fdo-mirrors/mesa
1
0
Fork 0
mirror of https://gitlab.freedesktop.org/mesa/mesa.git synced 2026-03-19 03:30:42 +01:00
Code Issues Projects Releases Packages Wiki Activity Actions 4
198121 commits 64 branches 926 tags 1.1 GiB
Commit graph

3438 commits

Author SHA1 Message Date
Daniel Schürmann
bb87832ce0 aco/insert_NOPs: add early exit to handle_valu_partial_forwarding_hazard_instr 
No need to continue if there was already a hazard found in
a different control flow path.

Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/32191>
 2024-11-22 08:46:32 +00:00
Daniel Schürmann
07df37ba01 aco/insert_NOPs: use RegCounterMap as replacement for the CounterMap implementation 
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/32191>
 2024-11-22 08:46:32 +00:00
Daniel Schürmann
fb5e5adfb3 aco/insert_NOPs: implement vector-based RegCounterMap as replacement for VGPRCounterMap 
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/32191>
 2024-11-22 08:46:32 +00:00
Georg Lehmann
f3926c9d4e aco/isel: use undef Operands for p_create_vector created from nir vecs 
Foz-DB Navi31:
Totals from 27464 (34.59% of 79395) affected shaders:
Instrs: 9595601 -> 9535260 (-0.63%); split: -0.63%, +0.00%
CodeSize: 47900112 -> 47658648 (-0.50%); split: -0.50%, +0.00%
Latency: 43928471 -> 43918448 (-0.02%); split: -0.05%, +0.02%
InvThroughput: 4940105 -> 4903447 (-0.74%); split: -0.75%, +0.01%
Copies: 667294 -> 604603 (-9.39%); split: -9.39%, +0.00%
VALU: 5282264 -> 5219604 (-1.19%); split: -1.19%, +0.00%
VOPD: 342 -> 311 (-9.06%)

Reviewed-by: Rhys Perry <pendingchaos02@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/32249>
 2024-11-21 14:09:52 +00:00
Daniel Schürmann
1ff9a0fe80 aco: remove Pseudo_instruction::tmp_in_scc 
This information is redundant, now.

No fossil-changes.

Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/32217>
 2024-11-20 11:04:32 +00:00
Daniel Schürmann
a1a4a6061c aco/ra: explicitly assign scratch SGPR for linear phis 
We are about to remove the branch definitions which previously
served this purpose. Also remove Block::scc_live_out.
Some changes due to round-robin RA.

Totals from 939 (1.18% of 79395) affected shaders: (Navi31)

Instrs: 5038786 -> 5038611 (-0.00%); split: -0.01%, +0.00%
CodeSize: 26153412 -> 26152904 (-0.00%); split: -0.00%, +0.00%
Latency: 41649989 -> 41650120 (+0.00%); split: -0.00%, +0.00%
InvThroughput: 6447508 -> 6447536 (+0.00%); split: -0.00%, +0.00%
SClause: 131319 -> 131276 (-0.03%); split: -0.03%, +0.00%
Copies: 359362 -> 359256 (-0.03%); split: -0.05%, +0.02%
SALU: 639275 -> 639169 (-0.02%); split: -0.03%, +0.01%
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/32217>
 2024-11-20 11:04:32 +00:00
Daniel Schürmann
17da551133 aco/ra: use bitset for sgpr_operands_alias_defs 
We cannot rely on SGPR Temps being fully aligned to 64 SGPRs.

Cc: mesa-stable
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/32217>
 2024-11-20 11:04:32 +00:00
Daniel Schürmann
a04e096339 aco/ra: set Pseudo_instruction::scratch_sgpr to SCC if it doesn't need to be preserved 
Also ensure that 'needs_scratch_reg' is always true if SCC might be overwritten.
Few changes, because some p_split_vector get SCC as scratch reg assigned,
and thus, can inhibit some postRA optimizations.

Totals from 3 (0.00% of 79395) affected shaders: (Navi31)
Instrs: 10501 -> 10500 (-0.01%); split: -0.02%, +0.01%
CodeSize: 51580 -> 51520 (-0.12%); split: -0.12%, +0.01%
Latency: 84166 -> 84174 (+0.01%)
InvThroughput: 13109 -> 13111 (+0.02%)
SALU: 859 -> 860 (+0.12%)

Cc: mesa-stable
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/32217>
 2024-11-20 11:04:32 +00:00
Samuel Pitoiset
2abbd361e2 radv,aco: dump LDS from the trap handler 
Can be useful for debugging.

Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/32182>
 2024-11-20 07:35:47 +00:00
Timur Kristóf
d94e65574f aco: Separate options for printing IR and recording disassembly. 
Signed-off-by: Timur Kristóf <timur.kristof@gmail.com>
Reviewed-by: Konstantin Seurer <konstantin.seurer@gmail.com>
Reviewed-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Acked-by: Daniel Schürmann <daniel@schuermann.dev>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/32079>
 2024-11-19 16:15:28 +01:00
Marek Olšák
b258a9aa4e aco: remove unused TCS fields from aco_shader_info 
Reviewed-by: Timur Kristóf <timur.kristof@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/32171>
 2024-11-16 21:58:26 -05:00
Georg Lehmann
3e037ac2a9 aco/gfx8: use ds_swizzle_b32 rotate mode 
Despite only being mentioned in the ISA docs since vega, rotate (and fft)
swizzle mode seem to exist since gfx8.

https://github.com/llvm/llvm-project/issues/28975#issuecomment-980964939

Reviewed-by: Timur Kristóf <timur.kristof@gmail.com>
Reviewed-by: Daniel Schürmann <daniel@schuermann.dev>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/31348>
 2024-11-14 15:34:48 +00:00
Samuel Pitoiset
b4b5f9eeb0 radv,aco: dump VGPRS from the trap handler shader 
Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/32090>
 2024-11-13 15:27:54 +00:00
Samuel Pitoiset
132b7a85c7 aco: drop the second M0 operand for s_set_gpr_idx_on 
Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/32090>
 2024-11-13 15:27:54 +00:00
Samuel Pitoiset
c712555a9f aco: save/restore VGPRS on GFX8 in the trap handler shader 
This will be needed for dumping VGPRs.

Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/32090>
 2024-11-13 15:27:54 +00:00
Samuel Pitoiset
a77af57e83 aco: use all invocations from the current wave in the trap handler 
Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/32090>
 2024-11-13 15:27:54 +00:00
Samuel Pitoiset
034014a165 aco: restore m0/exec before exiting the trap handler 
Dumping VGPRs will overwrite m0 and exec and they need to be restored
if we want to return to the shader.

Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/32090>
 2024-11-13 15:27:53 +00:00
Samuel Pitoiset
185a165a85 aco: fix validation for v_movrels_b32 and friends 
m0 is the second operand.

Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/32090>
 2024-11-13 15:27:53 +00:00
Samuel Pitoiset
40b343bbee aco: add a new variant for vop1() with two operands 
For v_movrels_b32 and friends which need a second operand for m0.

Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/32090>
 2024-11-13 15:27:53 +00:00
Samuel Pitoiset
f4cf6a71ed aco: use a 64-bit mov to save exec in the trap handler shader 
Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/32090>
 2024-11-13 15:27:53 +00:00
Rhys Perry
17cc8a5a54 aco: remove load byte_align 
8/16-bit loads given to instruction selection now always use VMEM and
scalar load instructions unless alignment easily allows a vector load.

Signed-off-by: Rhys Perry <pendingchaos02@gmail.com>
Reviewed-by: Georg Lehmann <dadschoorse@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/31904>
 2024-11-13 12:59:26 +00:00
Rhys Perry
d3ae1842a2 aco,ac/nir: flag loads to use smem in NIR 
This pass will be re-used later.

Signed-off-by: Rhys Perry <pendingchaos02@gmail.com>
Reviewed-by: Georg Lehmann <dadschoorse@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/31904>
 2024-11-13 12:59:26 +00:00
Rhys Perry
0619e4db63 nir,aco,ac/llvm: add nir_op_alignbyte_amd 
Signed-off-by: Rhys Perry <pendingchaos02@gmail.com>
Reviewed-by: Georg Lehmann <dadschoorse@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/31904>
 2024-11-13 12:59:26 +00:00
Rhys Perry
db0cbb7e9b aco: optimize nir_op_shfr with <32 src1 
No fossil-db changes.

Signed-off-by: Rhys Perry <pendingchaos02@gmail.com>
Reviewed-by: Georg Lehmann <dadschoorse@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/31904>
 2024-11-13 12:59:26 +00:00
Samuel Pitoiset
0c77469995 aco: fix saving/restoring VGPRS in the trap handler on GFX9 
When ADD_TID_ENABLE=1, DATA_FORMAT is STRIDE[14:17], so the stride
was too large.

Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/32109>
 2024-11-13 11:12:54 +00:00
Georg Lehmann
7e8a08ae77 aco: use nir_def_all_uses_ignore_sign_bit 
Reviewed-by: Alyssa Rosenzweig <alyssa@rosenzweig.io>
Reviewed-by: Daniel Schürmann <daniel@schuermann.dev>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/31844>
 2024-11-12 18:03:57 +00:00
Samuel Pitoiset
2d5df46c25 aco: emit nir_intrinsic_debug_break 
s_trap is used to enter the trap.

Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/32061>
 2024-11-12 16:05:17 +00:00
Samuel Pitoiset
5f79b8ea2d radv,aco: save/restore overwritten VGPRs in the trap handler shader 
The trap currently doesn't return to the shader but it will be needed
for example for the debug mode.

Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/32056>
 2024-11-12 11:16:13 +00:00
Samuel Pitoiset
6ec0c85908 radv,aco: use the trap handler layout struct while compiling the shader 
It's less error prone to rely on the layout for offsets.

Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/32056>
 2024-11-12 11:16:13 +00:00
Samuel Pitoiset
6bfd92123f aco: simplify postprocessing the trap handler shader 
Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/32056>
 2024-11-12 11:16:13 +00:00
Samuel Pitoiset
44dfeb4479 radv,aco: add a separate function to compile the trap handler shader 
Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/32056>
 2024-11-12 11:16:13 +00:00
Samuel Pitoiset
62e335c779 radv,aco: dump more SQ_WAVE regs from the trap handler 
Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/32056>
 2024-11-12 11:16:13 +00:00
Samuel Pitoiset
107f29c39a aco: do not reorder s_trap instructions 
Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/32055>
 2024-11-11 15:46:36 +00:00
Konstantin Seurer
69ebba82d4 aco: Pass debug information to the driver 
Reviewed-by: Timur Kristóf <timur.kristof@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/29298>
 2024-11-11 08:39:13 +00:00
Konstantin Seurer
f8ef1afec8 aco: Handle nir_debug_info_instr 
Propagated debug info using p_debug_info and Program::debug_info.
Offsets into the shader binary are gathered during assembly.
This will be usefull for mapping back the disassembled shader to
nir, glsl or spirv.

Reviewed-by: Timur Kristóf <timur.kristof@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/29298>
 2024-11-11 08:39:13 +00:00
Samuel Pitoiset
437bd63265 radv,aco: dump m0 and exec from the trap handler 
Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/32026>
 2024-11-08 14:00:15 +00:00
Samuel Pitoiset
d1d41be43f aco: declare phys regs for tba_hi/tma_hi 
Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/32026>
 2024-11-08 14:00:15 +00:00
Samuel Pitoiset
13bab450a2 aco: fix storing SQ_WAVE_STATUS in the trap handler shader 
SQ_WAVE_STATUS can change inside the trap because of SCC.

Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/32026>
 2024-11-08 14:00:14 +00:00
Samuel Pitoiset
494050d2ea aco: add a helper to dump SGPR to memory for the trap handler 
Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/32026>
 2024-11-08 14:00:14 +00:00
Samuel Pitoiset
8c6f2fef1b aco: use scalar buffer stores for dumping SGPRS from the trap on GFX8 
This avoids using any VGPRs on GFX8.

Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/32026>
 2024-11-08 14:00:14 +00:00
Samuel Pitoiset
17f6b4e51e aco: save/restore SCC in the trap handler shader 
SCC is only updated on GFX9+ but let's do it by default because the
trap handler shader is likely going to be more and more complex over
time.

Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/32026>
 2024-11-08 14:00:14 +00:00
Samuel Pitoiset
7b4386facd aco: cleanup using fixed registers in the trap handler shader 
It's easier to read and potentially less error prone.

Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/32026>
 2024-11-08 14:00:14 +00:00
Rhys Perry
215c44c124 aco: apply extract to v_cvt_f32_ubyte0 
No fossil-db changes.

Signed-off-by: Rhys Perry <pendingchaos02@gmail.com>
Reviewed-by: Georg Lehmann <dadschoorse@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/31762>
 2024-11-06 19:31:20 +00:00
Rhys Perry
f1a932bc29 aco: apply extract to p_extract_vector 
fossil-db (navi21):
Totals from 46 (0.06% of 79395) affected shaders:
Instrs: 80126 -> 79944 (-0.23%); split: -0.27%, +0.04%
CodeSize: 486860 -> 485668 (-0.24%); split: -0.31%, +0.06%
Latency: 1615395 -> 1614218 (-0.07%); split: -0.07%, +0.00%
InvThroughput: 705479 -> 705013 (-0.07%); split: -0.07%, +0.00%
Copies: 18934 -> 18797 (-0.72%); split: -0.98%, +0.25%
VALU: 52452 -> 52268 (-0.35%); split: -0.41%, +0.06%
SALU: 17253 -> 17255 (+0.01%); split: -0.02%, +0.03%

Signed-off-by: Rhys Perry <pendingchaos02@gmail.com>
Reviewed-by: Georg Lehmann <dadschoorse@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/31762>
 2024-11-06 19:31:20 +00:00
Rhys Perry
6cb9d39bc2 aco: combine extracts with sub-dword definitions 
fossil-db (navi21):
Totals from 23 (0.03% of 79395) affected shaders:
Instrs: 55133 -> 55099 (-0.06%)
CodeSize: 335744 -> 335512 (-0.07%)
Latency: 1709146 -> 1709031 (-0.01%)
InvThroughput: 613788 -> 613713 (-0.01%)
Copies: 14405 -> 14407 (+0.01%); split: -0.03%, +0.04%
VALU: 37038 -> 37000 (-0.10%)
SALU: 11125 -> 11131 (+0.05%)

Signed-off-by: Rhys Perry <pendingchaos02@gmail.com>
Reviewed-by: Georg Lehmann <dadschoorse@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/31762>
 2024-11-06 19:31:20 +00:00
Rhys Perry
30af7ae44f aco: add and use apply_extract_twice helper 
This will be used in the next commit.

Signed-off-by: Rhys Perry <pendingchaos02@gmail.com>
Reviewed-by: Georg Lehmann <dadschoorse@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/31762>
 2024-11-06 19:31:20 +00:00
Rhys Perry
05d0fa894e aco: allow applying sign-extended sel to p_extract more often 
In the case of v1=p_extract(v1=p_extract(src, 0, 16, 1), 0, 32, 0).
When we apply extracts with sub-dword definitions, this will also
include v2b=p_extract(v2b=p_extract(src, 0, 8, 1), 0, 16, 0).

No fossil-db changes.

Signed-off-by: Rhys Perry <pendingchaos02@gmail.com>
Reviewed-by: Georg Lehmann <dadschoorse@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/31762>
 2024-11-06 19:31:20 +00:00
Rhys Perry
e47bc3e750 aco: shrink code size of some p_extract 
fossil-db (navi21):
Totals from 37 (0.05% of 79395) affected shaders:
CodeSize: 2048204 -> 2047836 (-0.02%)

fossil-db (navi31):
Totals from 307 (0.39% of 79395) affected shaders:
CodeSize: 3075732 -> 3065236 (-0.34%); split: -0.34%, +0.00%

Signed-off-by: Rhys Perry <pendingchaos02@gmail.com>
Reviewed-by: Georg Lehmann <dadschoorse@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/31762>
 2024-11-06 19:31:20 +00:00
Rhys Perry
d285333800 aco: add a bit more p_extract/p_insert validation 
Signed-off-by: Rhys Perry <pendingchaos02@gmail.com>
Reviewed-by: Georg Lehmann <dadschoorse@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/31762>
 2024-11-06 19:31:20 +00:00
Rhys Perry
d3ac69f79b aco: handle SGPR limitations when applying extract 
We were already doing this, but missing it in a few places.

Signed-off-by: Rhys Perry <pendingchaos02@gmail.com>
Reviewed-by: Georg Lehmann <dadschoorse@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/31762>
 2024-11-06 19:31:20 +00:00