Commit Graph

5390 Commits

Author SHA1 Message Date
Rhys Perry 365d0aa6c5 aco: shrink mad_info
From 24 bytes to 16 bytes.

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/5281>
2020-06-18 14:26:01 +00:00
Rhys Perry 917260710f aco: make ssa_info::label 64-bit
We'll probably need these extra bits in the future.

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/5281>
2020-06-18 14:26:01 +00:00
Rhys Perry 47ca84a96d aco: shrink ssa_info
Reorder members so that it's 16 bytes instead of 24.

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/5281>
2020-06-18 14:26:01 +00:00
Daniel Schürmann 3817fa7a4d aco: fix WQM handling in nested loops
If on a nested loop
- the outer loop needs WQM but
- the inner loop doesn't need WQM and
- the break condition of the inner loop is computed in the outer loop
then it could happen that we transitioned to Exact before entering the inner loop
which could create an empty exec mask and lead to an infinite loop.

Fixes a GPU hang with RDR2

Cc: 20.1 <mesa-stable@lists.freedesktop.org>
Reviewed-by: Rhys Perry <pendingchaos02@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/5518>
2020-06-18 13:40:15 +00:00
Samuel Pitoiset fa149b996d radv: only requires LLVM 9 for GFX10 if not using ACO
In case someone links RADV with LLVM 8 and wants to use ACO.

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/5454>
2020-06-18 09:45:20 +00:00
Samuel Pitoiset 70cc80805c radv: compute CB_SHADER_MASK from the fragment shader outputs
The fragment shader doesn't necessarily output the number of components
expected by the target format.

Fixes new dEQP-VK.draw.output_location.*.

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/5427>
2020-06-18 09:16:04 +00:00
Samuel Pitoiset b848d88059 radv: make sure to set CB_SHADER_MASK correctly for internal CB operations
It should be always set to 0xf.

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/5427>
2020-06-18 09:16:04 +00:00
Samuel Pitoiset c4aa64b4c3 radv: lower discards to demote to workaround a RDR2 game bug
This fixes some sort of LOD issue.

Cc: 20.1 <mesa-stable@lists.freedesktop.org>
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/5525>
2020-06-18 08:39:28 +02:00
Eric Engestrom 405bffefe1 radv: replace all dup() with os_dupfd_cloexec()
Signed-off-by: Eric Engestrom <eric@engestrom.ch>
Reviewed-by: Kristian H. Kristensen <hoegsberg@google.com>
Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/5369>
2020-06-18 02:09:56 +00:00
Samuel Pitoiset 51fb3b09dc radv/aco: enable FP16 features/extensions on GFX9+
This enables shaderFloat16, VK_AMD_gpu_shader_half_float and
VK_AMD_gpu_shader_int16.

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/5347>
2020-06-17 18:12:51 +02:00
Rhys Perry abfe28a6bb aco: fix validation of opsel when set for the 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/5521>
2020-06-17 15:48:35 +00:00
Rhys Perry de7c6950b3 aco: fix sub-dword opsel/sdwa checks
These should all check if the operand has a regclass. The opsel check
should also be skipped post-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/5504>
2020-06-17 10:57:17 +00:00
Rhys Perry 1e791e51a6 aco: fix validation error from vgpr spill/restore code
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/5504>
2020-06-17 10:57:17 +00:00
Daniel Schürmann 8006feda09 aco: don't allow SGPRs on logical phis
aco_validate() is called after phi lowering, now.

Reviewed-by: Rhys Perry <pendingchaos02@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/5496>
2020-06-16 14:46:19 +01:00
Daniel Schürmann 0e47fe3fa2 aco: reorder calls to aco_validate() and cleanup aco_compile_shader()
The first call of aco_validate should happen after phi lowering.
Otherwise, subdword restrictions might be violated

Reviewed-by: Rhys Perry <pendingchaos02@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/5496>
2020-06-16 14:46:19 +01:00
Rob Clark 603d3c2991 radv: don't set num_components for non-vectorized intrinsics
Signed-off-by: Rob Clark <robdclark@chromium.org>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/5371>
2020-06-16 02:48:18 +00:00
Rhys Perry a02e7f6799 aco: fix encoding of certain s_setreg_imm32_b32 instructions
If the mode is too small, the operand will be an inline constant and the
literal dword won't be written.

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/5245>
2020-06-15 18:24:22 +00:00
Rhys Perry 82c265a514 aco: improve check for moving temporaries out of fixed definitions
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/5245>
2020-06-15 18:24:22 +00:00
Rhys Perry e9578e3033 aco: allow GFX9 partial writes with instructions which use opsel
Some instructions such as v_mad_f16 can do partial writes on GFX9.

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/5245>
2020-06-15 18:24:22 +00:00
Rhys Perry 82de70d06e aco: add more opcodes to can_swap_operands
fossil-db (Navi, fp16 enabled):
Totals from 310 (0.24% of 127638) affected shaders:
CodeSize: 1290508 -> 1289716 (-0.06%)

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/5245>
2020-06-15 18:24:22 +00:00
Samuel Pitoiset 3c1b55962e aco: allow to swap operands for some 16-bit float instructions
No fossil-db changes.

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/5245>
2020-06-15 18:24:22 +00:00
Rhys Perry 575b431c80 aco: validate sub-dword pseudo instructions
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/5245>
2020-06-15 18:24:22 +00:00
Rhys Perry d16a7190a3 aco: optimize 16-bit and 64-bit float comparisons
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/5245>
2020-06-15 18:24:22 +00:00
Rhys Perry 22d7122739 aco: copy-propagate constants through p_extract_vector/p_split_vector
fossil-db (Navi, fp16 enabled):
Totals from 1 (0.00% of 127638) affected shaders:
CodeSize: 4388 -> 4392 (+0.09%)
VMEM: 465 -> 458 (-1.51%)
Copies: 54 -> 55 (+1.85%)

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/5245>
2020-06-15 18:24:22 +00:00
Rhys Perry 3d6f67950d aco: improve 8/16-bit constants
fossil-db (Navi, fp16 enabled):
Totals from 1 (0.00% of 127638) affected shaders:
CodeSize: 4540 -> 4388 (-3.35%)
Instrs: 861 -> 830 (-3.60%)
Cycles: 3444 -> 3320 (-3.60%)
VMEM: 489 -> 465 (-4.91%)
SMEM: 107 -> 110 (+2.80%)
SClause: 31 -> 30 (-3.23%)
Copies: 58 -> 54 (-6.90%)

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/5245>
2020-06-15 18:24:22 +00:00
Rhys Perry 4784111abc aco: use 32-bit inline constants for 16-bit integer instructions
See https://reviews.llvm.org/D81841

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/5245>
2020-06-15 18:24:22 +00:00
Rhys Perry dd23345567 aco: fix half_pi constant for 16-bit fsin/fcos
This worked because the optimizer didn't consider that the 16-bit
instruction would interpret the inline constant differently. This will
change in the next commit.

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/5245>
2020-06-15 18:24:22 +00:00
Rhys Perry 9b69ed0bb9 aco: improve sub-dword check for sgpr/constant propagation
p_create_vector can have sub-dword operands with a v1 definition.

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/5245>
2020-06-15 18:24:22 +00:00
Rhys Perry 1210e0bd62 aco: create 16-bit input and output modifiers
fossil-db (Navi, fp16 enabled):
Totals from 1 (0.00% of 127638) affected shaders:
CodeSize: 4552 -> 4540 (-0.26%)
Instrs: 863 -> 861 (-0.23%)
Cycles: 3452 -> 3444 (-0.23%)
VMEM: 490 -> 489 (-0.20%)

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/5245>
2020-06-15 18:24:22 +00:00
Rhys Perry f5a5674178 aco: update comment about preserving fp16/fp64 denormals
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/5245>
2020-06-15 18:24:22 +00:00
Rhys Perry 7f511efa16 aco: create 16-bit mad/fma
fossil-db (Navi, fp16 enabled):
Totals from 1 (0.00% of 127638) affected shaders:
CodeSize: 4868 -> 4552 (-6.49%)
Instrs: 956 -> 863 (-9.73%)
Cycles: 3824 -> 3452 (-9.73%)
VMEM: 504 -> 490 (-2.78%)
SMEM: 109 -> 107 (-1.83%)
VClause: 19 -> 20 (+5.26%)
Copies: 54 -> 58 (+7.41%)
PreVGPRs: 43 -> 41 (-4.65%)

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/5245>
2020-06-15 18:24:22 +00:00
Rhys Perry 1b10764e50 aco: try to use fma instead of mad when denormals are enabled
v_mad_f32 doesn't support denormals but v_fma_f32 does.

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/5245>
2020-06-15 18:24:22 +00:00
Rhys Perry 6cb42cdd8f aco: create mads when signed zeros should be preserved
This check was added because I thought v_mad_f32 didn't preserve the
signess of zero, but I can't reproduce that and this isn't mentioned
anywhere in LLVM.

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/5245>
2020-06-15 18:24:22 +00:00
Rhys Perry 1b6a319c15 aco: add and set precise flag
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/5245>
2020-06-15 18:24:22 +00:00
Rhys Perry a8f800a836 aco: use p_as_uniform in emit_vop1_instruction
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/5245>
2020-06-15 18:24:22 +00:00
Rhys Perry b6d9e45f47 aco: improve code for f2{i,u}{8,16}
Use sub-dword definitions so that the RA can use SDWA

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/5245>
2020-06-15 18:24:22 +00:00
Rhys Perry 34d481fd1f aco: use num_opcodes instead of last_opcode
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/5245>
2020-06-15 18:24:22 +00:00
Samuel Pitoiset 013d096d15 ac: add ac_choose_spi_color_formats() to common code
It's similar between RADV and RadeonSI.

Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/5436>
2020-06-15 08:16:07 +02:00
Erik Faye-Lund cd01d0dee1 radv: update internal reference
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/4630>
2020-06-13 10:42:01 +00:00
Daniel Schürmann 1f98d8c804 aco: fix shared subdword loads
Shared subdword loads don't need byte alignment as they are split
into multiple loads if necessary.

Fixes: 5cde4989d3 ('aco: remove unnecessary split- and create_vector instructions for subdword loads')
Reviewed-by: Rhys Perry <pendingchaos02@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/5441>
2020-06-12 13:56:12 +00:00
Samuel Pitoiset 54e7ae569b radv/llvm: implement radv_enable_mrt_output_nan_fixup workaround
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/5359>
2020-06-12 14:43:58 +02:00
Samuel Pitoiset 7b44f549b3 aco: implement radv_enable_mrt_output_nan_fixup workaround
Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>
Reviewed-by: Rhys Perry <pendingchaos02@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/5359>
2020-06-12 14:43:57 +02:00
Samuel Pitoiset 6f21995f98 radv: add new drirc option radv_enable_mrt_output_nan_fixup
To replace NaN from FS with zeros to fix game bugs.

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/5359>
2020-06-12 14:43:31 +02:00
Samuel Pitoiset 07aefe8065 radv: set DB_SHADER_CONTROL.CONSERVATIVE_Z_EXPORT correctly
Use the SPIR-V execution modes if set.

Cc: 20.1 <mesa-stable@lists.freedesktop.org>
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/5404>
2020-06-12 08:22:47 +00:00
Mauro Rossi 7afdc549f4 android: aco: add aco_ir.cpp to Makefile.sources
Fixes the following building errors:

FAILED: out/target/product/x86_64/obj/SHARED_LIBRARIES/vulkan.radv_intermediates/LINKED/vulkan.radv.so
...
ld.lld: error: undefined symbol: aco::can_use_SDWA(chip_class, std::__1::unique_ptr<aco::Instruction, aco::instr_deleter_functor> const&)
...
ld.lld: error: undefined symbol: aco::can_use_opsel(chip_class, aco_opcode, int, bool)
...
clang-9: error: linker command failed with exit code 1 (use -v to see invocation)

Fixes: d9cfb8ad ("aco: validate instructions reading/writing upper halves/bytes")
Signed-off-by: Mauro Rossi <issor.oruam@gmail.com>
Reviewed-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/5425>
2020-06-11 20:00:16 +02:00
Marek Olšák 0b3e344212 ac/surface: don't free dcc_retile_map on failure
because the hash table now owns it.

Fixes: bd553f0546 - ac/surface: cache DCC retile maps (v2)

Reviewed-by: Pierre-Eric Pelloux-Prayer <pierre-eric.pelloux-prayer@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/5424>
2020-06-11 10:01:57 +00:00
Marek Olšák 56f2a77a41 ac/surface: enable DCC for the first level in the mip tail on gfx10
Acked-by: Pierre-Eric Pelloux-Prayer <pierre-eric.pelloux-prayer@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/5424>
2020-06-11 10:01:57 +00:00
Marek Olšák 7406ea37e6 ac/surface: require that gfx8 doesn't have DCC in order to be displayable
Acked-by: Pierre-Eric Pelloux-Prayer <pierre-eric.pelloux-prayer@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/5424>
2020-06-11 10:01:57 +00:00
Marek Olšák 374f6d568f ac/surface: don't set is_displayable if displayable DCC is missing
If flags.display isn't set, then displayable DCC will not be computed, so
is_displayable will always be false.

Reviewed-by: Pierre-Eric Pelloux-Prayer <pierre-eric.pelloux-prayer@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/5424>
2020-06-11 10:01:57 +00:00
Marek Olšák 0fcf55329b amd/addrlib: fix the C++ one definition rule violation
Closes: https://gitlab.freedesktop.org/mesa/mesa/-/issues/1854

Reviewed-by: Pierre-Eric Pelloux-Prayer <pierre-eric.pelloux-prayer@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/5414>
2020-06-11 05:11:50 -04:00