The previous matmul is using 1-dimensional shared array. This PR
adds the 2-dimensional shared array. The perf result shows that:
1. For 1-dimensional shared array, enabe-robustness wil bring almost 2x
regression both for matmulFloat and matmulVec4 on Intel CFL.
2. For 2-dimensional shared array, enabe-robustness will bring almost 2x
regression on matmulFloat. But have little impact on matmulVec4 on Intel
CFL.
Tested on Intel_R_UHD_Graphics_630.
shader enable robustness disable robustness
MatMulFloatOneDimSharedArray 5383 us 3105 us
MatMulFloatTwoDimSharedArray 4788 us 2608 us
MatMulVec4OneDimSharedArray 3070 us 1743 us
MatMulVec4TwoDimSharedArray 1840 us 1802 us
Bug: dawn:594
Change-Id: Ia29a78cf70649ef8d3ba8476db1ad4d6ded80840
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/50481
Reviewed-by: Jiawei Shao <jiawei.shao@intel.com>
Reviewed-by: Austin Eng <enga@chromium.org>
Commit-Queue: Jiajia Qin <jiajia.qin@intel.com>
Added in https://dawn-review.googlesource.com/c/tint/+/50742
Bindings may not collide except on the D3D12 backend where we use
the register offset as the binding number. Offsets may alias if
they are for different HLSL register types (buffer, texture, etc.)
Bug: tint:797
Change-Id: I020e3b5cd1d35fdf5678b587225507e0cf24f5b3
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/50801
Commit-Queue: Austin Eng <enga@chromium.org>
Commit-Queue: Ben Clayton <bclayton@google.com>
Auto-Submit: Austin Eng <enga@chromium.org>
Reviewed-by: Ben Clayton <bclayton@google.com>
Certain backends performs transformations on the tint::Program in the ShaderModule.
These transformations may raise diagnostics.
Tint diagnostics generated from an underlying tint::Source::File need that File kept alive.
Change-Id: I32deceb4cedb419e914631d342884fc27972b5cb
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/50581
Auto-Submit: Ben Clayton <bclayton@google.com>
Commit-Queue: Austin Eng <enga@chromium.org>
Reviewed-by: Austin Eng <enga@chromium.org>
Skip these tests on D3D12 because it is not implemented yet.
Bug: dawn:547
Change-Id: I0a009b00bd26bb04794bb627c53f4540462f1ca7
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/50243
Commit-Queue: Austin Eng <enga@chromium.org>
Reviewed-by: Yunchao He <yunchao.he@intel.com>
Bug: dawn:783
Change-Id: I3c7f0ffc3f45a0d67b411a39342e89c710604d54
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/50244
Commit-Queue: Austin Eng <enga@chromium.org>
Reviewed-by: Yunchao He <yunchao.he@intel.com>
Reviewed-by: Stephen White <senorblanco@chromium.org>
This is more consistent with the other backends that have a
CommandRecordingContext.
Bug: dawn:780
Change-Id: Icb689d00cc4d873a489fc4d4c247feeb15f8797c
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/50120
Commit-Queue: Stephen White <senorblanco@chromium.org>
Auto-Submit: Austin Eng <enga@chromium.org>
Reviewed-by: Corentin Wallez <cwallez@chromium.org>
Reviewed-by: Stephen White <senorblanco@chromium.org>
Adds UNASSIGNED-CoreValidation-DrawState-InvalidImageLayout (0x4dae5635)
initially, since we can see that it's causing issues with the WebGPU CTS
tests.
Change-Id: I862e3986d59ae590458f9011a16cc824bf752e72
Bug: dawn:785
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/50423
Commit-Queue: Brandon Jones <bajones@chromium.org>
Reviewed-by: Kai Ninomiya <kainino@chromium.org>
This patch skips CopyTests_T2B/CopyOneRowWithDepth32Float on all
D3D12 backends because it fails on many D3D12 drivers.
See https://crbug.com/dawn/727 for more details.
BUG=dawn:727
TEST=dawn_end2end_tests
Change-Id: I3d2c4cf363772939796a71c7efa165e57ed6ba41
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/50280
Commit-Queue: Corentin Wallez <cwallez@chromium.org>
Reviewed-by: Corentin Wallez <cwallez@chromium.org>
This is now handled inside the SPIR-V sanitizing transform and can be
enabled via a transform config option.
Fixed: tint:753
Change-Id: I897c60dd304dd4643cc905a8ca3001dd963c5fb0
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/50060
Auto-Submit: James Price <jrprice@google.com>
Commit-Queue: Corentin Wallez <cwallez@chromium.org>
Reviewed-by: Corentin Wallez <cwallez@chromium.org>
This CL complete the validation rules for CopyTextureForBrowser by:
- Restrict source texture must have |CopySrc| and |Sampled| usage.
- Restrict destinaton texture must have |CopyDst| and
|RenderAttachment| usage.
- Restrict sample counts of source texture and destination texture
must be 1.
- Restrict source copy origin.z must be 0.
- Restrict CopyTextureForBrowser() can only copy to single slice.
A validation unittest is added to check.
BUG=dawn:465
Change-Id: I5e645a4b69edeaf97ce1231bd7c8036027524ba8
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/49306
Reviewed-by: Corentin Wallez <cwallez@chromium.org>
Reviewed-by: Austin Eng <enga@chromium.org>
Commit-Queue: Jiawei Shao <jiawei.shao@intel.com>
Below are the list of all the individual changes, which are a good order
in which to review this CL.
Core changes:
- Change the tracking in the frontend to produce a synchronization
scope per dispatch instead of per compute pass. Some bindgroups might
not be part of any synchronization scopes so we also track all the
referenced resources on the side so they can be checked during
Queue::Submit validation.
- Fix clearing in the GL and Metal backends to use the per-dispatch
synchronization scopes.
- Fix the Vulkan backend to use the per dispatch synchronization scopes
to produce the correct pipeline barriers. This allows the removal of
previous logic that was subtly incorrect for Indirect buffer. This
allows the merging of the Compute and Render DescriptorSetTracker into
a single small helper class.
- D3D12 changes are similar to Vulkan, but the simplification is just a
the suppression of a branch with a lot of code in
BindGroupStateTracker.
Test changes:
- Fixup all the ResourceUsageTracking tests to follow the WebGPU spec
for synchronization scopes (fixing a lot of TODOs).
- Add additional tests checking that Indirect buffers are not allowed
to be used as a writeable storage in the same synchronization scope.
- Add tests for Queue::Submit validation correctly taking into account
resources that are bound but unused in compute passes.
- Add an end2end test for using a buffer as Indirect and Storage at the
same time in a DispatchIndirect, which would previously produce
incorrect barriers in the Vulkan and D3D12 backends.
Other small changes (that I was to lazy to put in a different CL):
- Add the utils::MakePipelineLayout helper function.
- Fix Indirect not being in the list of readonly buffer usages (caught
by a test added in this CL).
Bug: dawn:632
Change-Id: I77263c3535a4ba995faccbf26255da9a2f6ed3b5
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/49887
Commit-Queue: Corentin Wallez <cwallez@chromium.org>
Reviewed-by: Stephen White <senorblanco@chromium.org>
Add extensions to the appropriate (desktop or ES) LoadProc generation.
Extensions have no removed_procs, so don't check for them.
Bug: chromium:1205155
Change-Id: I705ef3fb700318f59fdbc6415fcf3909c62e39f4
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/50180
Reviewed-by: Corentin Wallez <cwallez@chromium.org>
Reviewed-by: Austin Eng <enga@chromium.org>
Commit-Queue: Stephen White <senorblanco@chromium.org>
This workaround was only necessary in the Metal backend but it was easy
to put it in the frontend so it was put there. However, this simplification
gets in the way of a validation change in the follow-up CL. So we move
the nooping to the Metal backend.
Bug: dawn:632
Change-Id: I0a9957a1bc9bdd7867eb8c60fcab710832c13af2
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/49886
Reviewed-by: Corentin Wallez <cwallez@chromium.org>
Reviewed-by: Stephen White <senorblanco@chromium.org>
Commit-Queue: Corentin Wallez <cwallez@chromium.org>
This change is a preparation for making the compute pass track the
synchronization scope usages per dispatch instead of for the whole pass.
This CL just separates the Compute and RenderPassResourceUsage types.
This requires making the difference between
SyncScope/ComputePass/RenderPass ResourceUsageTracker instead of having
a single combined tracker.
This change also duplicates SetBindGroup by removing the common handling
in ProgrammablePassEncoder and putting it in ComputePassEncoder and
RenderEncoderBase. This is necessary because the UsageTracker types are
now split, but it will also help have different handling of SetBindGroup
for compute and render in follow-up CLs.
There are no functional changes.
Bug: dawn:632
Change-Id: I482c04483d8b734fb10e44e717071eedcff2f15f
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/49884
Reviewed-by: Corentin Wallez <cwallez@chromium.org>
Commit-Queue: Corentin Wallez <cwallez@chromium.org>
This makes error messages from command buffers more useful because they
keep the whole stack trace instead of just showing that the error was
created in the CommandBuffer::Finish call.
Bug: dawn:632
Change-Id: I23e66045c3caa1ad086003a04eed78c40aefc562
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/49885
Reviewed-by: Corentin Wallez <cwallez@chromium.org>
Reviewed-by: Stephen White <senorblanco@chromium.org>
Commit-Queue: Corentin Wallez <cwallez@chromium.org>
This was causing the GL loader to not be rebuilt when
supported_extensions.json was written.
Bug: None
Change-Id: I1f085403ebabe3a4cf52a80bcc7d23217603c28b
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/49888
Reviewed-by: Stephen White <senorblanco@chromium.org>
Reviewed-by: Corentin Wallez <cwallez@chromium.org>
Commit-Queue: Corentin Wallez <cwallez@chromium.org>
This change is a preparation for making the compute pass track the
synchronization scope usages per dispatch instead of for the whole pass.
- Split the tracking of render and compute passes usages.
- Rename PassTextureUsage to TextureSubresourceUsage since is it not
per-pass
- Add SyncScopeResourceUsage as a subclass of PassResourceUsage to
start modifying some of the code to work with synchronization scopes
(even if syncscope == pass at the moment).
There are no functional changes.
Bug: dawn:632
Change-Id: Ieeb6d70a44dc1c726f26989eebcd87e63e732785
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/49883
Reviewed-by: Corentin Wallez <cwallez@chromium.org>
Commit-Queue: Corentin Wallez <cwallez@chromium.org>
This patch implements CallbackTask and CallbackTaskManager to store
the callbacks of Create*PipelineAsync().
In the futureCallbackTaskManager will manage all the callbacks that
should be called in Device.Tick().
BUG=dawn:529
Change-Id: I6ad4352371eb44515bc2d85cdc68220c9b758b8e
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/49060
Reviewed-by: Corentin Wallez <cwallez@chromium.org>
Reviewed-by: Austin Eng <enga@chromium.org>
Commit-Queue: Jiawei Shao <jiawei.shao@intel.com>
Mip dimension should be greater than or equal to 1, while width >> level
may lead to 0.
Bug: dawn:547
Change-Id: Ib3dfb9fbdbed0e922df6efa366598eff0ca10df2
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/49506
Commit-Queue: Yunchao He <yunchao.he@intel.com>
Reviewed-by: Austin Eng <enga@chromium.org>
Dawn Autoroller
Brandon Jones
Qin Jiajia
Austin Eng
Ben Clayton
Jiawei Shao
James Price
Yan
Corentin Wallez
Stephen White
陈俊嘉
Yunchao He