[hlsl] Only do subgroup matrix robustness transform on buffer with dynamic buffer offset

This patch disables subgroup matrix robustness transform on the storage
buffers without dynamic buffer offset on HLSL writer because HLSL SPEC
requires D3D12 should do bounds checks on the descriptor-based
linear-algebra matrix operations when the buffer is not bound as root
descriptors:
- `@dx.op.linAlgMatrixLoadFromDescriptor` substitutes zero for out
  of bounds reads.
- `@dx.op.linAlgMatrixStoreToDescriptor` turns out of bounds writes
  into no-ops.

The storage buffers with dynamic buffer offset are bound as root
descriptors so they still need robustness transforms in subgroup matrix
load and store operations.

Note that currently we don't need to consider uniform buffers as in
current Tint implementation a uniform buffer can never be the source of
a subgroup matrix load.

See the chapter `bounds-checking-behavior` of below link for more
details:
https://microsoft.github.io/hlsl-specs/proposals/0035-linalg-matrix

Bug: 507459351
Change-Id: Icc511aac86cfbb3886094c1b246417b382e4ed65
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/342395
Commit-Queue: Shao, Jiawei <jiawei.shao@intel.com>
Reviewed-by: James Price <jrprice@google.com>
4 files changed
tree: 604c510136fecc5c7b50c812ee03d7bb5c5bb83c
  1. .github/
  2. .vscode/
  3. agents/
  4. build_overrides/
  5. docs/
  6. generator/
  7. include/
  8. infra/
  9. scripts/
  10. src/
  11. test/
  12. third_party/
  13. tools/
  14. webgpu-cts/
  15. .bazelignore
  16. .bazelrc
  17. .bazelversion
  18. .clang-format
  19. .clang-format-ignore
  20. .clang-tidy
  21. .git-blame-ignore-revs
  22. .gitattributes
  23. .gitignore
  24. .gitmodules
  25. .gn
  26. .style.yapf
  27. AUTHORS
  28. BUILD.bazel
  29. BUILD.gn
  30. CMakeLists.txt
  31. CMakeSettings.json
  32. CODE_OF_CONDUCT.md
  33. codereview.settings
  34. CONTRIBUTING.md
  35. CPPLINT.cfg
  36. DEPS
  37. DIR_METADATA
  38. go.mod
  39. go.sum
  40. go_presubmit_support.py
  41. LICENSE
  42. MODULE.bazel
  43. OWNERS
  44. PRESUBMIT.py
  45. PRESUBMIT_test.py
  46. README.chromium
  47. README.md
  48. unsafe_buffers_paths.txt
  49. vpython.toml
  50. vpython.toml.uv.lock
  51. WATCHLISTS
  52. WORKSPACE.bazel
README.md

Build Status Matrix Space

Dawn, a WebGPU implementation

Dawn is an open-source and cross-platform implementation of the WebGPU standard. More precisely it implements webgpu.h that is a one-to-one mapping with the WebGPU IDL. Dawn is meant to be integrated as part of a larger system and is the underlying implementation of WebGPU in Chromium.

Dawn provides several WebGPU building blocks:

  • WebGPU C/C++ headers that applications and other building blocks use.
    • The webgpu.h version that Dawn implements.
    • A C++ wrapper for the webgpu.h.
  • A “native” implementation of WebGPU using platforms' GPU APIs: D3D12, Metal, Vulkan and OpenGL. See per API support for more details.
  • A client-server implementation of WebGPU for applications that are in a sandbox without access to native drivers
  • Tint is a compiler for the WebGPU Shader Language (WGSL) that can be used in standalone to convert shaders from and to WGSL.

Helpful links:

Documentation table of content

Developer documentation:

User documentation: (TODO, figure out what overlaps with the webgpu.h docs)

License

BSD 3-Clause License, please see LICENSE.

Disclaimer

This is not an officially supported Google product.