Repository navigation
[Vulkan] Keep unrepresentable and symbolic scalars out of the delegate - #23247
Merged
Merged
Conversation
…ghstack [ghstack-poisoned]
…ghstack [ghstack-poisoned]
…ghstack [ghstack-poisoned]
…ghstack [ghstack-poisoned]
…ghstack [ghstack-poisoned]
…ghstack [ghstack-poisoned]
…ghstack [ghstack-poisoned]
…ghstack [ghstack-poisoned]
Contributor
Author
mergennachin
added
the
module: vulkan
🔗 Helpful Links🧪 See artifacts and rendered test results at hud.pytorch.org/pr/pytorch/executorch/23247
Note: Links to docs will display an error until the docs builds have been completed. This comment was automatically generated by Dr. CI and updates every 15 minutes. |
mergennachin
temporarily deployed
to
cadence
September 29, 2026 15:03 — with 
GitHub Actions
Inactive
GitHub Actions
Inactive
mergennachin
temporarily deployed
to
cadence
September 29, 2026 15:03 — with
GitHub Actions
Inactive
GitHub Actions
Inactive
This was referenced Sep 29, 2026
meta-cla
Bot
added
the
CLA Signed
mergennachin
added a commit
that referenced
this pull request
Sep 29, 2026
…ormer blocks The dynamic transformer reproductions now lower to one Vulkan delegate across changing sequence lengths. This final PR makes `expand_copy` resizable and adds the eager-attention and SDPA integration tests, including fully masked rows on devices with 8-bit storage buffers. Fixes #23156. This is part 15 of a linear ghstack series. Review each PR against its selected base branch; this PR contains only the final expansion and transformer changes. The scalar-tensor issue #23158 is owned by part 10. | Part | PR | Change | | --- | --- | --- | | 1 | #23240 | Hardware Vulkan CI | | 2 | #23241 | Scalar cache type and signed-zero keys | | 3 | #23242 | Vulkan-local signed-zero serialization | | 4 | #23243 | GELU modes and view kwargs | | 5 | #23244 | Reduction clamp, NaN, and FP16 rounding | | 6 | #23245 | Reduction dimension guards | | 7 | #23246 | Bool staging and logical_not | | 8 | #23247 | Scalar representability and symbolic guards | | 9 | #23248 | 64-bit dtype and fusion policy | | 10 | #23249 | scalar_tensor with exact integer values | | 11 | #23250 | Typed, resizable full | | 12 | #23251 | Power special values and logical FP16 dtype | | 13 | #23252 | mul.Scalar | | 14 | #23253 | any.dim | | 15 | #23254 | Dynamic expand and transformer integration | The stack retains the FACTO and ATen conformance fixes and their regression tests: typed integer fills, NaN and signed-zero behavior, reduction range and accumulation fixes, safe scalar fallbacks, and the FP16 power contract. FP16 writes use explicit nearest-even conversion, and power keeps the requested tensor dtype when a device emulates FP16 storage with FP32. The test module is now `test_vulkan_dynamic.py`; CI and Buck references follow the rename. Each code patch was linted and tested before its original publication. Prior validation of the operator changes passed 116 FACTO cases against both ATen and portable kernels, plus 144 boundary cases against ATen, with texture and buffer preferences: all 520 Vulkan-configuration executions matched ATen, including CPU fallback where unsupported. The additional 239-case sweep covers all 91 available FACTO specifications for registered ATen overloads and 57 targeted examples. Exact-input replays of every candidate Vulkan failure on a separately built main at `a31838280f9309f04af1b375147375b702d29339` found no new regressions or unresolved comparisons. Validation uses the shared `backends/test/` harness on Apple M1 Pro / MoltenVK with Release portable CPU kernels and Vulkan; optimized kernels and XNNPACK are disabled. Buffer preference uses `texture_limits=(1,1,1)`, with actual storage and fallback recorded separately. Single-texel tensors can still select textures. FACTO revision: `3b8c778c99766a8b4d0d04563ae0b16cbb276829`, seed 0. The complete native regression run on the previously published head `a2543b8a82` passed **50 tests with one expected SwiftShader-only skip** on MoltenVK, covering all 35 dynamic tests, graph-builder and serialization tests, and PT2E quantized linear with downcasting disabled. The subsequent serializer refinement confines `--force-defaults` to Vulkan and restores the shared `exir` FlatBuffers API. That revised source tree passed **26 focused tests**, including shared/Vulkan serialization, signed-zero execution, and power special values. All other code patches are unchanged, and every replacement PR was checked against the tested source trees. Lintrunner and `git diff --check` passed. NVIDIA, SwiftShader, and Windows CI are pending verification on the new ghstack heads. Recreates #23162 through ghstack. Prior review discussion remains on that PR. Authored with OpenAI Codex; split planned with Claude Code. cc @SS-JIA @manuelcandales @digantdesai @cbilgin ghstack-source-id: c6a7e94 ghstack-comment-id: 5892899249 Pull-Request: #23254
[ghstack-poisoned]
[ghstack-poisoned]
mergennachin
added a commit
that referenced
this pull request
Oct 5, 2026
…ormer blocks The dynamic transformer reproductions now lower to one Vulkan delegate across changing sequence lengths. This final PR makes `expand_copy` resizable and adds the eager-attention and SDPA integration tests, including fully masked rows. Fixes #23156. This is part 15 of the original ghstack series. Parts 1–6 (#23240 through #23245) have landed in main; the remaining nine PRs run from #23246 through #23254. Review each PR against its selected base branch. This PR contains the final expansion and transformer changes; part 10 owns the scalar-tensor issue #23158. | Part | PR | Change | | --- | --- | --- | | 1 | #23240 (landed) | Hardware Vulkan CI | | 2 | #23241 (landed) | Scalar cache type and signed-zero keys | | 3 | #23242 (landed) | Vulkan-local signed-zero serialization | | 4 | #23243 (landed) | GELU modes and view kwargs | | 5 | #23244 (landed) | Reduction clamp, NaN, and FP16 rounding | | 6 | #23245 (landed) | Reduction and arg-reduction dimension/storage guards | | 7 | #23246 | Bool staging and logical_not | | 8 | #23247 | Scalar representability and symbolic guards | | 9 | #23248 | 64-bit dtype and fusion policy | | 10 | #23249 | scalar_tensor with exact integer values | | 11 | #23250 | Typed, resizable full | | 12 | #23251 | Power special values and logical FP16 dtype | | 13 | #23252 | mul.Scalar | | 14 | #23253 | any.dim textures with either keepdim setting | | 15 | #23254 | Dynamic expand and transformer integration | The stack retains the FACTO and ATen conformance fixes and their regression tests: typed integer fills, NaN and signed-zero behavior, reduction range and accumulation fixes, safe scalar fallbacks, and the FP16 power contract. On native FP16 devices, the scalar-tensor, full texture, single-dimension texture reduction, and binary scalar shaders round their FP16 outputs to nearest-even. The binary scalar shaders also preserve the requested FP16 dtype when storage is emulated with FP32. Other operators retain their existing FP32 intermediate behavior on those devices. Coverage includes scalar exponents 2.0001 and 2049 and a two-node FP16 mul.Scalar chain in both texture and buffer storage. PR4 introduces `test_vulkan_dynamic.py` and its CI/Buck references. In #23253, `any.dim` now supports both `keepdim` settings on textures. With `keepdim=False`, a reduction into a temporary texture is followed by a GPU view/repack into the squeezed output shape. Both nodes resize dynamically, and empty reduced axes produce false. This adds one temporary texture and one GPU dispatch without introducing a bool-buffer requirement for texture models. Unsupported scalar inputs and 4D batch/channel axes use CPU fallback. Current rebase: Rebased onto main at `91d26b314053` after #23245 landed. All nine remaining code patches are preserved; the only change to the combined source tree is main's Adreno UBO vector-indexing hardening (`b97239a0e4`). All 1153 shader variants compile with glslc, the 15 graph-builder/serialization tests pass, and lintrunner and `git diff --check` pass. Hardware and SwiftShader CI will rerun on the updated ghstack heads. Prior texture-path validation on main at `903cef063774`: the Release Vulkan/portable runtime built and eleven focused native tests passed on Apple M1 Pro / MoltenVK. Coverage included dynamic and chained any reductions, actual texture storage, scalar outputs, singleton and empty dimensions, growth after a zero-length dimension, transformer integration, supported and unsupported 4D reductions, special values, and FP16 rounding. An additional 27 FACTO-generated bool any.dim cases and six boundary cases matched ATen and portable kernels exactly through the shared `backends/test/` harness: 23 ran on Vulkan textures and 10 exercised expected CPU fallback. The subsequent review of #23245 with #23253 verified 784 support/storage combinations and eight mixed dynamic models; all nine review tests passed, with execution results matching ATen and portable kernels. These GPU execution results precede the latest upstream shader-indexing change. Earlier validation of the operator changes passed 116 FACTO cases against both ATen and portable kernels, plus 144 boundary cases against ATen, with texture and buffer preferences: all 520 Vulkan-configuration executions matched ATen, including CPU fallback where unsupported. The additional 239-case sweep covered all 91 available FACTO specifications for registered ATen overloads and 57 targeted examples. Exact-input replays of every candidate Vulkan failure on a separately built main at `a31838280f9309f04af1b375147375b702d29339` found no new regressions or unresolved comparisons. The prior full native run passed 51 tests with one expected SwiftShader-only skip. Subsequent review coverage added exact FP16 rounding boundaries, int32 buffer amax/amin range checks against both ATen and portable kernels, and 28 argmax/argmin execution and fallback cases. These are prior validation results, not a repeat of the full suite for this update. Validation uses Release portable CPU kernels and Vulkan; optimized kernels and XNNPACK are disabled. Buffer preference uses `texture_limits=(1,1,1)`, with actual storage and fallback recorded separately. Single-texel tensors can still select textures. FACTO revision: `3b8c778c99766a8b4d0d04563ae0b16cbb276829`, seed 0. Recreates #23162 through ghstack. Prior review discussion remains on that PR. Authored with OpenAI Codex; split planned with Claude Code. cc @SS-JIA @manuelcandales @digantdesai @cbilgin ghstack-source-id: 4c0dff7 ghstack-comment-id: 5892899249 Pull-Request: #23254
mergennachin
marked this pull request as ready for review
October 5, 2026 01:29
SS-JIA
approved these changes
Oct 6, 2026
Contributor
|
LGTM, but there appears to be some merge conflicts; most likely those can be resolved by updating the base branch to point to the most recent |
[ghstack-poisoned]
mergennachin
added a commit
that referenced
this pull request
Oct 7, 2026
…ormer blocks The dynamic transformer reproductions now lower to one Vulkan delegate across changing sequence lengths. This final PR makes `expand_copy` resizable and adds the eager-attention and SDPA integration tests, including fully masked rows. Fixes #23156. This is part 15 of the original ghstack series. Parts 1–7 (#23240 through #23246) have landed in main; the remaining eight PRs run from #23247 through #23254. Review each PR against its selected base branch. This PR contains the final expansion and transformer changes; part 10 owns the scalar-tensor issue #23158. | Part | PR | Change | | --- | --- | --- | | 1 | #23240 (landed) | Hardware Vulkan CI | | 2 | #23241 (landed) | Scalar cache type and signed-zero keys | | 3 | #23242 (landed) | Vulkan-local signed-zero serialization | | 4 | #23243 (landed) | GELU modes and view kwargs | | 5 | #23244 (landed) | Reduction clamp, NaN, and FP16 rounding | | 6 | #23245 (landed) | Reduction and arg-reduction dimension/storage guards | | 7 | #23246 (landed) | Bool staging and logical_not | | 8 | #23247 | Scalar representability and symbolic guards | | 9 | #23248 | 64-bit dtype and fusion policy | | 10 | #23249 | scalar_tensor with exact integer values | | 11 | #23250 | Typed, resizable full | | 12 | #23251 | Power special values and logical FP16 dtype | | 13 | #23252 | mul.Scalar | | 14 | #23253 | any.dim textures with either keepdim setting | | 15 | #23254 | Dynamic expand and transformer integration | The stack retains the FACTO and ATen conformance fixes and their regression tests: typed integer fills, NaN and signed-zero behavior, reduction range and accumulation fixes, safe scalar fallbacks, and the FP16 power contract. On native FP16 devices, the scalar-tensor, full texture, single-dimension texture reduction, and binary scalar shaders round their FP16 outputs to nearest-even. The binary scalar shaders also preserve the requested FP16 dtype when storage is emulated with FP32. Other operators retain their existing FP32 intermediate behavior on those devices. Coverage includes scalar exponents 2.0001 and 2049 and a two-node FP16 mul.Scalar chain in both texture and buffer storage. PR4 introduces `test_vulkan_dynamic.py` and its CI/Buck references. In #23253, `any.dim` now supports both `keepdim` settings on textures. With `keepdim=False`, a reduction into a temporary texture is followed by a GPU view/repack into the squeezed output shape. Both nodes resize dynamically, and empty reduced axes produce false. This adds one temporary texture and one GPU dispatch without introducing a bool-buffer requirement for texture models. Unsupported scalar inputs and 4D batch/channel axes use CPU fallback. Current rebase: Rebased onto main at `5f72739cf26e` after #23246 landed. All eight remaining code patches are unchanged, and the rebase had no conflicts. All 1153 shader variants compile with glslc, the 15 graph-builder/serialization tests pass, and lintrunner and `git diff --check` pass. Native GPU execution was validated on an earlier base; hardware and SwiftShader CI will rerun on the updated ghstack heads. Prior texture-path validation on main at `903cef063774`: the Release Vulkan/portable runtime built and eleven focused native tests passed on Apple M1 Pro / MoltenVK. Coverage included dynamic and chained any reductions, actual texture storage, scalar outputs, singleton and empty dimensions, growth after a zero-length dimension, transformer integration, supported and unsupported 4D reductions, special values, and FP16 rounding. An additional 27 FACTO-generated bool any.dim cases and six boundary cases matched ATen and portable kernels exactly through the shared `backends/test/` harness: 23 ran on Vulkan textures and 10 exercised expected CPU fallback. The subsequent review of #23245 with #23253 verified 784 support/storage combinations and eight mixed dynamic models; all nine review tests passed, with execution results matching ATen and portable kernels. These GPU execution results precede the latest upstream shader-indexing change. Earlier validation of the operator changes passed 116 FACTO cases against both ATen and portable kernels, plus 144 boundary cases against ATen, with texture and buffer preferences: all 520 Vulkan-configuration executions matched ATen, including CPU fallback where unsupported. The additional 239-case sweep covered all 91 available FACTO specifications for registered ATen overloads and 57 targeted examples. Exact-input replays of every candidate Vulkan failure on a separately built main at `a31838280f9309f04af1b375147375b702d29339` found no new regressions or unresolved comparisons. The prior full native run passed 51 tests with one expected SwiftShader-only skip. Subsequent review coverage added exact FP16 rounding boundaries, int32 buffer amax/amin range checks against both ATen and portable kernels, and 28 argmax/argmin execution and fallback cases. These are prior validation results, not a repeat of the full suite for this update. Validation uses Release portable CPU kernels and Vulkan; optimized kernels and XNNPACK are disabled. Buffer preference uses `texture_limits=(1,1,1)`, with actual storage and fallback recorded separately. Single-texel tensors can still select textures. FACTO revision: `3b8c778c99766a8b4d0d04563ae0b16cbb276829`, seed 0. Recreates #23162 through ghstack. Prior review discussion remains on that PR. Authored with OpenAI Codex; split planned with Claude Code. cc @SS-JIA @manuelcandales @digantdesai @cbilgin ghstack-source-id: e508cde ghstack-comment-id: 5892899249 Pull-Request: #23254
This PR needs a
|
mergennachin
added a commit
that referenced
this pull request
Oct 7, 2026
…ormer blocks The dynamic transformer reproductions now lower to one Vulkan delegate across changing sequence lengths. This final PR makes `expand_copy` resizable and adds the eager-attention and SDPA integration tests, including fully masked rows. Fixes #23156. This is part 15 of the original ghstack series. Parts 1–7 (#23240 through #23246) have landed in main; the remaining eight PRs run from #23247 through #23254. Review each PR against its selected base branch. This PR contains the final expansion and transformer changes; part 10 owns the scalar-tensor issue #23158. | Part | PR | Change | | --- | --- | --- | | 1 | #23240 (landed) | Hardware Vulkan CI | | 2 | #23241 (landed) | Scalar cache type and signed-zero keys | | 3 | #23242 (landed) | Vulkan-local signed-zero serialization | | 4 | #23243 (landed) | GELU modes and view kwargs | | 5 | #23244 (landed) | Reduction clamp, NaN, and FP16 rounding | | 6 | #23245 (landed) | Reduction and arg-reduction dimension/storage guards | | 7 | #23246 (landed) | Bool staging and logical_not | | 8 | #23247 | Scalar representability and symbolic guards | | 9 | #23248 | 64-bit dtype and fusion policy | | 10 | #23249 | scalar_tensor with exact integer values | | 11 | #23250 | Typed, resizable full | | 12 | #23251 | Power special values and logical FP16 dtype | | 13 | #23252 | mul.Scalar | | 14 | #23253 | any.dim textures with either keepdim setting | | 15 | #23254 | Dynamic expand and transformer integration | The stack retains the FACTO and ATen conformance fixes and their regression tests: typed integer fills, NaN and signed-zero behavior, reduction range and accumulation fixes, safe scalar fallbacks, and the FP16 power contract. On native FP16 devices, the scalar-tensor, full texture, single-dimension texture reduction, and binary scalar shaders round their FP16 outputs to nearest-even. The binary scalar shaders also preserve the requested FP16 dtype when storage is emulated with FP32. Other operators retain their existing FP32 intermediate behavior on those devices. Coverage includes scalar exponents 2.0001 and 2049 and a two-node FP16 mul.Scalar chain in both texture and buffer storage. PR4 introduces `test_vulkan_dynamic.py` and its CI/Buck references. In #23253, `any.dim` now supports both `keepdim` settings on textures. With `keepdim=False`, a reduction into a temporary texture is followed by a GPU view/repack into the squeezed output shape. Both nodes resize dynamically, and empty reduced axes produce false. This adds one temporary texture and one GPU dispatch without introducing a bool-buffer requirement for texture models. Unsupported scalar inputs and 4D batch/channel axes use CPU fallback. Current rebase: Rebased onto main at `5f72739cf26e` after #23246 landed. All eight remaining code patches are unchanged, and the rebase had no conflicts. All 1153 shader variants compile with glslc, the 15 graph-builder/serialization tests pass, and lintrunner and `git diff --check` pass. Native GPU execution was validated on an earlier base; hardware and SwiftShader CI will rerun on the updated ghstack heads. Prior texture-path validation on main at `903cef063774`: the Release Vulkan/portable runtime built and eleven focused native tests passed on Apple M1 Pro / MoltenVK. Coverage included dynamic and chained any reductions, actual texture storage, scalar outputs, singleton and empty dimensions, growth after a zero-length dimension, transformer integration, supported and unsupported 4D reductions, special values, and FP16 rounding. An additional 27 FACTO-generated bool any.dim cases and six boundary cases matched ATen and portable kernels exactly through the shared `backends/test/` harness: 23 ran on Vulkan textures and 10 exercised expected CPU fallback. The subsequent review of #23245 with #23253 verified 784 support/storage combinations and eight mixed dynamic models; all nine review tests passed, with execution results matching ATen and portable kernels. These GPU execution results precede the latest upstream shader-indexing change. Earlier validation of the operator changes passed 116 FACTO cases against both ATen and portable kernels, plus 144 boundary cases against ATen, with texture and buffer preferences: all 520 Vulkan-configuration executions matched ATen, including CPU fallback where unsupported. The additional 239-case sweep covered all 91 available FACTO specifications for registered ATen overloads and 57 targeted examples. Exact-input replays of every candidate Vulkan failure on a separately built main at `a31838280f9309f04af1b375147375b702d29339` found no new regressions or unresolved comparisons. The prior full native run passed 51 tests with one expected SwiftShader-only skip. Subsequent review coverage added exact FP16 rounding boundaries, int32 buffer amax/amin range checks against both ATen and portable kernels, and 28 argmax/argmin execution and fallback cases. These are prior validation results, not a repeat of the full suite for this update. Validation uses Release portable CPU kernels and Vulkan; optimized kernels and XNNPACK are disabled. Buffer preference uses `texture_limits=(1,1,1)`, with actual storage and fallback recorded separately. Single-texel tensors can still select textures. FACTO revision: `3b8c778c99766a8b4d0d04563ae0b16cbb276829`, seed 0. Recreates #23162 through ghstack. Prior review discussion remains on that PR. Authored with OpenAI Codex; split planned with Claude Code. cc @SS-JIA @manuelcandales @digantdesai @cbilgin ghstack-source-id: fe79288 ghstack-comment-id: 5892899249 Pull-Request: #23254
mergennachin
added a commit
that referenced
this pull request
Oct 7, 2026
…ormer blocks The dynamic transformer reproductions now lower to one Vulkan delegate across changing sequence lengths. This final PR makes `expand_copy` resizable and adds the eager-attention and SDPA integration tests, including fully masked rows. Fixes #23156. This is part 15 of the original ghstack series. Parts 1–7 (#23240 through #23246) have landed in main; the remaining eight PRs run from #23247 through #23254. Review each PR against its selected base branch. This PR contains the final expansion and transformer changes; part 10 owns the scalar-tensor issue #23158. | Part | PR | Change | | --- | --- | --- | | 1 | #23240 (landed) | Hardware Vulkan CI | | 2 | #23241 (landed) | Scalar cache type and signed-zero keys | | 3 | #23242 (landed) | Vulkan-local signed-zero serialization | | 4 | #23243 (landed) | GELU modes and view kwargs | | 5 | #23244 (landed) | Reduction clamp, NaN, and FP16 rounding | | 6 | #23245 (landed) | Reduction and arg-reduction dimension/storage guards | | 7 | #23246 (landed) | Bool staging and logical_not | | 8 | #23247 | Scalar representability and symbolic guards | | 9 | #23248 | 64-bit dtype and fusion policy | | 10 | #23249 | scalar_tensor with exact integer values | | 11 | #23250 | Typed, resizable full | | 12 | #23251 | Power special values and logical FP16 dtype | | 13 | #23252 | mul.Scalar | | 14 | #23253 | any.dim textures with either keepdim setting | | 15 | #23254 | Dynamic expand and transformer integration | The stack retains the FACTO and ATen conformance fixes and their regression tests: typed integer fills, NaN and signed-zero behavior, reduction range and accumulation fixes, safe scalar fallbacks, and the FP16 power contract. On native FP16 devices, the scalar-tensor, full texture, single-dimension texture reduction, and binary scalar shaders round their FP16 outputs to nearest-even. The binary scalar shaders also preserve the requested FP16 dtype when storage is emulated with FP32. Other operators retain their existing FP32 intermediate behavior on those devices. Coverage includes scalar exponents 2.0001 and 2049 and a two-node FP16 mul.Scalar chain in both texture and buffer storage. PR4 introduces `test_vulkan_dynamic.py` and its CI/Buck references. In #23253, `any.dim` now supports both `keepdim` settings on textures. With `keepdim=False`, a reduction into a temporary texture is followed by a GPU view/repack into the squeezed output shape. Both nodes resize dynamically, and empty reduced axes produce false. This adds one temporary texture and one GPU dispatch without introducing a bool-buffer requirement for texture models. Unsupported scalar inputs and 4D batch/channel axes use CPU fallback. Current rebase: Rebased onto main at `5f72739cf26e` after #23246 landed. All eight remaining code patches are unchanged, and the rebase had no conflicts. All 1153 shader variants compile with glslc, the 15 graph-builder/serialization tests pass, and lintrunner and `git diff --check` pass. Native GPU execution was validated on an earlier base; hardware and SwiftShader CI will rerun on the updated ghstack heads. Prior texture-path validation on main at `903cef063774`: the Release Vulkan/portable runtime built and eleven focused native tests passed on Apple M1 Pro / MoltenVK. Coverage included dynamic and chained any reductions, actual texture storage, scalar outputs, singleton and empty dimensions, growth after a zero-length dimension, transformer integration, supported and unsupported 4D reductions, special values, and FP16 rounding. An additional 27 FACTO-generated bool any.dim cases and six boundary cases matched ATen and portable kernels exactly through the shared `backends/test/` harness: 23 ran on Vulkan textures and 10 exercised expected CPU fallback. The subsequent review of #23245 with #23253 verified 784 support/storage combinations and eight mixed dynamic models; all nine review tests passed, with execution results matching ATen and portable kernels. These GPU execution results precede the latest upstream shader-indexing change. Earlier validation of the operator changes passed 116 FACTO cases against both ATen and portable kernels, plus 144 boundary cases against ATen, with texture and buffer preferences: all 520 Vulkan-configuration executions matched ATen, including CPU fallback where unsupported. The additional 239-case sweep covered all 91 available FACTO specifications for registered ATen overloads and 57 targeted examples. Exact-input replays of every candidate Vulkan failure on a separately built main at `a31838280f9309f04af1b375147375b702d29339` found no new regressions or unresolved comparisons. The prior full native run passed 51 tests with one expected SwiftShader-only skip. Subsequent review coverage added exact FP16 rounding boundaries, int32 buffer amax/amin range checks against both ATen and portable kernels, and 28 argmax/argmin execution and fallback cases. These are prior validation results, not a repeat of the full suite for this update. Validation uses Release portable CPU kernels and Vulkan; optimized kernels and XNNPACK are disabled. Buffer preference uses `texture_limits=(1,1,1)`, with actual storage and fallback recorded separately. Single-texel tensors can still select textures. FACTO revision: `3b8c778c99766a8b4d0d04563ae0b16cbb276829`, seed 0. Recreates #23162 through ghstack. Prior review discussion remains on that PR. Authored with OpenAI Codex; split planned with Claude Code. cc @SS-JIA @manuelcandales @digantdesai @cbilgin ghstack-source-id: a7c2766 ghstack-comment-id: 5892899249 Pull-Request: #23254
mergennachin
added a commit
that referenced
this pull request
Oct 7, 2026
With downcast_64_bit disabled, the graph builder refuses int64 and float64 tensors, but the partitioner still claimed those nodes, so lowering failed late or the delegate aborted at load. The partitioner now rejects nodes that need 64-bit dtypes in that mode, and keeps a fusable pattern entirely on CPU when it touches a 64-bit non-constant tensor rather than splitting it. Constant parameters are exempt because quantization parameters are folded away during fusion. Part 9/15 of the Vulkan transformer and operator-conformance stack. Depends on #23247; review against the selected base branch. Integration PR: #23254. Validation: 4 passed, 2 warnings in 32.37s. Lintrunner and git diff --check pass. Native tests use MoltenVK with portable CPU kernels; hardware and SwiftShader CI are pending. Recreates #23209 through ghstack. Prior review discussion remains on that PR. Authored with OpenAI Codex; split planned with Claude Code. cc @SS-JIA @manuelcandales @digantdesai @cbilgin
mergennachin
added a commit
that referenced
this pull request
Oct 7, 2026
…ormer blocks The dynamic transformer reproductions now lower to one Vulkan delegate across changing sequence lengths. This final PR makes `expand_copy` resizable and adds the eager-attention and SDPA integration tests, including fully masked rows. Fixes #23156. This is part 15 of the original ghstack series. Parts 1–7 (#23240 through #23246) have landed in main; the remaining eight PRs run from #23247 through #23254. Review each PR against its selected base branch. This PR contains the final expansion and transformer changes; part 10 owns the scalar-tensor issue #23158. | Part | PR | Change | | --- | --- | --- | | 1 | #23240 (landed) | Hardware Vulkan CI | | 2 | #23241 (landed) | Scalar cache type and signed-zero keys | | 3 | #23242 (landed) | Vulkan-local signed-zero serialization | | 4 | #23243 (landed) | GELU modes and view kwargs | | 5 | #23244 (landed) | Reduction clamp, NaN, and FP16 rounding | | 6 | #23245 (landed) | Reduction and arg-reduction dimension/storage guards | | 7 | #23246 (landed) | Bool staging and logical_not | | 8 | #23247 | Scalar representability and symbolic guards | | 9 | #23248 | 64-bit dtype and fusion policy | | 10 | #23249 | scalar_tensor with exact integer values | | 11 | #23250 | Typed, resizable full | | 12 | #23251 | Power special values and logical FP16 dtype | | 13 | #23252 | mul.Scalar | | 14 | #23253 | any.dim textures with either keepdim setting | | 15 | #23254 | Dynamic expand and transformer integration | The stack retains the FACTO and ATen conformance fixes and their regression tests: typed integer fills, NaN and signed-zero behavior, reduction range and accumulation fixes, safe scalar fallbacks, and the FP16 power contract. On native FP16 devices, the scalar-tensor, full texture, single-dimension texture reduction, and binary scalar shaders round their FP16 outputs to nearest-even. The binary scalar shaders also preserve the requested FP16 dtype when storage is emulated with FP32. Other operators retain their existing FP32 intermediate behavior on those devices. Coverage includes scalar exponents 2.0001 and 2049 and a two-node FP16 mul.Scalar chain in both texture and buffer storage. PR4 introduces `test_vulkan_dynamic.py` and its CI/Buck references. In #23253, `any.dim` now supports both `keepdim` settings on textures. With `keepdim=False`, a reduction into a temporary texture is followed by a GPU view/repack into the squeezed output shape. Both nodes resize dynamically, and empty reduced axes produce false. This adds one temporary texture and one GPU dispatch without introducing a bool-buffer requirement for texture models. Unsupported scalar inputs and 4D batch/channel axes use CPU fallback. Current rebase: Rebased onto main at `5f72739cf26e` after #23246 landed. All eight remaining code patches are unchanged, and the rebase had no conflicts. All 1153 shader variants compile with glslc, the 15 graph-builder/serialization tests pass, and lintrunner and `git diff --check` pass. Native GPU execution was validated on an earlier base; hardware and SwiftShader CI will rerun on the updated ghstack heads. Prior texture-path validation on main at `903cef063774`: the Release Vulkan/portable runtime built and eleven focused native tests passed on Apple M1 Pro / MoltenVK. Coverage included dynamic and chained any reductions, actual texture storage, scalar outputs, singleton and empty dimensions, growth after a zero-length dimension, transformer integration, supported and unsupported 4D reductions, special values, and FP16 rounding. An additional 27 FACTO-generated bool any.dim cases and six boundary cases matched ATen and portable kernels exactly through the shared `backends/test/` harness: 23 ran on Vulkan textures and 10 exercised expected CPU fallback. The subsequent review of #23245 with #23253 verified 784 support/storage combinations and eight mixed dynamic models; all nine review tests passed, with execution results matching ATen and portable kernels. These GPU execution results precede the latest upstream shader-indexing change. Earlier validation of the operator changes passed 116 FACTO cases against both ATen and portable kernels, plus 144 boundary cases against ATen, with texture and buffer preferences: all 520 Vulkan-configuration executions matched ATen, including CPU fallback where unsupported. The additional 239-case sweep covered all 91 available FACTO specifications for registered ATen overloads and 57 targeted examples. Exact-input replays of every candidate Vulkan failure on a separately built main at `a31838280f9309f04af1b375147375b702d29339` found no new regressions or unresolved comparisons. The prior full native run passed 51 tests with one expected SwiftShader-only skip. Subsequent review coverage added exact FP16 rounding boundaries, int32 buffer amax/amin range checks against both ATen and portable kernels, and 28 argmax/argmin execution and fallback cases. These are prior validation results, not a repeat of the full suite for this update. Validation uses Release portable CPU kernels and Vulkan; optimized kernels and XNNPACK are disabled. Buffer preference uses `texture_limits=(1,1,1)`, with actual storage and fallback recorded separately. Single-texel tensors can still select textures. FACTO revision: `3b8c778c99766a8b4d0d04563ae0b16cbb276829`, seed 0. Recreates #23162 through ghstack. Prior review discussion remains on that PR. Authored with OpenAI Codex; split planned with Claude Code. cc @SS-JIA @manuelcandales @digantdesai @cbilgin ghstack-source-id: 6b879b2 ghstack-comment-id: 5892899249 Pull-Request: #23254
mergennachin
added a commit
that referenced
this pull request
Oct 7, 2026
…ormer blocks The dynamic transformer reproductions now lower to one Vulkan delegate across changing sequence lengths. This final PR makes `expand_copy` resizable and adds the eager-attention and SDPA integration tests, including fully masked rows. Fixes #23156. This is part 15 of the original ghstack series. Parts 1–7 (#23240 through #23246) have landed in main; the remaining eight PRs run from #23247 through #23254. Review each PR against its selected base branch. This PR contains the final expansion and transformer changes; part 10 owns the scalar-tensor issue #23158. | Part | PR | Change | | --- | --- | --- | | 1 | #23240 (landed) | Hardware Vulkan CI | | 2 | #23241 (landed) | Scalar cache type and signed-zero keys | | 3 | #23242 (landed) | Vulkan-local signed-zero serialization | | 4 | #23243 (landed) | GELU modes and view kwargs | | 5 | #23244 (landed) | Reduction clamp, NaN, and FP16 rounding | | 6 | #23245 (landed) | Reduction and arg-reduction dimension/storage guards | | 7 | #23246 (landed) | Bool staging and logical_not | | 8 | #23247 | Scalar representability and symbolic guards | | 9 | #23248 | 64-bit dtype and fusion policy | | 10 | #23249 | scalar_tensor with exact integer values | | 11 | #23250 | Typed, resizable full | | 12 | #23251 | Power special values and logical FP16 dtype | | 13 | #23252 | mul.Scalar | | 14 | #23253 | any.dim textures with either keepdim setting | | 15 | #23254 | Dynamic expand and transformer integration | The stack retains the FACTO and ATen conformance fixes and their regression tests: typed integer fills, NaN and signed-zero behavior, reduction range and accumulation fixes, safe scalar fallbacks, and the FP16 power contract. On native FP16 devices, the scalar-tensor, full texture, single-dimension texture reduction, and binary scalar shaders round their FP16 outputs to nearest-even. The binary scalar shaders also preserve the requested FP16 dtype when storage is emulated with FP32. Other operators retain their existing FP32 intermediate behavior on those devices. Coverage includes scalar exponents 2.0001 and 2049 and a two-node FP16 mul.Scalar chain in both texture and buffer storage. PR4 introduces `test_vulkan_dynamic.py` and its CI/Buck references. In #23253, `any.dim` now supports both `keepdim` settings on textures. With `keepdim=False`, a reduction into a temporary texture is followed by a GPU view/repack into the squeezed output shape. Both nodes resize dynamically, and empty reduced axes produce false. This adds one temporary texture and one GPU dispatch without introducing a bool-buffer requirement for texture models. Unsupported scalar inputs and 4D batch/channel axes use CPU fallback. Current rebase: Rebased onto main at `5f72739cf26e` after #23246 landed. All eight remaining code patches are unchanged, and the rebase had no conflicts. All 1153 shader variants compile with glslc, the 15 graph-builder/serialization tests pass, and lintrunner and `git diff --check` pass. Native GPU execution was validated on an earlier base; hardware and SwiftShader CI will rerun on the updated ghstack heads. Prior texture-path validation on main at `903cef063774`: the Release Vulkan/portable runtime built and eleven focused native tests passed on Apple M1 Pro / MoltenVK. Coverage included dynamic and chained any reductions, actual texture storage, scalar outputs, singleton and empty dimensions, growth after a zero-length dimension, transformer integration, supported and unsupported 4D reductions, special values, and FP16 rounding. An additional 27 FACTO-generated bool any.dim cases and six boundary cases matched ATen and portable kernels exactly through the shared `backends/test/` harness: 23 ran on Vulkan textures and 10 exercised expected CPU fallback. The subsequent review of #23245 with #23253 verified 784 support/storage combinations and eight mixed dynamic models; all nine review tests passed, with execution results matching ATen and portable kernels. These GPU execution results precede the latest upstream shader-indexing change. Earlier validation of the operator changes passed 116 FACTO cases against both ATen and portable kernels, plus 144 boundary cases against ATen, with texture and buffer preferences: all 520 Vulkan-configuration executions matched ATen, including CPU fallback where unsupported. The additional 239-case sweep covered all 91 available FACTO specifications for registered ATen overloads and 57 targeted examples. Exact-input replays of every candidate Vulkan failure on a separately built main at `a31838280f9309f04af1b375147375b702d29339` found no new regressions or unresolved comparisons. The prior full native run passed 51 tests with one expected SwiftShader-only skip. Subsequent review coverage added exact FP16 rounding boundaries, int32 buffer amax/amin range checks against both ATen and portable kernels, and 28 argmax/argmin execution and fallback cases. These are prior validation results, not a repeat of the full suite for this update. Validation uses Release portable CPU kernels and Vulkan; optimized kernels and XNNPACK are disabled. Buffer preference uses `texture_limits=(1,1,1)`, with actual storage and fallback recorded separately. Single-texel tensors can still select textures. FACTO revision: `3b8c778c99766a8b4d0d04563ae0b16cbb276829`, seed 0. Recreates #23162 through ghstack. Prior review discussion remains on that PR. Authored with OpenAI Codex; split planned with Claude Code. cc @SS-JIA @manuelcandales @digantdesai @cbilgin ghstack-source-id: 0705e5f ghstack-comment-id: 5892899249 Pull-Request: #23254
SS-JIA
added a commit
that referenced
this pull request
Oct 7, 2026
Summary: ## Ulterior Motive Restore the "Test WebGPU Native (Dawn)" workflow on `main`, which has failed on every commit since "[Vulkan] Keep unrepresentable and symbolic scalars out of the delegate (#23247)". ## Rationale **What**: The Vulkan partitioner no longer applies its SymBool/SymFloat rejection to assert ops (`_assert_scalar`, `sym_constrain_range_for_size`) or to nodes whose only users are assert ops. **Why**: #23247 rejects every node that produces or consumes a SymBool. Models that guard an `.item()` value with `torch._check` produce `u <= N` / `u >= 0` comparisons feeding `_assert_scalar`. Rejecting them cascades through the new "symbolic scalar has an unsupported consumer" check to the `_local_scalar_dense` producer, splitting the graph into multiple delegates. `test_rope_hf.py` then fails with `AssertionError: expected one delegate, got 2`. Asserts are stripped by `RemoveAssertsTransform` during preprocessing, so these SymBool values never need a runtime representation. ## Details - Assert-only SymBool chains - **Before**: Comparisons and asserts rejected; `.item()` producer rejected as a consequence; graph split into multiple delegates. - **After**: Delegated, then removed during preprocessing; graph stays in one delegate. - SymBool/SymFloat values feeding non-assert ops - **Before**: Rejected. - **After**: Unchanged (still rejected). Test Plan: ## Observed Results | Validation | Device | Result | |---|---|---| | Vulkan partition count for `DynamicHfRope` (`test_rope_hf.py`), parent commit | macOS (Apple Silicon) host, partitioner only | 2 partitions; skip logs match the CI failure | | Vulkan partition count for `DynamicHfRope`, this commit | macOS (Apple Silicon) host, partitioner only | 1 partition, no skipped nodes | | `test_rope_hf.py` end-to-end ("Test WebGPU Native (Dawn)") | — | Not run: needs the Dawn/SwiftShader runtime; relying on CI | | `backends/vulkan/test/test_vulkan_dynamic.py` | — | Not run: needs the Vulkan runtime; relying on CI | | Lint | — | Not run: lint tooling not installed locally; relying on CI | ## Reproduction Guide ### Setup ```bash git clone https://github.com/pytorch/executorch.git cd executorch ./install_executorch.sh ``` ### Count Vulkan partitions for the dynamic HF RoPE model ```bash cat > /tmp/count_partitions.py <<'PY' import torch from executorch.exir import to_edge from executorch.backends.vulkan.partitioner.vulkan_partitioner import VulkanPartitioner from executorch.backends.webgpu.test.ops import test_rope_hf as t inputs = t._dynamic_inputs() with torch._dynamo.config.patch(capture_scalar_outputs=True): ep = torch.export.export(t.DynamicHfRope().eval(), inputs) res = VulkanPartitioner().partition(to_edge(ep).exported_program()) tags = { n.meta["delegation_tag"] for n in res.tagged_exported_program.graph.nodes if n.meta.get("delegation_tag") } print("partitions:", len(tags)) PY python /tmp/count_partitions.py 2>&1 | grep -E "Vulkan Partitioner|partitions:" ``` Run once on this commit (expect `partitions: 1`) and once on its parent (`git checkout HEAD~1`, expect `partitions: 2` plus "symbolic float or bool values are not supported" skip logs). Reviewers: This PR was authored with Claude Code.
This branch was previously deployed
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
The binary-scalar and comparison kernels receive scalars as int32 or float, yet the partitioner accepted NaN, integers outside int32 and symbolic scalars, which lowered with wrong values or failed in the graph builder. is_scalar_value_supported rejects them for pow.Tensor_Scalar and the comparison scalar ops. The partitioner also rejects SymFloat and SymBool values and keeps SymInt producers on CPU when a consumer is unsupported; node support is memoized so that check stays linear.
Part 8/15 of the Vulkan transformer and operator-conformance stack. Depends on #23246; review against the selected base branch. Integration PR: #23254.
Validation: 5 passed, 2 warnings in 150.36s (0:02:30). Lintrunner and git diff --check pass. Native tests use MoltenVK with portable CPU kernels; hardware and SwiftShader CI are pending.
Recreates #23208 through ghstack. Prior review discussion remains on that PR.
Authored with OpenAI Codex; split planned with Claude Code.
cc @SS-JIA @manuelcandales @digantdesai @cbilgin