Instrs: 3264949 -> 3257527 (-0.23%); split: -0.23%, +0.00%
CodeSize: 17835692 -> 17806004 (-0.17%); split: -0.17%, +0.00%
Latency: 45990060 -> 45987924 (-0.00%); split: -0.00%, +0.00%
InvThroughput: 7643850 -> 7643835 (-0.00%); split: -0.00%, +0.00%
Copies: 193641 -> 186219 (-3.83%); split: -3.84%, +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/15244>
 2022-03-08 12:49:59 +00:00
Rhys Perry
42a5be975a aco/insert_exec_mask: use get_exec_op 
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/15244>
 2022-03-08 12:49:59 +00:00
Rhys Perry
aa55ecc296 aco/insert_exec_mask: fix top-level to-exact with non-global exact mask 
After transitioning to exact after a discard, the exec stack might be:
[exact|global, wqm, exact]

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/15244>
 2022-03-08 12:49:59 +00:00
Rhys Perry
d28b6b6856 aco: rework removal of jumps over branches 
Only allow this in situations where we know it's safe. In particular, this
stops removal of unconditional branches like with
block_kind_continue_or_break.

Fixes dEQP-VK.graphicsfuzz.fragcoord-control-flow hang.

fossil-db (Sienna Cichlid):
Totals from 34 (0.02% of 162293) affected shaders:
Instrs: 84115 -> 84178 (+0.07%); split: -0.00%, +0.08%
CodeSize: 463372 -> 463624 (+0.05%); split: -0.00%, +0.06%
Latency: 3467316 -> 3467652 (+0.01%)
InvThroughput: 3085493 -> 3085578 (+0.00%)
Branches: 3221 -> 3284 (+1.96%); split: -0.03%, +1.99%

Signed-off-by: Rhys Perry <pendingchaos02@gmail.com>
Reviewed-by: Daniel Schürmann <daniel@schuermann.dev>
Fixes: f030b75b7d ("aco: relax condition to remove branches in case of few instructions")
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/15214>
 2022-03-04 12:32:36 +00:00
Samuel Pitoiset
9b113f1b6c aco: implement nir_op_pack_{uint,sint}_2x16 
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/15231>
 2022-03-04 08:06:56 +00:00
Daniel Schürmann
ccf4bcd162 aco/ra: don't immediately assign a register for p_branch 
These get now assigned after handling phis.

Totals from 564 (0.42% of 134913) affected shaders: (GFX10.3)
CodeSize: 5519744 -> 5515308 (-0.08%)
Instrs: 1063045 -> 1061936 (-0.10%)
Latency: 11880452 -> 11875904 (-0.04%)
InvThroughput: 2259933 -> 2259581 (-0.02%); split: -0.02%, +0.00%
Copies: 86908 -> 85799 (-1.28%)

Reviewed-by: Rhys Perry <pendingchaos02@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/13432>
 2022-03-03 20:21:08 +00:00
Rhys Perry
c3070773f8 aco/tests: add test for branch definition RA 
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/13432>
 2022-03-03 20:21:08 +00:00
Rhys Perry
32d0bae8ec aco: fix branch definition validation 
Like how they have to be register allocated differently, branch
definitions at merge block predecessors need to be validated differently.

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/13432>
 2022-03-03 20:21:08 +00:00
Rhys Perry
bed5a31005 aco: add validate_instr_defs() 
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/13432>
 2022-03-03 20:21:08 +00:00
Rhys Perry
d5349a99c2 aco/ra: fix register allocation of branch definitions 
fossil-db (Sienna Cichlid):
Totals from 704 (0.52% of 134913) affected shaders:
CodeSize: 7177288 -> 7182072 (+0.07%); split: -0.00%, +0.07%
Instrs: 1371781 -> 1372977 (+0.09%); split: -0.00%, +0.09%
Latency: 17993572 -> 18001344 (+0.04%); split: -0.00%, +0.04%
InvThroughput: 4198996 -> 4199569 (+0.01%); split: -0.00%, +0.01%
Copies: 122456 -> 123516 (+0.87%); split: -0.01%, +0.88%
Branches: 43815 -> 43818 (+0.01%); split: -0.02%, +0.03%

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/13432>
 2022-03-03 20:21:08 +00:00
Rhys Perry
608d48b787 aco/ra: add get_reg_phi() helper 
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/13432>
 2022-03-03 20:21:08 +00:00
Rhys Perry
ceca5e68c4 aco: remove vcc hint from branch definitions 
This doesn't seem to have much benefit anymore.

fossil-db (Sienna Cichlid):
Totals from 198 (0.15% of 134913) affected shaders:
CodeSize: 2610536 -> 2610872 (+0.01%); split: -0.01%, +0.02%
Instrs: 479001 -> 479085 (+0.02%); split: -0.01%, +0.03%
Latency: 7310684 -> 7300735 (-0.14%); split: -0.16%, +0.02%
InvThroughput: 2439084 -> 2437446 (-0.07%); split: -0.07%, +0.00%
SClause: 14760 -> 14722 (-0.26%)

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/13432>
 2022-03-03 20:21:08 +00:00
Samuel Pitoiset
516aee64cc radv,aco: do not lower nir_op_pack_{unorm,snorm}_2x16 
v_cvt_pknorm_{u16,i16}_f32 can be emitted instead, it's supported on
all generations.

No fossils-db changes.

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/15215>
 2022-03-03 14:54:12 +01:00
Timur Kristóf
11957d3863 aco: Remove superfluous code for mesh shader workgroup ID. 
Signed-off-by: Timur Kristóf <timur.kristof@gmail.com>
Reviewed-by: Rhys Perry <pendingchaos02@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/15199>
 2022-03-01 15:37:12 +00:00
Daniel Schürmann
f030b75b7d aco: relax condition to remove branches in case of few instructions 
This patch relaxes the conditions under which
we remove branch instructions.

Totals from 27246 (20.20% of 134913) affected shaders: (GFX10.3)
CodeSize: 193413312 -> 192924928 (-0.25%)
Instrs: 36146788 -> 36024692 (-0.34%)
Latency: 528374112 -> 528469044 (+0.02%); split: -0.01%, +0.02%
InvThroughput: 106198759 -> 106216583 (+0.02%); split: -0.00%, +0.02%
Branches: 1040640 -> 918543 (-11.73%)

Reviewed-by: Timur Kristóf <timur.kristof@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/8647>
 2022-02-25 15:38:08 +00:00
Timur Kristóf
1ca6b2f216 aco: Support memory modes properly with load/store_buffer_amd. 
Signed-off-by: Timur Kristóf <timur.kristof@gmail.com>
Reviewed-by: Rhys Perry <pendingchaos02@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/15161>
 2022-02-25 14:08:39 +01:00
Timur Kristóf
ba4b48e787 aco: Support task_payload with barriers, refactor allowed storage class. 
Signed-off-by: Timur Kristóf <timur.kristof@gmail.com>
Reviewed-by: Rhys Perry <pendingchaos02@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/15161>
 2022-02-25 14:08:36 +01:00
Timur Kristóf
cd0dd5d6b7 aco: Add storage class for Task Shader payload. 
Signed-off-by: Timur Kristóf <timur.kristof@gmail.com>
Reviewed-by: Rhys Perry <pendingchaos02@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/15161>
 2022-02-25 13:20:08 +01:00
Timur Kristóf
9cc9cf77a8 aco: Fix multiview view index for mesh shaders. 
Signed-off-by: Timur Kristóf <timur.kristof@gmail.com>
Reviewed-by: Rhys Perry <pendingchaos02@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/15034>
 2022-02-25 06:31:33 +00:00
Timur Kristóf
082b691141 aco: Fix workgroup_id.y and .z for NV_mesh_shader. 
Signed-off-by: Timur Kristóf <timur.kristof@gmail.com>
Reviewed-by: Rhys Perry <pendingchaos02@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/15034>
 2022-02-25 06:31:33 +00:00
Timur Kristóf
10ebfb3bf2 aco: Allow 1-byte loads and stores with load/store_buffer_amd 
Signed-off-by: Timur Kristóf <timur.kristof@gmail.com>
Reviewed-by: Rhys Perry <pendingchaos02@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/15034>
 2022-02-25 06:31:33 +00:00
Georg Lehmann
d223b7f096 radv, aco: Add u_foreach_bit to .clang-format. 
Signed-off-by: Georg Lehmann <dadschoorse@gmail.com>
Reviewed-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/15083>
 2022-02-22 14:57:29 +00:00
Samuel Pitoiset
369b8cffea radv,aco,llvm: lower adjusting vertex alpha in NIR 
Instead of duplicating the same lowering in both compiler backends.
This pass will be used to do more VS input lowering.

fossils-db (Polaris10):
Totals from 48 (0.04% of 135960) affected shaders:
VGPRs: 1692 -> 1684 (-0.47%)
CodeSize: 54016 -> 53964 (-0.10%); split: -0.11%, +0.01%
MaxWaves: 339 -> 341 (+0.59%)
Instrs: 11260 -> 11247 (-0.12%); split: -0.13%, +0.02%
Latency: 88165 -> 88113 (-0.06%); split: -0.07%, +0.01%
InvThroughput: 36153 -> 36093 (-0.17%)
Copies: 583 -> 568 (-2.57%)

fossils-db (Pitcairn):
Totals from 43 (0.03% of 135960) affected shaders:
VGPRs: 1548 -> 1552 (+0.26%)
CodeSize: 47900 -> 47820 (-0.17%)
Instrs: 10751 -> 10731 (-0.19%)
Latency: 83029 -> 82873 (-0.19%)
VClause: 168 -> 164 (-2.38%)
SClause: 393 -> 391 (-0.51%)
Copies: 705 -> 685 (-2.84%)

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/15076>
 2022-02-22 08:08:42 +00:00
Samuel Pitoiset
cbd5724a6d aco: implement nir_intrinsic_load_vrs_rates_amd 
Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-by: Timur Kristóf <timur.kristof@gmail.com>
Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/14713>
 2022-02-16 08:11:19 +01:00
Daniel Schürmann
1bbbabedb7 aco/insert_exec_mask: refactor and remove some unnecessary WQM handling code 
Some cases cannot happen and don't need to be handled anymore.

Reviewed-by: Rhys Perry <pendingchaos02@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/14951>
 2022-02-11 19:05:30 +00:00
Daniel Schürmann
d7d7b9974a aco/insert_exec_mask: refactor and simplify get_block_needs() 
Reviewed-by: Rhys Perry <pendingchaos02@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/14951>
 2022-02-11 19:05:30 +00:00
Daniel Schürmann
fcc5dec8d6 aco/insert_exec_mask: remove ever_again_needs and Exact_Branch 
This information is not required anymore.

Reviewed-by: Rhys Perry <pendingchaos02@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/14951>
 2022-02-11 19:05:30 +00:00
Daniel Schürmann
cbb1b095ca aco/insert_exec_mask: remove some unnecessary WQM loop handling code 
These workarounds are were necessary to prevent infinite loops
with helper lane registers containing wrong data.

Reviewed-by: Rhys Perry <pendingchaos02@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/14951>
 2022-02-11 19:05:30 +00:00
Daniel Schürmann
580a63b4ac aco/insert_exec_mask: remove Preserve_WQM flag 
If WQM is needed anywhere after discard_if(), it will also
be flagged as WQM. We can rely on that to preserve the WQM mask.

Reviewed-by: Rhys Perry <pendingchaos02@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/14951>
 2022-02-11 19:05:30 +00:00
Daniel Schürmann
a5a40beefa aco: don't emit WQM for bool_to_scalar_condition 
This was only necessary to ensure that the source is computed in WQM
if the current exec mask is in WQM.

Totals from 23170 (17.17% of 134913) affected shaders: (GFX10.3)
VGPRs: 1384464 -> 1383400 (-0.08%); split: -0.08%, +0.01%
SpillSGPRs: 7575 -> 7574 (-0.01%)
CodeSize: 146444752 -> 146317104 (-0.09%); split: -0.13%, +0.04%
MaxWaves: 429870 -> 429868 (-0.00%)
Instrs: 27202586 -> 27170316 (-0.12%); split: -0.17%, +0.05%
Latency: 379488313 -> 379335412 (-0.04%); split: -0.07%, +0.03%
InvThroughput: 69500561 -> 69487704 (-0.02%); split: -0.03%, +0.01%
VClause: 473080 -> 473038 (-0.01%); split: -0.02%, +0.01%
SClause: 1080576 -> 1080571 (-0.00%); split: -0.06%, +0.06%
Copies: 1492865 -> 1504604 (+0.79%); split: -0.40%, +1.19%
Branches: 711295 -> 716849 (+0.78%); split: -0.01%, +0.79%
PreSGPRs: 1331362 -> 1328402 (-0.22%)

Reviewed-by: Rhys Perry <pendingchaos02@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/14951>
 2022-02-11 19:05:30 +00:00
Daniel Schürmann
f816dd1be7 aco: don't propagate WQM for p_as_uniform 
This was needed, so that in case of active helper lanes,
these contain the correct value. It is now handled implicitly.

Totals from 1004 (0.74% of 134913) affected shaders: (GFX10.3)
CodeSize: 7581020 -> 7580892 (-0.00%); split: -0.00%, +0.00%
Instrs: 1454940 -> 1454908 (-0.00%); split: -0.00%, +0.00%
Latency: 12984953 -> 12984894 (-0.00%); split: -0.00%, +0.00%
InvThroughput: 3173037 -> 3173049 (+0.00%); split: -0.00%, +0.00%
PreSGPRs: 47498 -> 47273 (-0.47%)

Reviewed-by: Rhys Perry <pendingchaos02@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/14951>
 2022-02-11 19:05:30 +00:00