AlexIndustrial/mesa
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity
153,316 Commits 1 Branch 0 Tags
69cceab7182be4fee4f1dbba2d3fe5370ba2333a
Commit Graph

1887 Commits

Author SHA1 Message Date
Georg Lehmann 69cceab718 aco: Remove D16 zero components from image stores. 
No foz-db changes.

Signed-off-by: Georg Lehmann <dadschoorse@gmail.com>
Reviewed-by: Daniel Schürmann <daniel@schuermann.dev>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/15179>
 2022-05-04 09:58:03 +00:00
Georg Lehmann 7a6dbe0c77 aco: Implement image_load d16. 
Signed-off-by: Georg Lehmann <dadschoorse@gmail.com>
Reviewed-by: Rhys Perry <pendingchaos02@gmail.com>
Reviewed-by: Daniel Schürmann <daniel@schuermann.dev>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/15179>
 2022-05-04 09:58:03 +00:00
Georg Lehmann 7ffcaf9187 aco: Implement image_store d16. 
Signed-off-by: Georg Lehmann <dadschoorse@gmail.com>
Reviewed-by: Rhys Perry <pendingchaos02@gmail.com>
Reviewed-by: Daniel Schürmann <daniel@schuermann.dev>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/15179>
 2022-05-04 09:58:03 +00:00
Joshua Ashton ce7966fcb4 aco: Use movk for AddressHi bits in vertex prolog 
Eliminates a useless extra dword by transforming
        s_mov_b32 s47, 0xffff8000                                   ; beaf03ff ffff8000
to
        s_movk_i32 s47, 0x8000                                      ; b02f8000

which does the same thing.

Reviewed-By: Tatsuyuki Ishi <ishitatsuyuki@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/15415>
 2022-05-03 10:15:36 +00:00
Daniel Schürmann 58bd9a379e aco/ra: fix live-range splits of phi definitions 
It could happen that at the time of a live-range split,
a phi was not yet placed in the new instruction vector,
and thus, instead of renamed, a new phi was created.

Fixes: dEQP-VK.subgroups.ballot_broadcast.compute.subgroupbroadcast_i8vec2
Cc: mesa-stable

Reviewed-by: Rhys Perry <pendingchaos02@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/16248>
 2022-05-03 09:54:30 +00:00
Daniel Schürmann 12d7f911c9 aco/optimizer: prevent any overflow between SGPR and const offset on MUBUF 
Apparently, if the SGPR offset + const offset overflows,
it doesn't work.

Totals from 145 (0.11% of 134913) affected shaders: (GFX10.3)
SpillSGPRs: 134 -> 104 (-22.39%)
CodeSize: 1632676 -> 1645916 (+0.81%); split: -0.03%, +0.84%
Instrs: 316920 -> 320252 (+1.05%); split: -0.01%, +1.07%
Latency: 1456285 -> 1459686 (+0.23%); split: -0.02%, +0.25%
InvThroughput: 165785 -> 166086 (+0.18%); split: -0.02%, +0.20%
VClause: 6815 -> 6875 (+0.88%); split: -0.03%, +0.91%
SClause: 19089 -> 19079 (-0.05%); split: -0.06%, +0.01%
PreSGPRs: 7302 -> 7304 (+0.03%); split: -0.01%, +0.04%

Fixes: KHR-GL45.shader_storage_buffer_object.basic-operations-case1-cs
Reviewed-by: Rhys Perry <pendingchaos02@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/15866>
 2022-04-29 15:57:57 +00:00
Daniel Schürmann d5dc0c0392 aco: adjust num_waves for LDS before scheduling 
Totals from 67 (0.05% of 134913) affected shaders: (GFX10.3)
VGPRs: 2024 -> 2136 (+5.53%); split: -0.40%, +5.93%
CodeSize: 162364 -> 162348 (-0.01%); split: -0.08%, +0.07%
MaxWaves: 1882 -> 1816 (-3.51%); split: +0.11%, -3.61%
Instrs: 29176 -> 29162 (-0.05%); split: -0.09%, +0.04%
Latency: 329984 -> 327272 (-0.82%); split: -0.88%, +0.06%
InvThroughput: 54653 -> 54672 (+0.03%); split: -0.01%, +0.04%
VClause: 782 -> 761 (-2.69%); split: -2.81%, +0.13%
SClause: 833 -> 824 (-1.08%); split: -2.28%, +1.20%
Copies: 1872 -> 1873 (+0.05%); split: -0.37%, +0.43%

Reviewed-by: Rhys Perry <pendingchaos02@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/16039>
 2022-04-29 15:39:10 +00:00
Daniel Schürmann 8d8c59b4cd aco: split num_waves adjustment into separate function 
Reviewed-by: Rhys Perry <pendingchaos02@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/16039>
 2022-04-29 15:39:10 +00:00
Daniel Schürmann 6220046ad1 aco: remove 'max_waves' and use 'num_waves' to adjust for LDS and workgroup size 
Totals from 21 (0.02% of 134913) affected shaders: (GFX10.3)
VGPRs: 1024 -> 1176 (+14.84%)
CodeSize: 127824 -> 127664 (-0.13%); split: -0.17%, +0.04%
MaxWaves: 416 -> 378 (-9.13%)
Instrs: 22521 -> 22502 (-0.08%); split: -0.17%, +0.09%
Latency: 146386 -> 143154 (-2.21%); split: -2.21%, +0.00%
InvThroughput: 28379 -> 28944 (+1.99%); split: -0.23%, +2.22%
VClause: 575 -> 579 (+0.70%); split: -0.87%, +1.57%
SClause: 692 -> 645 (-6.79%)
Copies: 780 -> 747 (-4.23%); split: -4.74%, +0.51%

Reviewed-by: Rhys Perry <pendingchaos02@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/16039>
 2022-04-29 15:39:10 +00:00
Samuel Pitoiset 6873da0e42 aco: fix load_barycentric_at_{sample,offset} on GFX6-7 
The computation was wrong.

Fixes dEQP-GLES31.functional.shaders.multisample_interpolation.interpolate_at_offset.*
with Zink on GFX6 (Pitcairn).

Cc: 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/16099>
 2022-04-25 07:17:39 +00:00
Jason Ekstrand 1b8a43a0ba util: Remove util_cpu_detect 
util_cpu_detect is an anti-pattern: it relies on callers high up in the call
chain initializing a local implementation detail. As a real example, I added:

...a Mali compiler unit test
...that called bi_imm_f16() to construct an FP16 immediate
...that calls _mesa_float_to_half internally
...that calls util_get_cpu_caps internally, but only on x86_64!
...that relies on util_cpu_detect having been called before.

As a consequence, this unit test:

...crashes on x86_64 with USE_X86_64_ASM set
...passes on every other architecture
...works on my local arm64 workstation and on my test board
...failed CI which runs on x86_64
...needed to have a random util_cpu_detect() call sprinkled in.

This is a bad design decision. It pollutes the tree with magic, it causes
mysterious CI failures especially for non-x86_64 developers, and it is not
justified by a micro-optimization.

Instead, let's call util_cpu_detect directly from util_get_cpu_caps, avoiding
the footgun where it fails to be called.  This cleans up Mesa's design,
simplifies the tree, and avoids a class of a (possibly platform-specific)
failures. To mitigate the added overhead, wrap it all in a (fast) atomic
load check and declare the whole thing as ATTRIBUTE_CONST so the
compiler will CSE calls to util_cpu_detect.

Co-authored-by: Alyssa Rosenzweig <alyssa@collabora.com>
Reviewed-by: Marek Olšák <maraeo@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/15580>
 2022-04-20 18:44:35 +00:00
Georg Lehmann d12b5e7633 aco: Reuse previous -1 result in find_msb to avoid using VOP3. 
Totals:
CodeSize: 388934388 -> 388933712 (-0.00%)

Totals from 208 (0.15% of 134913) affected shaders:
CodeSize: 2008016 -> 2007340 (-0.03%)

Signed-off-by: Georg Lehmann <dadschoorse@gmail.com>
Reviewed-by: Rhys Perry <pendingchaos02@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/16011>
 2022-04-19 15:18:58 +00:00
Georg Lehmann a8b29094c2 aco: Remove some old comments in aco_opcodes.py. 
s_cmovk_i32 isn't GFX8_GFX9 only and s_version doesn't need a comment to say
it's GFX10+ exclusive. The encoding list is enough to provide this information,
as for other GFX10+ instructions.

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/16006>
 2022-04-18 15:59:38 +00:00
Rhys Perry 63e40adf8c aco: fix disassembly of SMEM with both SGPR and constant offset 
Signed-off-by: Rhys Perry <pendingchaos02@gmail.com>
Reviewed-by: Timur Kristóf <timur.kristof@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/15890>
 2022-04-14 20:58:36 +00:00
Rhys Perry c883abda76 aco: implement load_shared2_amd/store_shared2_amd 
Signed-off-by: Rhys Perry <pendingchaos02@gmail.com>
Reviewed-by: Timur Kristóf <timur.kristof@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/13778>
 2022-04-13 23:08:07 +00:00
Rhys Perry 5aa5af7776 aco: handle read2st64/write2st64 in optimizer 
Signed-off-by: Rhys Perry <pendingchaos02@gmail.com>
Reviewed-by: Timur Kristóf <timur.kristof@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/13778>
 2022-04-13 23:08:07 +00:00
Rhys Perry 2135c88d9c aco: fix signedness of DS_instruction::offset0/1 
Signed-off-by: Rhys Perry <pendingchaos02@gmail.com>
Reviewed-by: Timur Kristóf <timur.kristof@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/13778>
 2022-04-13 23:08:07 +00:00
Daniel Schürmann d703a0e808 aco: remove register hints entirely 
Reviewed-by: Rhys Perry <pendingchaos02@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/15408>
 2022-04-13 21:52:43 +00:00
Daniel Schürmann 2fe005a3fe aco: remove occurences of VCC hint 
Reviewed-by: Rhys Perry <pendingchaos02@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/15408>
 2022-04-13 21:52:43 +00:00
Daniel Schürmann b10c4d7dee aco: make program->needs_vcc independent of VCC hints 
Totals from 5 (0.00% of 135048) affected shaders: (GFX9)
SGPRs: 208 -> 160 (-23.08%)
CodeSize: 2700 -> 2692 (-0.30%)
Instrs: 533 -> 531 (-0.38%)
Latency: 41688 -> 41680 (-0.02%)

Reviewed-by: Rhys Perry <pendingchaos02@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/15408>
 2022-04-13 21:52:43 +00:00
Daniel Schürmann 415a3820fc aco/ra: omit VCC affinity on VOPC_SDWA for GFX9+ 
VOPC_SDWA can also use arbitrary SGPR pairs on GFX9+.

Totals from 5607 (4.16% of 134913) affected shaders: (GFX10.3)
CodeSize: 42470760 -> 42452988 (-0.04%)
Instrs: 7943174 -> 7942883 (-0.00%)
Latency: 102887029 -> 102886305 (-0.00%); split: -0.00%, +0.00%
InvThroughput: 20454456 -> 20454338 (-0.00%); split: -0.00%, +0.00%
Copies: 376818 -> 376865 (+0.01%); split: -0.00%, +0.01%

Reviewed-by: Rhys Perry <pendingchaos02@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/15408>
 2022-04-13 21:52:43 +00:00
Daniel Schürmann 6ebc61d71b aco/ra: create VCC-affinities during RA 
instead of using register hints.

Totals from 88367 (65.50% of 134913) affected shaders: (GFX10.3)
CodeSize: 322492184 -> 322252912 (-0.07%); split: -0.08%, +0.01%
Instrs: 60615809 -> 60541260 (-0.12%); split: -0.12%, +0.00%
Latency: 557067980 -> 557009210 (-0.01%); split: -0.01%, +0.00%
InvThroughput: 109676757 -> 109674804 (-0.00%); split: -0.00%, +0.00%
SClause: 1939703 -> 1939924 (+0.01%); split: -0.01%, +0.02%
Copies: 4557567 -> 4487530 (-1.54%); split: -1.54%, +0.00%
Branches: 1941123 -> 1937453 (-0.19%)

Reviewed-by: Rhys Perry <pendingchaos02@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/15408>
 2022-04-13 21:52:43 +00:00
Daniel Schürmann 44fb9ba84a aco/ra: only use VCC if program->needs_vcc == true 
A future commit will make VCC register assignment independent
from register hints. Up to GFX9, VCC can alternatively be used
as regular SGPR, so prevent overlap.

Reviewed-by: Rhys Perry <pendingchaos02@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/15408>
 2022-04-13 21:52:43 +00:00
Rhys Perry 7478b00c7c aco: remove old global access intrinsics 
No fossil-db changes.

Signed-off-by: Rhys Perry <pendingchaos02@gmail.com>
Reviewed-by: Timur Kristóf <timur.kristof@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/14124>
 2022-04-13 16:23:35 +00:00
Rhys Perry 9d1bab3615 aco: increase global_load_params.max_const_offset_plus_one 
The callback now supports this. This shouldn't have any effect yet except
on GFX6 with 12 byte loads.

fossil-db (Pitcairn):
Totals from 246 (0.18% of 135668) affected shaders:
VGPRs: 14684 -> 14768 (+0.57%); split: -0.44%, +1.01%
CodeSize: 1765792 -> 1738040 (-1.57%)
Instrs: 344605 -> 340055 (-1.32%)
Latency: 4892904 -> 4861942 (-0.63%)
InvThroughput: 2479599 -> 2446070 (-1.35%)
VClause: 8782 -> 8735 (-0.54%)
SClause: 9854 -> 9853 (-0.01%)
Copies: 47327 -> 45401 (-4.07%); split: -4.08%, +0.01%

Signed-off-by: Rhys Perry <pendingchaos02@gmail.com>
Reviewed-by: Timur Kristóf <timur.kristof@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/14124>
 2022-04-13 16:23:35 +00:00
Rhys Perry 3e9517c757 aco: implement _amd global access intrinsics 
fossil-db (Sienna Cichlid):
Totals from 7 (0.01% of 134621) affected shaders:
VGPRs: 760 -> 776 (+2.11%)
CodeSize: 222000 -> 222044 (+0.02%); split: -0.01%, +0.03%
Instrs: 40959 -> 40987 (+0.07%); split: -0.01%, +0.08%
Latency: 874811 -> 886609 (+1.35%); split: -0.00%, +1.35%
InvThroughput: 437405 -> 443303 (+1.35%); split: -0.00%, +1.35%
VClause: 1242 -> 1240 (-0.16%)
SClause: 1050 -> 1049 (-0.10%); split: -0.19%, +0.10%
Copies: 4953 -> 4973 (+0.40%); split: -0.04%, +0.44%
Branches: 1947 -> 1957 (+0.51%); split: -0.05%, +0.56%
PreVGPRs: 741 -> 747 (+0.81%)

fossil-db changes seem to be noise.

Signed-off-by: Rhys Perry <pendingchaos02@gmail.com>
Reviewed-by: Timur Kristóf <timur.kristof@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/14124>
 2022-04-13 16:23:35 +00:00
Rhys Perry 391bf3ea30 aco: don't expand smem/mubuf global loads 
For example, dwordx3->dwordx4 or ubyte3->dwordx2.

Global loads don't have the bounds checking that buffer loads have that
makes this safe.

The alignment checks are added to global_load_callback() in case
byte_align_loads=false, align=1 and bytes_needed=3. Without them, the
callback will create a dword load.

fossil-db (Sienna Cichlid):
Totals from 267 (0.20% of 134621) affected shaders:
CodeSize: 1603352 -> 1606568 (+0.20%)
Instrs: 294946 -> 295482 (+0.18%); split: -0.00%, +0.18%
Latency: 2997003 -> 2997052 (+0.00%); split: -0.02%, +0.02%
InvThroughput: 526645 -> 526659 (+0.00%)
SClause: 9179 -> 9185 (+0.07%); split: -0.02%, +0.09%
Copies: 25363 -> 25375 (+0.05%); split: -0.08%, +0.13%
Branches: 8298 -> 8299 (+0.01%)

fossil-db (Polaris10):
Totals from 267 (0.20% of 135668) affected shaders:
CodeSize: 1636672 -> 1638756 (+0.13%); split: -0.00%, +0.13%
Instrs: 308484 -> 308733 (+0.08%); split: -0.01%, +0.09%
Latency: 3446045 -> 3446904 (+0.02%); split: -0.00%, +0.03%
InvThroughput: 1206722 -> 1206828 (+0.01%); split: -0.00%, +0.01%
SClause: 9308 -> 9311 (+0.03%); split: -0.08%, +0.11%
Copies: 36933 -> 36921 (-0.03%); split: -0.08%, +0.05%

fossil-db (Pitcairn):
Totals from 275 (0.20% of 135668) affected shaders:
SGPRs: 17616 -> 17520 (-0.54%); split: -0.64%, +0.09%
VGPRs: 15428 -> 15540 (+0.73%); split: -0.23%, +0.96%
CodeSize: 1885792 -> 1929120 (+2.30%); split: -0.00%, +2.30%
MaxWaves: 1284 -> 1285 (+0.08%)
Instrs: 368963 -> 376095 (+1.93%); split: -0.00%, +1.94%
Latency: 5122922 -> 5168398 (+0.89%); split: -0.01%, +0.90%
InvThroughput: 2562866 -> 2604279 (+1.62%)
VClause: 9268 -> 9296 (+0.30%); split: -0.13%, +0.43%
SClause: 10702 -> 10705 (+0.03%); split: -0.05%, +0.07%
Copies: 48620 -> 50629 (+4.13%); split: -0.08%, +4.21%

Signed-off-by: Rhys Perry <pendingchaos02@gmail.com>
Reviewed-by: Timur Kristóf <timur.kristof@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/14124>
 2022-04-13 16:23:35 +00:00
Rhys Perry 6baad09711 aco: use saddr for global access with sgpr address 
fossil-db (Sienna Cichlid):
Totals from 38 (0.03% of 134621) affected shaders:
CodeSize: 237196 -> 237060 (-0.06%); split: -0.09%, +0.03%
Instrs: 43895 -> 43894 (-0.00%); split: -0.02%, +0.01%
Latency: 914633 -> 916263 (+0.18%); split: -0.01%, +0.19%
InvThroughput: 468215 -> 468971 (+0.16%); split: -0.02%, +0.18%
SClause: 1239 -> 1242 (+0.24%)
PreSGPRs: 997 -> 1003 (+0.60%)
PreVGPRs: 936 -> 923 (-1.39%); split: -1.50%, +0.11%

Regression seems to be RA noise, creating a waitcnt.

Signed-off-by: Rhys Perry <pendingchaos02@gmail.com>
Reviewed-by: Timur Kristóf <timur.kristof@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/14124>
 2022-04-13 16:23:35 +00:00
Rhys Perry d957730b9b aco: use vcc for 64-bit vgpr addition 
fossil-db (Sienna Cichlid):
Totals from 229 (0.17% of 134621) affected shaders:
CodeSize: 1520192 -> 1517644 (-0.17%)

Signed-off-by: Rhys Perry <pendingchaos02@gmail.com>
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/14124>
 2022-04-13 16:23:35 +00:00
Timur Kristóf b874a2ed61 aco: Fix VOP2 instruction format in visit_tex. 
There was a v_or_b32 that accidentally used SOP2.
It should use VOP2.

Issue found by looking at a gfxreconstruct trace posted by a user
in this bug: https://gitlab.freedesktop.org/mesa/mesa/-/issues/5838

Cc: mesa-stable
Fixes: 93c8ebfa78 "aco: Initial commit of independent AMD compiler"

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/15923>
 2022-04-13 13:06:49 +00:00
Rhys Perry 773c7cbcbc radv,aco: implement 64-bit inline push constants 
fossil-db (Sienna Cichlid):
Totals from 21 (0.02% of 134621) affected shaders:
CodeSize: 1932 -> 1560 (-19.25%)
Instrs: 357 -> 303 (-15.13%)
Latency: 6576 -> 5883 (-10.54%)
InvThroughput: 26304 -> 23532 (-10.54%)
SClause: 42 -> 24 (-42.86%)
Copies: 90 -> 105 (+16.67%); split: -10.00%, +26.67%
PreSGPRs: 144 -> 201 (+39.58%)

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/12145>
 2022-04-12 11:44:30 +00:00
Rhys Perry 7f6262bb85 radv: allow holes in inline push constants 
Use a dword mask instead of a range to track which push constants to
inline.

fossil-db (Sienna Cichlid):
Totals from 5724 (4.25% of 134621) affected shaders:
CodeSize: 20894044 -> 20815748 (-0.37%); split: -0.39%, +0.02%
Instrs: 4002568 -> 3988385 (-0.35%); split: -0.38%, +0.02%
Latency: 29285060 -> 29224414 (-0.21%); split: -0.22%, +0.01%
InvThroughput: 5529700 -> 5526893 (-0.05%); split: -0.05%, +0.00%
VClause: 78093 -> 78240 (+0.19%); split: -0.23%, +0.41%
SClause: 135495 -> 131027 (-3.30%); split: -3.30%, +0.00%
Copies: 330856 -> 324552 (-1.91%); split: -2.37%, +0.46%
PreSGPRs: 226031 -> 224778 (-0.55%); split: -0.61%, +0.05%

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/12145>
 2022-04-12 11:44:30 +00:00
Erik Faye-Lund 28dbabec8e aco: do not use designated initializers 
Designated initializers are a C++20 feature, but we don't use C++20. In
fact, enabling C++20 for ACO triggers new compiler errors due to some
equality semantics details.

So let's instead stop using designated initializers here.

Reviewed-by: Daniel Schürmann <daniel@schuermann.dev>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/15706>
 2022-04-05 16:58:56 +00:00
Rhys Perry 5b4e41e4db aco: don't use v_mad_mix on GFX9 if 16-bit denormals must be preserved 
This probably effectively disables the v_mad_mix optimization on GFX9.

fossil-db (Vega):
Totals from 11545 (7.15% of 161366) affected shaders:
MaxWaves: 43025 -> 42780 (-0.57%); split: +0.06%, -0.63%
Instrs: 18571635 -> 18734201 (+0.88%); split: -0.00%, +0.88%
CodeSize: 96483568 -> 96611012 (+0.13%); split: -0.11%, +0.24%
SGPRs: 1079056 -> 1077616 (-0.13%); split: -0.14%, +0.01%
VGPRs: 819248 -> 821868 (+0.32%); split: -0.04%, +0.36%
SpillSGPRs: 13313 -> 12464 (-6.38%)
Latency: 293804093 -> 295046122 (+0.42%); split: -0.09%, +0.51%
InvThroughput: 110002239 -> 110994978 (+0.90%); split: -0.03%, +0.93%
VClause: 342458 -> 342596 (+0.04%); split: -0.12%, +0.16%
SClause: 648566 -> 648046 (-0.08%); split: -0.12%, +0.04%
Copies: 1728225 -> 1726679 (-0.09%); split: -0.66%, +0.57%
Branches: 552973 -> 552963 (-0.00%); split: -0.02%, +0.02%
PreSGPRs: 862360 -> 856820 (-0.64%); split: -0.69%, +0.05%
PreVGPRs: 773689 -> 776818 (+0.40%); split: -0.02%, +0.42%

Signed-off-by: Rhys Perry <pendingchaos02@gmail.com>
Reviewed-by: Timur Kristóf <timur.kristof@gmail.com>
Closes: https://gitlab.freedesktop.org/mesa/mesa/-/issues/6178
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/15718>
 2022-04-04 19:27:12 +00:00
Daniel Schürmann 6d383159d4 aco/optimizer: check recursively if we can eliminate s_and exec 
Totals from 2860 (2.12% of 134913) affected shaders: (GFX10.3)
CodeSize: 5990728 -> 5979164 (-0.19%); split: -0.20%, +0.01%
Instrs: 1094562 -> 1091653 (-0.27%); split: -0.28%, +0.01%
Latency: 8689841 -> 8684523 (-0.06%); split: -0.07%, +0.00%
InvThroughput: 1840533 -> 1840527 (-0.00%); split: -0.00%, +0.00%
SClause: 51437 -> 51439 (+0.00%)
Copies: 82461 -> 82472 (+0.01%)
PreSGPRs: 83136 -> 83172 (+0.04%)

Reviewed-by: Timur Kristóf <timur.kristof@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/15675>
 2022-04-01 15:35:26 +02:00
Daniel Schürmann db8c401f71 aco/ra: fix stride check on subdword parallelcopies for create_vector 
On GFX6/7, info.rc is in full dwords.

Fixes: 9476986e6f ('aco/ra: special-case get_reg_for_create_vector_copy()')
Reviewed-by: Rhys Perry <pendingchaos02@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/15393>
 2022-03-31 08:55:07 +00:00
Samuel Pitoiset d32656bc65 radv: lower has_multiview_view_index in NIR 
This lowering is done in a new NIR pass where the layer is written
before emit_vertex_with_counter for geometry shaders and after the
position for other vertex stages.

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/15456>
 2022-03-31 07:34:21 +00:00
Georg Lehmann 141ca78634 radv, aco: Packed iadd_sat/uadd_sat. 
Signed-off-by: Georg Lehmann <dadschoorse@gmail.com>
Reviewed-by: Daniel Schürmann <daniel@schuermann.dev>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/15421>
 2022-03-28 20:02:52 +00:00
Georg Lehmann 50f585254c aco: Implement scalar iadd_sat. 
Signed-off-by: Georg Lehmann <dadschoorse@gmail.com>
Reviewed-by: Daniel Schürmann <daniel@schuermann.dev>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/15421>
 2022-03-28 20:02:52 +00:00
Georg Lehmann 989f2b1785 aco: Implement 64bit uadd_sat. 
Signed-off-by: Georg Lehmann <dadschoorse@gmail.com>
Reviewed-by: Daniel Schürmann <daniel@schuermann.dev>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/15421>
 2022-03-28 20:02:52 +00:00
Georg Lehmann 01fad6aa72 aco: Remove 0 data components from image stores. 
Image stores always write a full texel.
The hardware writes zero for components not included in dmask.

Totals from 387 (0.29% of 134913) affected shaders: (GFX10.3)
VGPRs: 17216 -> 17136 (-0.46%)
CodeSize: 1987652 -> 1981504 (-0.31%)
MaxWaves: 9054 -> 9058 (+0.04%)
Instrs: 361883 -> 361115 (-0.21%); split: -0.21%, +0.00%
Latency: 5383187 -> 5381227 (-0.04%); split: -0.04%, +0.00%
InvThroughput: 1373830 -> 1372097 (-0.13%)
VClause: 9031 -> 9038 (+0.08%); split: -0.04%, +0.12%
Copies: 25923 -> 25153 (-2.97%); split: -2.98%, +0.01%
PreVGPRs: 14643 -> 14555 (-0.60%)

Signed-off-by: Georg Lehmann <dadschoorse@gmail.com>
Reviewed-by: Daniel Schürmann <daniel@schuermann.dev>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/14986>
 2022-03-28 19:43:18 +00:00
Samuel Pitoiset 4cfb5332d6 radv: lower adjusting gl_FragCoord.z for VRS in NIR 
fossils-db (Sienna Cichlid):
Totals from 4432 (3.29% of 134913) affected shaders:
VGPRs: 231232 -> 231880 (+0.28%)
CodeSize: 24738224 -> 24718008 (-0.08%); split: -0.08%, +0.00%
MaxWaves: 93120 -> 93000 (-0.13%)
Instrs: 4540970 -> 4541062 (+0.00%); split: -0.01%, +0.01%
Latency: 49658353 -> 49641444 (-0.03%); split: -0.05%, +0.01%
InvThroughput: 9604328 -> 9603041 (-0.01%); split: -0.02%, +0.01%
VClause: 66497 -> 66498 (+0.00%)
SClause: 209530 -> 209532 (+0.00%); split: -0.01%, +0.01%
Copies: 276135 -> 276249 (+0.04%); split: -0.14%, +0.18%
PreSGPRs: 189409 -> 189415 (+0.00%)
PreVGPRs: 207368 -> 207458 (+0.04%)

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/15450>
 2022-03-28 16:17:37 +02:00
Samuel Pitoiset a42b6a4d39 radv: lower load_sample_mask_in in NIR 
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/15450>
 2022-03-28 16:17:34 +02:00
Samuel Pitoiset 8c51874af4 radv,aco: lower color exports in NIR 
fossils-db (Sienna Cichlid):
Totals from 27108 (20.09% of 134913) affected shaders:
VGPRs: 1260608 -> 1261424 (+0.06%); split: -0.00%, +0.07%
CodeSize: 112795868 -> 112785892 (-0.01%); split: -0.05%, +0.04%
MaxWaves: 628608 -> 628448 (-0.03%); split: +0.00%, -0.03%
Instrs: 20750003 -> 20749314 (-0.00%); split: -0.01%, +0.00%
Latency: 288088081 -> 288015865 (-0.03%); split: -0.06%, +0.04%
InvThroughput: 53944847 -> 53961693 (+0.03%); split: -0.01%, +0.04%
VClause: 396463 -> 396467 (+0.00%); split: -0.02%, +0.02%
SClause: 842088 -> 842150 (+0.01%); split: -0.03%, +0.04%
Copies: 1244982 -> 1259026 (+1.13%); split: -0.01%, +1.14%
PreSGPRs: 1251949 -> 1251909 (-0.00%)
PreVGPRs: 1099647 -> 1100879 (+0.11%); split: -0.03%, +0.14%

fossils-db (Polaris10):
Totals from 23928 (17.60% of 135960) affected shaders:
SGPRs: 1751792 -> 1751024 (-0.04%); split: -0.05%, +0.01%
VGPRs: 1098964 -> 1098556 (-0.04%); split: -0.13%, +0.09%
CodeSize: 99893472 -> 99837940 (-0.06%); split: -0.06%, +0.00%
MaxWaves: 138322 -> 138306 (-0.01%); split: +0.03%, -0.04%
Instrs: 19213995 -> 19211980 (-0.01%); split: -0.02%, +0.01%
Latency: 273026926 -> 273109402 (+0.03%); split: -0.01%, +0.04%
InvThroughput: 111160907 -> 111195187 (+0.03%); split: -0.04%, +0.07%
VClause: 343058 -> 343097 (+0.01%); split: -0.02%, +0.03%
SClause: 802756 -> 802884 (+0.02%); split: -0.04%, +0.06%
Copies: 1729387 -> 1739208 (+0.57%); split: -0.04%, +0.61%
PreSGPRs: 1090264 -> 1090303 (+0.00%); split: -0.00%, +0.01%
PreVGPRs: 959490 -> 960600 (+0.12%); split: -0.04%, +0.15%

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/15263>
 2022-03-28 11:47:35 +00:00
Rhys Perry 1ead285d92 aco: fix RA validation of 16-bit fma_mix operands 
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/15562>
 2022-03-28 11:05:25 +01:00
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 2d1e6b756e aco: remove 'high' parameter from can_use_opsel() 
No fossil-db changes.

Reviewed-by: Rhys Perry <pendingchaos02@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/15551>
 2022-03-25 22:02:50 +00:00
Daniel Schürmann b98a9dcc36 aco/optimizer: fix call to can_use_opsel() in apply_insert() 
The definition index is -1.

Fixes: 54292e99c7 ('aco: optimize 32-bit extracts and inserts using SDWA ')
Reviewed-by: Rhys Perry <pendingchaos02@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/15551>
 2022-03-25 22:02:50 +00:00
Boris Brezillon de039537da aco: Fix an MSVC warning 
'warning C4804: '<<': unsafe use of type 'bool' in operation'

Fixes: 9934c86761 ("aco: use v_fma_mix to combine mul/add/fma input conversions")
Acked-by: Daniel Stone <daniels@collabora.com>
Acked-by: Erik Faye-Lund <erik.faye-lund@collabora.com>
Acked-by: Jason Ekstrand <jason.ekstrand@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/15518>
 2022-03-24 09:11:13 +00:00
Rhys Perry 543a5487cf radv,aco: lower image descriptor loads in NIR 
fossil-db (Sienna Cichlid):
Totals from 2926 (1.80% of 162293) affected shaders:
Instrs: 2315110 -> 2306644 (-0.37%); split: -0.37%, +0.00%
CodeSize: 12581592 -> 12546588 (-0.28%); split: -0.28%, +0.00%
VGPRs: 130216 -> 130208 (-0.01%)
SpillSGPRs: 477 -> 474 (-0.63%); split: -5.03%, +4.40%
Latency: 29686188 -> 29678804 (-0.02%); split: -0.05%, +0.02%
InvThroughput: 6926545 -> 6926286 (-0.00%); split: -0.02%, +0.02%
SClause: 73761 -> 72996 (-1.04%); split: -1.16%, +0.12%
Copies: 144068 -> 137279 (-4.71%); split: -4.78%, +0.07%
Branches: 47466 -> 47483 (+0.04%); split: -0.01%, +0.04%
PreSGPRs: 118042 -> 117377 (-0.56%); split: -1.34%, +0.77%

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