Michel Zou [Sun, 10 Apr 2022 07:51:52 +0000 (09:51 +0200)]
gallium: fix unused symbols warnings
fixes:
d760a915
Reviewed-by: Adam Jackson <ajax@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/15840>
Jason Ekstrand [Fri, 15 Apr 2022 20:20:17 +0000 (15:20 -0500)]
nir/opcodes: fisfinite32 should return bool32
Otherwise constant-folding will fold it to 0/1 instead of 0/~0.
Fixes:
330e28155f79 ("nir: add 32-bit bool of fisfinite")
Reviewed-by: Alyssa Rosenzweig <alyssa@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/15984>
Jason Ekstrand [Fri, 15 Apr 2022 20:06:40 +0000 (15:06 -0500)]
nir,microsoft: Move scale_fdiv into a common NIR pass
While we're at it, convert to nir_shader_instructions_pass() to get rid
of some boilerplate and get metadata correct.
Reviewed-by: Jesse Natalie <jenatali@microsoft.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/15983>
Jason Ekstrand [Fri, 15 Apr 2022 20:24:11 +0000 (15:24 -0500)]
intel/compiler: Set lower_fisnormal
Reviewed-by: Ian Romanick <ian.d.romanick@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/15985>
Rhys Perry [Fri, 25 Mar 2022 19:55:30 +0000 (19:55 +0000)]
radv: use inot(a) instead of ine(a, true)/ieq(a, false)
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/15854>
Rhys Perry [Fri, 25 Mar 2022 19:53:38 +0000 (19:53 +0000)]
radv: use nir_op_imm helpers
Signed-off-by: Rhys Perry <pendingchaos02@gmail.com>
Acked-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/15854>
Rhys Perry [Wed, 30 Mar 2022 10:56:01 +0000 (11:56 +0100)]
radv: use radv_meta_build_nir_vs_generate_vertices() more
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/15854>
Rhys Perry [Wed, 30 Mar 2022 10:52:58 +0000 (11:52 +0100)]
radv: initialize meta shader options earlier
We might need this for NIR builders.
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/15854>
Rhys Perry [Tue, 29 Mar 2022 19:16:01 +0000 (20:16 +0100)]
radv: remove unnecessary align_mul/offset initialization
Since
af51efe195b, these builders fill these in with scalar alignment if
not provided.
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/15854>
Rhys Perry [Tue, 29 Mar 2022 19:16:54 +0000 (20:16 +0100)]
nir/builder: add nir_{ine,ibfe,ubfe}_imm() helper
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/15854>
Rhys Perry [Thu, 14 Apr 2022 18:20:40 +0000 (19:20 +0100)]
nir/gather_info: fix system_value_read for rt/mesh system values
Signed-off-by: Rhys Perry <pendingchaos02@gmail.com>
Reviewed-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Fixes:
c7eaf030687 ("radv: use shader_info::system_values_read")
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/15952>
Rhys Perry [Thu, 14 Apr 2022 17:57:51 +0000 (18:57 +0100)]
ac/nir: properly handle large global access constant offsets
Signed-off-by: Rhys Perry <pendingchaos02@gmail.com>
Reviewed-by: Timur Kristóf <timur.kristof@gmail.com>
Fixes:
61ac5acca3b ("radv,ac/nir: lower global access to _amd global access intrinsics")
Closes: https://gitlab.freedesktop.org/mesa/mesa/-/issues/6321
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/15951>
Lionel Landwerlin [Wed, 1 Sep 2021 10:01:11 +0000 (13:01 +0300)]
anv: use shadow surface for stencil input attachment on gfx7
This fixes a number of tests like :
dEQP-VK.renderpass*.suballocation.multisample.s8_uint.*
dEQP-VK.renderpass*.suballocation.multisample.separate_stencil_usage.d24_unorm_s8_uint.*.test_stencil
dEQP-VK.renderpass*.suballocation.multisample.d24_unorm_s8_uint.*
dEQP-VK.renderpass*.suballocation.multisample.d32_sfloat_s8_uint.*
Because the driver asserts when generating RENDER_SURFACE_STATE with a
8 Valign value for stencil buffer (only 2 & 4 are supported).
Signed-off-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Reviewed-by: Nanley Chery <nanley.g.chery@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12670>
Samuel Pitoiset [Wed, 13 Apr 2022 14:37:09 +0000 (16:37 +0200)]
radv: add support for VkShaderModuleCreateInfo as pNext
With VK_EXT_graphics_pipeline_library, modules can be passed via the
pNext of VkpipelineShaderStageCreateInfo.
Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-by: Mike Blumenkrantz <michael.blumenkrantz@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/15926>
Samuel Pitoiset [Mon, 11 Apr 2022 10:05:48 +0000 (12:05 +0200)]
radv: add support for independent descriptor set layouts
With VK_EXT_graphics_pipeline_library, pipeline layouts created with
VK_PIPELINE_LAYOUT_CREATE_INDEPENDENT_SETS_BIT_EXT might contain NULL
descriptor sets.
Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-by: Mike Blumenkrantz <michael.blumenkrantz@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/15926>
Samuel Pitoiset [Fri, 18 Mar 2022 07:34:26 +0000 (08:34 +0100)]
radv: fix enabling adjust_frag_coord_z and apply per-pipeline
Fossilize always enables all supported extensions, that means that
adjust_frag_coord_z would always be enabled on RDNA2, even if the
application doesn't enable it. The pipeline key would then be different
and precompilation wouldn't work. Move this per-pipeline since we can
know if VRS will be used.
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/15444>
M Henning [Tue, 8 Feb 2022 06:00:02 +0000 (01:00 -0500)]
nouveau/nir: Allow up to 6 nested joins
This matches what the tgsi path does and doesn't regress any tests. (For
comparison, unlimited join nesting does regress tests in deqp and piglit)
Fixes graphical artifacts from stack overflows in
https://www.shadertoy.com/view/Xds3zN
with nir on kepler
Reviewed-by: Emma Anholt <emma@anholt.net>
Acked-by: Karol Herbst <kherbst@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/15597>
Enrico Galli [Fri, 11 Feb 2022 01:47:32 +0000 (17:47 -0800)]
microsoft/compiler: Fix when using a shadow sampler more than once
Reusing the shadow sampler's variable causes problems when the sampler
is used more than once. The remaining `deref_var`s will be using the
wrong type.
Reviewed-by: Boris Brezillon <boris.brezillon@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/14988>
Enrico Galli [Thu, 10 Feb 2022 22:37:32 +0000 (14:37 -0800)]
microsoft/compiler: Add support for unnamed ubos
Reviewed-by: Jesse Natalie <jenatali@microsoft.com>
Reviewed-by: Boris Brezillon <boris.brezillon@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/14988>
Enrico Galli [Thu, 10 Feb 2022 22:30:12 +0000 (14:30 -0800)]
microsoft/compiler: Fix for arrays of 1 element
The DXIL validator doesn't like dynamic indexing into resources if the
resource was not declared as an array type. This commit makes it so that
we always generate array resource types if the original type was
declared as an array instead, not just when the number of elements is
greater than 1.
Reviewed-by: Boris Brezillon <boris.brezillon@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/14988>
Enrico Galli [Wed, 30 Mar 2022 15:25:22 +0000 (08:25 -0700)]
microsoft/spirv_to_dxil: Add pass to lower dynamic accesses on ubo[1]
Reviewed-by: Boris Brezillon <boris.brezillon@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/14988>
Mike Blumenkrantz [Wed, 13 Apr 2022 20:44:18 +0000 (16:44 -0400)]
zink: nv ci update
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/15957>
Mike Blumenkrantz [Thu, 14 Apr 2022 21:29:02 +0000 (17:29 -0400)]
zink: radv baseline fix
typo
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/15957>
Emma Anholt [Tue, 12 Apr 2022 18:59:00 +0000 (11:59 -0700)]
ci/virgl: Add piglit and GL4.3 testing to the virpipe path.
This will be helpful in regression-testing the nir-to-tgsi transition, and
with the big runners at google we have plenty of capacity to do it.
I dropped the GL3.0-3.2 caselists because GL4.3 should be a superset of
them.
Acked-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/15158>
Mike Blumenkrantz [Tue, 12 Apr 2022 15:46:47 +0000 (11:46 -0400)]
zink: unset PIPE_CAP_TEXRECT
this isn't supported by vulkan, and though it can be implicitly converted
to 2d in the driver, this still leaves validation spam
Reviewed-by: Dave Airlie <airlied@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/15895>
Mike Blumenkrantz [Tue, 12 Apr 2022 15:45:33 +0000 (11:45 -0400)]
mesa/st: set normalized coords for RECT samplers if rects are unsupported
the shaders will never see these, so set the expected value for 2D
cc: mesa-stable
Reviewed-by: Dave Airlie <airlied@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/15895>
Mike Blumenkrantz [Tue, 12 Apr 2022 15:42:14 +0000 (11:42 -0400)]
nir/lower_tex: fix rect queries with lower_rect set
queries still need the sampler_dim changed
Fixes:
682e14d3eae ("nir: lower_tex: Don't normalize coordinates for TXF with RECT")
Reviewed-by: Dave Airlie <airlied@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/15895>
Emma Anholt [Wed, 13 Apr 2022 23:10:49 +0000 (16:10 -0700)]
Revert "ci: Disable Google's lab"
This reverts commit
8506c2b7ee447faabcda7c4b123717d6e375ee9c. Network
admins say the fixed things, and we see the runners phoning home again.
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/15936>
Jason Ekstrand [Wed, 13 Apr 2022 21:52:17 +0000 (16:52 -0500)]
clc: Declare LLVMContexts on the stack
This prevents more use-after-free errors. Passing them around using
std::unique_ptr ensures that the LLVMContext gets destroyed but doesn't
ensure destruction order. Declaring it on the stack ensures that the
context doesn't get destroyed until right before the the function
returns which is after any other LLVM stuff is destroyed.
Reviewed-by: Jesse Natalie <jenatali@microsoft.com>
Reviewed-by: Icecream95 <ixn@disroot.org>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/15937>
Jason Ekstrand [Wed, 13 Apr 2022 21:16:30 +0000 (16:16 -0500)]
clc: Rework logging a bit
First, separate out the LLVM context logging to make it take a
clc_logger instead of passing in a string stream. Currently, the LLVM
context may outlive the string stream which we assign which may lead to
use-after-free errors. Second, use a separate string stream for clang
diagnosticl logging which we intentionally declare before the compiler
so the compiler can't outlive it.
Reviewed-by: Jesse Natalie <jenatali@microsoft.com>
Reviewed-by: Icecream95 <ixn@disroot.org>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/15937>
Jason Ekstrand [Fri, 8 Apr 2022 21:20:43 +0000 (16:20 -0500)]
clc: Only initialize LLVM once
Reviewed-by: Jesse Natalie <jenatali@microsoft.com>
Reviewed-by: Icecream95 <ixn@disroot.org>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/15937>
Dave Airlie [Wed, 13 Oct 2021 01:25:15 +0000 (11:25 +1000)]
clc: initialise one more llvm stage
Reviewed-by: Jason Ekstrand <jason.ekstrand@collabora.com>
Reviewed-by: Jesse Natalie <jenatali@microsoft.com>
Reviewed-by: Icecream95 <ixn@disroot.org>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/15937>
Dave Airlie [Wed, 13 Oct 2021 01:18:37 +0000 (11:18 +1000)]
clc: add simple llvm initialise API
This just calls some of the LLVM init functions in a common place
Reviewed-by: Jason Ekstrand <jason.ekstrand@collabora.com>
Reviewed-by: Jesse Natalie <jenatali@microsoft.com>
Reviewed-by: Icecream95 <ixn@disroot.org>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/15937>
Rhys Perry [Tue, 12 Apr 2022 14:19:40 +0000 (15:19 +0100)]
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>
Samuel Pitoiset [Fri, 11 Mar 2022 08:53:46 +0000 (09:53 +0100)]
radv: use shader_info::system_values_read
Signed-off-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/15343>
Samuel Pitoiset [Fri, 11 Mar 2022 08:29:28 +0000 (09:29 +0100)]
radv: use shader_info::outputs_written/per_primitive_outputs for VS outputs
Signed-off-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/15343>
Samuel Pitoiset [Fri, 11 Mar 2022 07:54:11 +0000 (08:54 +0100)]
radv: use shader_info::inputs_read/outputs_written for FS IO
Signed-off-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/15343>
Samuel Pitoiset [Thu, 10 Mar 2022 17:20:47 +0000 (18:20 +0100)]
radv: use shader_info::gs::active_stream_mask
Signed-off-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/15343>
Samuel Pitoiset [Thu, 10 Mar 2022 16:43:55 +0000 (17:43 +0100)]
radv: use shader_info::writes_memory
Signed-off-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/15343>
Boris Brezillon [Mon, 11 Apr 2022 12:20:04 +0000 (05:20 -0700)]
dzn: Make sure sampler heaps don't contain more than 2048 samplers
The spec says "The maximum number of samplers in a shader visible
descriptor heap is 2048.". Let's make sure we follow this rule
in dozen.
Acked-by: Erik Faye-Lund <erik.faye-lund@collabora.com>
Reviewed-by: Jesse Natalie <jenatali@microsoft.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/15884>
Boris Brezillon [Mon, 11 Apr 2022 11:46:31 +0000 (04:46 -0700)]
dzn: Skip binding entries with zero descriptors
D3D12 doesn't like empty descriptor ranges, so let's skip those
at set layout creation time.
Acked-by: Erik Faye-Lund <erik.faye-lund@collabora.com>
Reviewed-by: Jesse Natalie <jenatali@microsoft.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/15884>
Emma Anholt [Sat, 26 Mar 2022 04:06:35 +0000 (21:06 -0700)]
nouveau/nir: Don't try to emit OP_FMA pre-nvc0.
The TGSI backend avoids TGSI_OPCODE_FMA (and thus OP_FMA) pre-nvc0,
replacing it with TGSI_OPCODE_MAD in that case.
Noticed when looking at native-NIR stats and finding that load
optimization wasn't taking place on the unsupported opcode.
Reviewed-by: Karol Herbst <kherbst@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/15543>
Gert Wollny [Wed, 13 Apr 2022 10:04:33 +0000 (12:04 +0200)]
virgl: Extend integer write out output fix to all non-move integers ops
The host virglrenderer can only handle moves to integer outputs, all
ALU opt that create integer outputs are created with extra code to convert
to float for the temporaries, and this breaks the output write
handling.
Fixes:
spec@arb_sample_shading@builtin-gl-sample-mask *
spec@arb_sample_shading@builtin-gl-sample-mask-simple *
Signed-off-by: Gert Wollny <gert.wollny@collabora.com>
Reviewed-by: Emma Anholt <emma@anholt.net>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/15921>
Samuel Pitoiset [Wed, 13 Apr 2022 09:12:28 +0000 (11:12 +0200)]
radv: exclude PRIMITIVE_{COUNT,INDICES} from the per-vertex output mask
They should be excluded for the primitive and vertex output masks.
Signed-off-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/15918>
Icecream95 [Wed, 19 Jan 2022 08:41:23 +0000 (21:41 +1300)]
clc: Use stringstream for printing spirv errors
The type of the spv_position_t components can differ across platforms,
it's simpler to just let C++ overloading handle it.
Reviewed-by: Karol Herbst <kherbst@redhat.com>
Reviewed-by: Jason Ekstrand <jason.ekstrand@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/15437>
Dylan Baker [Wed, 13 Apr 2022 22:54:22 +0000 (15:54 -0700)]
docs: truncate new_features.txt
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/15935>
Dylan Baker [Wed, 13 Apr 2022 22:52:31 +0000 (15:52 -0700)]
VERSION: bump to 22.2-devel for next cycle
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/15935>
Rhys Perry [Fri, 12 Nov 2021 10:28:24 +0000 (10:28 +0000)]
radv: use load_shared2_amd/store_shared2_amd
fossil-db (Sienna Cichlid):
Totals from 376 (0.23% of 162293) affected shaders:
MaxWaves: 9620 -> 9596 (-0.25%); split: +0.08%, -0.33%
Instrs: 207533 -> 203901 (-1.75%); split: -1.76%, +0.01%
CodeSize: 1130904 -> 1106420 (-2.16%); split: -2.17%, +0.01%
VGPRs: 14016 -> 14120 (+0.74%); split: -0.34%, +1.08%
Latency: 2143281 -> 2132212 (-0.52%); split: -0.56%, +0.05%
InvThroughput: 389116 -> 387990 (-0.29%); split: -0.34%, +0.05%
VClause: 4483 -> 4485 (+0.04%); split: -0.11%, +0.16%
SClause: 5780 -> 5778 (-0.03%); split: -0.17%, +0.14%
Copies: 15319 -> 15331 (+0.08%); split: -0.53%, +0.61%
Branches: 5561 -> 5563 (+0.04%)
PreSGPRs: 11776 -> 11775 (-0.01%)
PreVGPRs: 11393 -> 11497 (+0.91%); split: -0.13%, +1.04%
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>
Rhys Perry [Fri, 12 Nov 2021 10:45:46 +0000 (10:45 +0000)]
ac/llvm: 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>
Rhys Perry [Fri, 12 Nov 2021 10:28:13 +0000 (10:28 +0000)]
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>
Rhys Perry [Mon, 15 Nov 2021 16:40:53 +0000 (16:40 +0000)]
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>
Rhys Perry [Wed, 10 Nov 2021 15:02:24 +0000 (15:02 +0000)]
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>
Rhys Perry [Fri, 12 Nov 2021 10:27:13 +0000 (10:27 +0000)]
nir/opt_load_store_vectorize: create 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>
Rhys Perry [Thu, 11 Nov 2021 16:07:20 +0000 (16:07 +0000)]
nir/opt_load_store_vectorize: fix broken indentation
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>
Rhys Perry [Fri, 12 Nov 2021 10:26:30 +0000 (10:26 +0000)]
nir: add load_shared2_amd and 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>
Konstantin Seurer [Wed, 13 Apr 2022 19:02:55 +0000 (21:02 +0200)]
radv: Fix barriers with cp dma
We need to wait for cp dma if VK_PIPELINE_STAGE_2_ALL_TRANSFER_BIT or
VK_PIPELINE_STAGE_2_ALL_COMMANDS_BIT are set.
Closes: #5911
Fixes:
4b9bc4791b5 ("radv: only sync CP DMA for transfer operations or bottom pipe")
Signed-off-by: Konstantin Seurer <konstantin.seurer@gmail.com>
Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/15933>
Daniel Schürmann [Tue, 15 Mar 2022 15:28:06 +0000 (16:28 +0100)]
aco: remove register hints entirely
Reviewed-by: Rhys Perry <pendingchaos02@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/15408>
Daniel Schürmann [Tue, 15 Mar 2022 13:49:32 +0000 (14:49 +0100)]
aco: remove occurences of VCC hint
Reviewed-by: Rhys Perry <pendingchaos02@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/15408>
Daniel Schürmann [Wed, 16 Mar 2022 09:56:26 +0000 (10:56 +0100)]
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>
Daniel Schürmann [Tue, 15 Mar 2022 12:06:48 +0000 (13:06 +0100)]
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>
Daniel Schürmann [Tue, 15 Mar 2022 11:15:44 +0000 (12:15 +0100)]
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>
Daniel Schürmann [Wed, 16 Mar 2022 09:59:52 +0000 (10:59 +0100)]
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>
Lionel Landwerlin [Wed, 13 Apr 2022 10:06:43 +0000 (13:06 +0300)]
anv: stop using old entrypoint/struct/enum names for 1.3
v2: More replacements
Signed-off-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Reviewed-by: Jason Ekstrand <jason.ekstrand@collabora.com>
Reviewed-by: Tapani Pälli <tapani.palli@intel.com> (v1)
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/15920>
Emma Anholt [Tue, 12 Apr 2022 17:33:14 +0000 (10:33 -0700)]
nir_to_tgsi: Do the required cleanup for nir_opt_find_array_copies().
If we made a copy deref, then we need to do dead-write elimination for the
pervious writes or we'll just emit the same copy deref again next time
around. And, at the end of the opt loop, we need to lower copy derefs
because later passes (locals_to_regs, notably) depend on it.
Fixes infinite opt loop on fs-function-inout-array with virgl on NTT.
Reviewed-by: Gert Wollny <gert.wollny@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/15899>
Jason Ekstrand [Tue, 12 Apr 2022 18:34:26 +0000 (13:34 -0500)]
iris: More gracefully fail in resource_from_user_memory
rusticl (and clover) would like to get a graceful fail here so they can
fall back to a shadow copy instead of us asserting. We also start
rejecting arrayed surface because isl doesn't allow selecting a QPitch
yet. Even if it did, QPitch is horribly restrictive, even for linear
surfaces, that it likely wouldn't be that useful.
Fixes:
e81f3edf76b0 ("iris: Allow userptr on 1D and 2D images")
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/15903>
Mike Blumenkrantz [Wed, 13 Apr 2022 15:36:53 +0000 (11:36 -0400)]
zink: set optimal tiling on swapchain images
this otherwise breaks kopper
fixes #6294
Reviewed-by: Erik Faye-Lund <erik.faye-lund@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/15928>
Louis-Francis Ratté-Boulianne [Thu, 10 Feb 2022 16:03:57 +0000 (11:03 -0500)]
dzn: Add CI target for vulkan driver
A custom branch of `deqp` is used to have proper results when
crashing. See:
https://github.com/KhronosGroup/VK-GL-CTS/issues/311
A custom branch of `deqp-runner` with Windows support is also
used until the changes are merged into the main repository.
The `api`, `info`, `draw`, `query-pool` and `memory` test cases are
executed for now.
Signed-off-by: Louis-Francis Ratté-Boulianne <lfrb@collabora.com>
Acked-by: Boris Brezillon <boris.brezillon@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/15742>
Louis-Francis Ratté-Boulianne [Mon, 4 Apr 2022 18:45:20 +0000 (14:45 -0400)]
dzn: Add a debug flag to enable D3D12 debug layer
Signed-off-by: Louis-Francis Ratté-Boulianne <lfrb@collabora.com>
Reviewed-by: Boris Brezillon <boris.brezillon@collabora.com>
Reviewed-by: Erik Faye-Lund <erik.faye-lund@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/15742>
Karmjit Mahil [Wed, 23 Feb 2022 11:48:05 +0000 (11:48 +0000)]
pvr: Implement vkCreateQueryPool() and vkDestroyQueryPool().
Signed-off-by: Karmjit Mahil <Karmjit.Mahil@imgtec.com>
Reviewed-by: Frank Binns <frank.binns@imgtec.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/15880>
Karmjit Mahil [Wed, 23 Feb 2022 15:43:54 +0000 (15:43 +0000)]
pvr: Add pvrsrvkm visibility test heap.
Signed-off-by: Karmjit Mahil <Karmjit.Mahil@imgtec.com>
Reviewed-by: Frank Binns <frank.binns@imgtec.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/15880>
Karmjit Mahil [Wed, 23 Feb 2022 13:51:55 +0000 (13:51 +0000)]
pvr: Add core count info and pvr_device_runtime_info.
Signed-off-by: Karmjit Mahil <Karmjit.Mahil@imgtec.com>
Reviewed-by: Frank Binns <frank.binns@imgtec.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/15880>
Jason Ekstrand [Mon, 4 Apr 2022 23:37:26 +0000 (18:37 -0500)]
v3dv: Add emulated timeline semaphore support
This is trivial thanks to the emulated timelines provided in common
code. "Real" timeline semaphores which can be shared across processes
will require kernel support.
Reviewed-by: Iago Toral Quiroga <itoral@igalia.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/15704>
Jason Ekstrand [Mon, 4 Apr 2022 23:33:55 +0000 (18:33 -0500)]
v3dv: Use the core version property helpers
vulkaninfo is the same before and after.
Reviewed-by: Iago Toral Quiroga <itoral@igalia.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/15704>
Jason Ekstrand [Mon, 4 Apr 2022 23:24:30 +0000 (18:24 -0500)]
v3dv: Use the core version feature helpers
vulkaninfo is the same before and after.
Reviewed-by: Iago Toral Quiroga <itoral@igalia.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/15704>
Jason Ekstrand [Tue, 29 Mar 2022 22:52:32 +0000 (17:52 -0500)]
v3dv: Switch to the common submit framework
Reviewed-by: Iago Toral Quiroga <itoral@igalia.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/15704>
Jason Ekstrand [Thu, 7 Apr 2022 15:17:26 +0000 (10:17 -0500)]
v3dv: Always wait on last_job_syncs if job->serialize
Even if we're the first job on some queue, there may be no wait
semaphores but we still need to ensure things happen in-order. (See
the "Implicit Synchronization Guarantees" section of the Vulkan spec.)
The client can submit back-to-back command buffers with no semaphores
between them and it needs to adt the same as if there were a semaphore.
If job->serialize is set because of a barrier or something, we still
need to synchronize across HW queues by waiting on last_job_syncs.
Reviewed-by: Iago Toral Quiroga <itoral@igalia.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/15704>
Jason Ekstrand [Mon, 4 Apr 2022 15:25:15 +0000 (10:25 -0500)]
v3dv: Add a condition variable for queries
In order to properly wait for a query to be complete, we need to first
wait for the end query job to flush through on the queue. Since query
end is always handled on the CPU, we can do this with a condition
variable. The 2s timeout is taken from ANV.
Reviewed-by: Alejandro Piñeiro <apinheiro@igalia.com>
Reviewed-by: Iago Toral Quiroga <itoral@igalia.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/15704>
Jason Ekstrand [Mon, 4 Apr 2022 15:22:45 +0000 (10:22 -0500)]
v3dv: Use util/os_time helpers
Reviewed-by: Alejandro Piñeiro <apinheiro@igalia.com>
Reviewed-by: Iago Toral Quiroga <itoral@igalia.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/15704>
Jason Ekstrand [Mon, 4 Apr 2022 14:50:26 +0000 (09:50 -0500)]
v3dv: Switch to the common device lost tracking
Vulkan requires that, once the device has been lost, you keep returning
VK_ERROR_DEVICE_LOST. We've got tracking for this in common code; it
just needs to be wired up.
Reviewed-by: Alejandro Piñeiro <apinheiro@igalia.com>
Reviewed-by: Iago Toral Quiroga <itoral@igalia.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/15704>
Jason Ekstrand [Mon, 4 Apr 2022 13:44:53 +0000 (08:44 -0500)]
v3dv: Destroy the device mutex on the teardown path
Reviewed-by: Alejandro Piñeiro <apinheiro@igalia.com>
Reviewed-by: Iago Toral Quiroga <itoral@igalia.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/15704>
Jason Ekstrand [Mon, 4 Apr 2022 13:40:30 +0000 (08:40 -0500)]
v3dv: Don't use pthread functions on c11 mutexes
This only works because c11/threads.h is typedeffing the c11 stuff to
ptrheads.
Reviewed-by: Alejandro Piñeiro <apinheiro@igalia.com>
Reviewed-by: Iago Toral Quiroga <itoral@igalia.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/15704>
Jason Ekstrand [Fri, 1 Apr 2022 21:10:12 +0000 (16:10 -0500)]
v3dv: Put indirect compute CSD jobs in the job list
Instead of having the CPU job execute the CSD job, put both jobs on the
list with the CPU job first which modifies the GPU job which gets kicked
off next. This gives the queue code more visibility into what types of
jobs are actually in the list. In particular, if an indirect compute
job is the last job in a batch buffer, it currently appears as if the
batch ends with CPU work which isn't true because it kicks off GPU work.
In that case, the last job on the list is now a GPU job, which better
matches reality.
Reviewed-by: Iago Toral Quiroga <itoral@igalia.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/15704>
Jason Ekstrand [Tue, 29 Mar 2022 22:55:27 +0000 (17:55 -0500)]
v3dv: Stop directly setting vk_device::alloc
vk_device_init() will do this.
Reviewed-by: Alejandro Piñeiro <apinheiro@igalia.com>
Reviewed-by: Iago Toral Quiroga <itoral@igalia.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/15704>
Jason Ekstrand [Thu, 31 Mar 2022 20:29:30 +0000 (15:29 -0500)]
vulkan/drm_syncobj: Implement WAIT_PENDING with a sync_file lookup
The v3dv kernel driver doesn't support timelines yet but we want
threaded submit and that requires WAIT_PENDING. Fortunately, it should
never sit in this loop for long in practice. The primary use-case is
sorting out dependencies and these checks will always trivially succeed
for non-shared semaphores because v3dv only has a single queue.
Acked-by: Alejandro Piñeiro <apinheiro@igalia.com>
Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/15704>
Rhys Perry [Thu, 2 Dec 2021 14:38:57 +0000 (14:38 +0000)]
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>
Rhys Perry [Thu, 2 Dec 2021 14:35:15 +0000 (14:35 +0000)]
radv,ac/nir: lower global access to _amd global access intrinsics
fossil-db (Sienna Cichlid):
Totals from 400 (0.30% of 134621) affected shaders:
VGPRs: 18696 -> 18688 (-0.04%)
CodeSize: 2031348 -> 1946640 (-4.17%)
Instrs: 374703 -> 360226 (-3.86%)
Latency: 4200727 -> 4108628 (-2.19%); split: -2.20%, +0.01%
InvThroughput: 1059935 -> 1029441 (-2.88%); split: -2.88%, +0.00%
VClause: 5777 -> 5771 (-0.10%)
SClause: 11890 -> 10891 (-8.40%); split: -8.57%, +0.17%
Copies: 34035 -> 33259 (-2.28%); split: -2.98%, +0.70%
Branches: 11108 -> 11100 (-0.07%); split: -0.08%, +0.01%
PreSGPRs: 15999 -> 15942 (-0.36%); split: -0.44%, +0.08%
PreVGPRs: 16994 -> 16970 (-0.14%)
fossil-db (Polaris10):
Totals from 400 (0.29% of 135668) affected shaders:
SGPRs: 23799 -> 22919 (-3.70%); split: -4.30%, +0.61%
VGPRs: 18480 -> 18472 (-0.04%)
CodeSize: 2090316 -> 2041592 (-2.33%)
Instrs: 395461 -> 385747 (-2.46%); split: -2.46%, +0.00%
Latency: 5045768 -> 5020196 (-0.51%); split: -0.53%, +0.02%
InvThroughput: 2694320 -> 2689886 (-0.16%); split: -0.23%, +0.07%
VClause: 5982 -> 5968 (-0.23%)
SClause: 12064 -> 10823 (-10.29%); split: -10.33%, +0.04%
Copies: 48233 -> 48322 (+0.18%); split: -0.47%, +0.65%
PreSGPRs: 16409 -> 16358 (-0.31%); split: -0.39%, +0.08%
fossil-db (Pitcairn):
Totals from 400 (0.29% of 135668) affected shaders:
SGPRs: 22431 -> 22215 (-0.96%); split: -2.60%, +1.64%
VGPRs: 18776 -> 18560 (-1.15%); split: -1.21%, +0.06%
CodeSize: 2104440 -> 2017708 (-4.12%)
MaxWaves: 2363 -> 2367 (+0.17%)
Instrs: 413099 -> 397446 (-3.79%)
Latency: 5507707 -> 5450251 (-1.04%); split: -1.12%, +0.07%
InvThroughput: 2838867 -> 2786903 (-1.83%); split: -1.83%, +0.00%
VClause: 10334 -> 10097 (-2.29%)
SClause: 12346 -> 11005 (-10.86%); split: -10.89%, +0.02%
Copies: 54034 -> 52065 (-3.64%); split: -3.99%, +0.35%
PreSGPRs: 17916 -> 17857 (-0.33%); split: -0.40%, +0.07%
PreVGPRs: 16917 -> 16893 (-0.14%)
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>
Rhys Perry [Fri, 3 Dec 2021 13:48:28 +0000 (13:48 +0000)]
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>
Rhys Perry [Thu, 2 Dec 2021 14:34:52 +0000 (14:34 +0000)]
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>
Rhys Perry [Fri, 3 Dec 2021 16:07:24 +0000 (16:07 +0000)]
ac/llvm: implement _amd global access intrinsics
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>
Rhys Perry [Thu, 2 Dec 2021 14:33:17 +0000 (14:33 +0000)]
nir: add _amd global access intrinsics
These are the same as the normal ones, but they take an unsigned 32-bit
offset in BASE and another unsigned 32-bit offset in the last source.
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>
Rhys Perry [Thu, 2 Dec 2021 10:57:35 +0000 (10:57 +0000)]
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>
Rhys Perry [Tue, 9 Mar 2021 16:09:15 +0000 (16:09 +0000)]
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>
Rhys Perry [Tue, 9 Mar 2021 16:40:23 +0000 (16:40 +0000)]
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>
Rhys Perry [Thu, 15 Apr 2021 13:22:11 +0000 (14:22 +0100)]
radv: don't require robust vectorization for nir_var_mem_global
Robust vectorization is to prevent vectorization of loads using the near
maximum offset with loads of offset 0. Global loads can't read from offset
0 (NULL) anyways, so this isn't necessary.
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>
Jason Ekstrand [Tue, 12 Apr 2022 16:45:41 +0000 (11:45 -0500)]
iris: Don't leak scratch BOs
Fixes:
4d219b0eb3d6 ("iris: implement scratch space!")
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/15897>
Timur Kristóf [Wed, 13 Apr 2022 12:54:30 +0000 (14:54 +0200)]
radv: Only use TES vertex offset 2 for triangles and quads.
Signed-off-by: Timur Kristóf <timur.kristof@gmail.com>
Reviewed-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/15837>
Timur Kristóf [Sat, 9 Apr 2022 20:00:10 +0000 (22:00 +0200)]
radv: Fix gs_vgpr_comp_cnt for NGG VS without passthrough mode.
When not in passthrough mode, the NGG shader needs to calculate the
primitive export value from the input primitive's vertex indices.
So, GS vertex offset 2 is needed when NGG has triangles
and isn't in passthrough mode.
Fixes:
7ad69e2f7ee10c0e7afc302b9324e7a320424dcb "radv: stop loading invocation ID for NGG vertex shaders"
Signed-off-by: Timur Kristóf <timur.kristof@gmail.com>
Reviewed-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/15837>
Timur Kristóf [Wed, 6 Apr 2022 16:53:20 +0000 (18:53 +0200)]
nir: Handle out of bounds access in nir_vectorize_tess_levels.
Replace out of bounds loads with undef.
Then, delete instructions with out of bounds access.
Fixes:
f5adf27fb926a330a13af716f0a03da1a224656d "nir,radv: add and use nir_vectorize_tess_levels()"
Closes: https://gitlab.freedesktop.org/mesa/mesa/-/issues/6264
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/15775>
Timur Kristóf [Wed, 13 Apr 2022 12:11:18 +0000 (14:11 +0200)]
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:
93c8ebfa780ebd1495095e794731881aef29e7d3 "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>
Rohan Garg [Tue, 12 Apr 2022 19:07:13 +0000 (21:07 +0200)]
iris: set a default EDSC flag
anv sets the default EDSC flag, do the same for iris too
Fixes:
5ae278da18b6 ("iris: use vtbl to avoid multiple symbols, fix state base address")
Signed-off-by: Rohan Garg <rohan.garg@intel.com>
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
Acked-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/15905>