[ExecuTorch][WebGPU] Add dynamic-shape resize hook to cat (output mismatch) - #21232
Merged
Conversation
🔗 Helpful Links🧪 See artifacts and rendered test results at hud.pytorch.org/pr/pytorch/executorch/21232
Note: Links to docs will display an error until the docs builds have been completed. ⏳ No Failures, 46 PendingAs of commit 38272a1 with merge base 4a26c64 ( This comment was automatically generated by Dr. CI and updates every 15 minutes. |
This was referenced Jul 22, 2026
JCNTH
added a commit
that referenced
this pull request
Jul 24, 2026
…match) Pull Request resolved: #21232 Under dynamic shapes `cat` produced a wrong (scrambled + tail-garbage) output. `cat_impl` baked every shape-dependent quantity - each input `in_meta` (strides/numel), the shared `out_meta` (strides), the per-input `off_k`, and each dispatch's `workgroup_count` - from the MAX (upper-bound) build shape and registered NO resize hook, then released the UBOs. Under a smaller live shape the kernel then decoded coords with max strides, scattered with max out-strides, and looped over the max numel. This is the `[dynamic]` output-mismatch behind `test_inception_v3`, `test_squeezenet1_1`, `test_densenet161` (all channel-cat on spatial-dynamic feature maps; the concat dim is static so `off_k` was already correct - the defect is the stale spatial strides/numel). Fix (mirrors the WebGPUGraph SwiGLU/QKV resize templates and the shipped `mul` hook): - Keep the per-input `in_meta`/`params` UBOs and the shared `out_meta` UBO alive via `own_uniform_buffer` (previously released after build); collect their handles plus each `add_dispatch` index. - Register an idempotent `add_tensor_resize_hook` on every input id: from `cur_dims` recompute live out dims (`set_cur_dims(out_id, ...)` to cascade to consumers + fix the delegate-output shape), rebuild `out_meta` + each input's `in_meta`/`params` and `wgpuQueueWriteBuffer` them, and rewrite each dispatch's `workgroup_count_x` via `compute_1d_workgroup_count`. On a static graph `cur_dims == dims`, so the hook rewrites identical values (no behavior change). Applied identically to both the `xplat/` and `fbcode/` mirrors (byte-identical). Co-authored-with: Claude Code. ghstack-source-id: 406388769 @exported-using-ghexport Differential Revision: [D113319596](https://our.internmc.facebook.com/intern/diff/D113319596/)
JCNTH
added a commit
that referenced
this pull request
Jul 24, 2026
…match) Pull Request resolved: #21232 Under dynamic shapes `cat` produced a wrong (scrambled + tail-garbage) output. `cat_impl` baked every shape-dependent quantity - each input `in_meta` (strides/numel), the shared `out_meta` (strides), the per-input `off_k`, and each dispatch's `workgroup_count` - from the MAX (upper-bound) build shape and registered NO resize hook, then released the UBOs. Under a smaller live shape the kernel then decoded coords with max strides, scattered with max out-strides, and looped over the max numel. This is the `[dynamic]` output-mismatch behind `test_inception_v3`, `test_squeezenet1_1`, `test_densenet161` (all channel-cat on spatial-dynamic feature maps; the concat dim is static so `off_k` was already correct - the defect is the stale spatial strides/numel). Fix (mirrors the WebGPUGraph SwiGLU/QKV resize templates and the shipped `mul` hook): - Keep the per-input `in_meta`/`params` UBOs and the shared `out_meta` UBO alive via `own_uniform_buffer` (previously released after build); collect their handles plus each `add_dispatch` index. - Register an idempotent `add_tensor_resize_hook` on every input id: from `cur_dims` recompute live out dims (`set_cur_dims(out_id, ...)` to cascade to consumers + fix the delegate-output shape), rebuild `out_meta` + each input's `in_meta`/`params` and `wgpuQueueWriteBuffer` them, and rewrite each dispatch's `workgroup_count_x` via `compute_1d_workgroup_count`. On a static graph `cur_dims == dims`, so the hook rewrites identical values (no behavior change). Applied identically to both the `xplat/` and `fbcode/` mirrors (byte-identical). Co-authored-with: Claude Code. ghstack-source-id: 406388769 @exported-using-ghexport Differential Revision: [D113319596](https://our.internmc.facebook.com/intern/diff/D113319596/)
JCNTH
added a commit
that referenced
this pull request
Jul 24, 2026
…match) Pull Request resolved: #21232 Under dynamic shapes `cat` produced a wrong (scrambled + tail-garbage) output. `cat_impl` baked every shape-dependent quantity - each input `in_meta` (strides/numel), the shared `out_meta` (strides), the per-input `off_k`, and each dispatch's `workgroup_count` - from the MAX (upper-bound) build shape and registered NO resize hook, then released the UBOs. Under a smaller live shape the kernel then decoded coords with max strides, scattered with max out-strides, and looped over the max numel. This is the `[dynamic]` output-mismatch behind `test_inception_v3`, `test_squeezenet1_1`, `test_densenet161` (all channel-cat on spatial-dynamic feature maps; the concat dim is static so `off_k` was already correct - the defect is the stale spatial strides/numel). Fix (mirrors the WebGPUGraph SwiGLU/QKV resize templates and the shipped `mul` hook): - Keep the per-input `in_meta`/`params` UBOs and the shared `out_meta` UBO alive via `own_uniform_buffer` (previously released after build); collect their handles plus each `add_dispatch` index. - Register an idempotent `add_tensor_resize_hook` on every input id: from `cur_dims` recompute live out dims (`set_cur_dims(out_id, ...)` to cascade to consumers + fix the delegate-output shape), rebuild `out_meta` + each input's `in_meta`/`params` and `wgpuQueueWriteBuffer` them, and rewrite each dispatch's `workgroup_count_x` via `compute_1d_workgroup_count`. On a static graph `cur_dims == dims`, so the hook rewrites identical values (no behavior change). Applied identically to both the `xplat/` and `fbcode/` mirrors (byte-identical). Co-authored-with: Claude Code. ghstack-source-id: 406388769 @exported-using-ghexport Differential Revision: [D113319596](https://our.internmc.facebook.com/intern/diff/D113319596/)
JCNTH
added a commit
that referenced
this pull request
Jul 24, 2026
…match) Pull Request resolved: #21232 Under dynamic shapes `cat` produced a wrong (scrambled + tail-garbage) output. `cat_impl` baked every shape-dependent quantity - each input `in_meta` (strides/numel), the shared `out_meta` (strides), the per-input `off_k`, and each dispatch's `workgroup_count` - from the MAX (upper-bound) build shape and registered NO resize hook, then released the UBOs. Under a smaller live shape the kernel then decoded coords with max strides, scattered with max out-strides, and looped over the max numel. This is the `[dynamic]` output-mismatch behind `test_inception_v3`, `test_squeezenet1_1`, `test_densenet161` (all channel-cat on spatial-dynamic feature maps; the concat dim is static so `off_k` was already correct - the defect is the stale spatial strides/numel). Fix (mirrors the WebGPUGraph SwiGLU/QKV resize templates and the shipped `mul` hook): - Keep the per-input `in_meta`/`params` UBOs and the shared `out_meta` UBO alive via `own_uniform_buffer` (previously released after build); collect their handles plus each `add_dispatch` index. - Register an idempotent `add_tensor_resize_hook` on every input id: from `cur_dims` recompute live out dims (`set_cur_dims(out_id, ...)` to cascade to consumers + fix the delegate-output shape), rebuild `out_meta` + each input's `in_meta`/`params` and `wgpuQueueWriteBuffer` them, and rewrite each dispatch's `workgroup_count_x` via `compute_1d_workgroup_count`. On a static graph `cur_dims == dims`, so the hook rewrites identical values (no behavior change). Applied identically to both the `xplat/` and `fbcode/` mirrors (byte-identical). Co-authored-with: Claude Code. ghstack-source-id: 406388769 @exported-using-ghexport Differential Revision: [D113319596](https://our.internmc.facebook.com/intern/diff/D113319596/)
JCNTH
added a commit
that referenced
this pull request
Jul 24, 2026
…match) Pull Request resolved: #21232 Under dynamic shapes `cat` produced a wrong (scrambled + tail-garbage) output. `cat_impl` baked every shape-dependent quantity - each input `in_meta` (strides/numel), the shared `out_meta` (strides), the per-input `off_k`, and each dispatch's `workgroup_count` - from the MAX (upper-bound) build shape and registered NO resize hook, then released the UBOs. Under a smaller live shape the kernel then decoded coords with max strides, scattered with max out-strides, and looped over the max numel. This is the `[dynamic]` output-mismatch behind `test_inception_v3`, `test_squeezenet1_1`, `test_densenet161` (all channel-cat on spatial-dynamic feature maps; the concat dim is static so `off_k` was already correct - the defect is the stale spatial strides/numel). Fix (mirrors the WebGPUGraph SwiGLU/QKV resize templates and the shipped `mul` hook): - Keep the per-input `in_meta`/`params` UBOs and the shared `out_meta` UBO alive via `own_uniform_buffer` (previously released after build); collect their handles plus each `add_dispatch` index. - Register an idempotent `add_tensor_resize_hook` on every input id: from `cur_dims` recompute live out dims (`set_cur_dims(out_id, ...)` to cascade to consumers + fix the delegate-output shape), rebuild `out_meta` + each input's `in_meta`/`params` and `wgpuQueueWriteBuffer` them, and rewrite each dispatch's `workgroup_count_x` via `compute_1d_workgroup_count`. On a static graph `cur_dims == dims`, so the hook rewrites identical values (no behavior change). Applied identically to both the `xplat/` and `fbcode/` mirrors (byte-identical). Co-authored-with: Claude Code. ghstack-source-id: 406388769 @exported-using-ghexport Differential Revision: [D113319596](https://our.internmc.facebook.com/intern/diff/D113319596/)
JCNTH
added a commit
that referenced
this pull request
Jul 24, 2026
…match) Pull Request resolved: #21232 Under dynamic shapes `cat` produced a wrong (scrambled + tail-garbage) output. `cat_impl` baked every shape-dependent quantity - each input `in_meta` (strides/numel), the shared `out_meta` (strides), the per-input `off_k`, and each dispatch's `workgroup_count` - from the MAX (upper-bound) build shape and registered NO resize hook, then released the UBOs. Under a smaller live shape the kernel then decoded coords with max strides, scattered with max out-strides, and looped over the max numel. This is the `[dynamic]` output-mismatch behind `test_inception_v3`, `test_squeezenet1_1`, `test_densenet161` (all channel-cat on spatial-dynamic feature maps; the concat dim is static so `off_k` was already correct - the defect is the stale spatial strides/numel). Fix (mirrors the WebGPUGraph SwiGLU/QKV resize templates and the shipped `mul` hook): - Keep the per-input `in_meta`/`params` UBOs and the shared `out_meta` UBO alive via `own_uniform_buffer` (previously released after build); collect their handles plus each `add_dispatch` index. - Register an idempotent `add_tensor_resize_hook` on every input id: from `cur_dims` recompute live out dims (`set_cur_dims(out_id, ...)` to cascade to consumers + fix the delegate-output shape), rebuild `out_meta` + each input's `in_meta`/`params` and `wgpuQueueWriteBuffer` them, and rewrite each dispatch's `workgroup_count_x` via `compute_1d_workgroup_count`. On a static graph `cur_dims == dims`, so the hook rewrites identical values (no behavior change). Applied identically to both the `xplat/` and `fbcode/` mirrors (byte-identical). Co-authored-with: Claude Code. ghstack-source-id: 406388769 @exported-using-ghexport Differential Revision: [D113319596](https://our.internmc.facebook.com/intern/diff/D113319596/)
JCNTH
added a commit
that referenced
this pull request
Jul 24, 2026
…match) Pull Request resolved: #21232 Under dynamic shapes `cat` produced a wrong (scrambled + tail-garbage) output. `cat_impl` baked every shape-dependent quantity - each input `in_meta` (strides/numel), the shared `out_meta` (strides), the per-input `off_k`, and each dispatch's `workgroup_count` - from the MAX (upper-bound) build shape and registered NO resize hook, then released the UBOs. Under a smaller live shape the kernel then decoded coords with max strides, scattered with max out-strides, and looped over the max numel. This is the `[dynamic]` output-mismatch behind `test_inception_v3`, `test_squeezenet1_1`, `test_densenet161` (all channel-cat on spatial-dynamic feature maps; the concat dim is static so `off_k` was already correct - the defect is the stale spatial strides/numel). Fix (mirrors the WebGPUGraph SwiGLU/QKV resize templates and the shipped `mul` hook): - Keep the per-input `in_meta`/`params` UBOs and the shared `out_meta` UBO alive via `own_uniform_buffer` (previously released after build); collect their handles plus each `add_dispatch` index. - Register an idempotent `add_tensor_resize_hook` on every input id: from `cur_dims` recompute live out dims (`set_cur_dims(out_id, ...)` to cascade to consumers + fix the delegate-output shape), rebuild `out_meta` + each input's `in_meta`/`params` and `wgpuQueueWriteBuffer` them, and rewrite each dispatch's `workgroup_count_x` via `compute_1d_workgroup_count`. On a static graph `cur_dims == dims`, so the hook rewrites identical values (no behavior change). Applied identically to both the `xplat/` and `fbcode/` mirrors (byte-identical). Co-authored-with: Claude Code. ghstack-source-id: 406388769 @exported-using-ghexport Differential Revision: [D113319596](https://our.internmc.facebook.com/intern/diff/D113319596/)
JCNTH
added a commit
that referenced
this pull request
Jul 24, 2026
…match) Pull Request resolved: #21232 Under dynamic shapes `cat` produced a wrong (scrambled + tail-garbage) output. `cat_impl` baked every shape-dependent quantity - each input `in_meta` (strides/numel), the shared `out_meta` (strides), the per-input `off_k`, and each dispatch's `workgroup_count` - from the MAX (upper-bound) build shape and registered NO resize hook, then released the UBOs. Under a smaller live shape the kernel then decoded coords with max strides, scattered with max out-strides, and looped over the max numel. This is the `[dynamic]` output-mismatch behind `test_inception_v3`, `test_squeezenet1_1`, `test_densenet161` (all channel-cat on spatial-dynamic feature maps; the concat dim is static so `off_k` was already correct - the defect is the stale spatial strides/numel). Fix (mirrors the WebGPUGraph SwiGLU/QKV resize templates and the shipped `mul` hook): - Keep the per-input `in_meta`/`params` UBOs and the shared `out_meta` UBO alive via `own_uniform_buffer` (previously released after build); collect their handles plus each `add_dispatch` index. - Register an idempotent `add_tensor_resize_hook` on every input id: from `cur_dims` recompute live out dims (`set_cur_dims(out_id, ...)` to cascade to consumers + fix the delegate-output shape), rebuild `out_meta` + each input's `in_meta`/`params` and `wgpuQueueWriteBuffer` them, and rewrite each dispatch's `workgroup_count_x` via `compute_1d_workgroup_count`. On a static graph `cur_dims == dims`, so the hook rewrites identical values (no behavior change). Applied identically to both the `xplat/` and `fbcode/` mirrors (byte-identical). Co-authored-with: Claude Code. ghstack-source-id: 406388769 @exported-using-ghexport Differential Revision: [D113319596](https://our.internmc.facebook.com/intern/diff/D113319596/)
JCNTH
added a commit
that referenced
this pull request
Jul 24, 2026
…match) Pull Request resolved: #21232 Under dynamic shapes `cat` produced a wrong (scrambled + tail-garbage) output. `cat_impl` baked every shape-dependent quantity - each input `in_meta` (strides/numel), the shared `out_meta` (strides), the per-input `off_k`, and each dispatch's `workgroup_count` - from the MAX (upper-bound) build shape and registered NO resize hook, then released the UBOs. Under a smaller live shape the kernel then decoded coords with max strides, scattered with max out-strides, and looped over the max numel. This is the `[dynamic]` output-mismatch behind `test_inception_v3`, `test_squeezenet1_1`, `test_densenet161` (all channel-cat on spatial-dynamic feature maps; the concat dim is static so `off_k` was already correct - the defect is the stale spatial strides/numel). Fix (mirrors the WebGPUGraph SwiGLU/QKV resize templates and the shipped `mul` hook): - Keep the per-input `in_meta`/`params` UBOs and the shared `out_meta` UBO alive via `own_uniform_buffer` (previously released after build); collect their handles plus each `add_dispatch` index. - Register an idempotent `add_tensor_resize_hook` on every input id: from `cur_dims` recompute live out dims (`set_cur_dims(out_id, ...)` to cascade to consumers + fix the delegate-output shape), rebuild `out_meta` + each input's `in_meta`/`params` and `wgpuQueueWriteBuffer` them, and rewrite each dispatch's `workgroup_count_x` via `compute_1d_workgroup_count`. On a static graph `cur_dims == dims`, so the hook rewrites identical values (no behavior change). Applied identically to both the `xplat/` and `fbcode/` mirrors (byte-identical). Co-authored-with: Claude Code. ghstack-source-id: 406388769 @exported-using-ghexport Differential Revision: [D113319596](https://our.internmc.facebook.com/intern/diff/D113319596/)
JCNTH
added a commit
that referenced
this pull request
Jul 24, 2026
…match) Pull Request resolved: #21232 Under dynamic shapes `cat` produced a wrong (scrambled + tail-garbage) output. `cat_impl` baked every shape-dependent quantity - each input `in_meta` (strides/numel), the shared `out_meta` (strides), the per-input `off_k`, and each dispatch's `workgroup_count` - from the MAX (upper-bound) build shape and registered NO resize hook, then released the UBOs. Under a smaller live shape the kernel then decoded coords with max strides, scattered with max out-strides, and looped over the max numel. This is the `[dynamic]` output-mismatch behind `test_inception_v3`, `test_squeezenet1_1`, `test_densenet161` (all channel-cat on spatial-dynamic feature maps; the concat dim is static so `off_k` was already correct - the defect is the stale spatial strides/numel). Fix (mirrors the WebGPUGraph SwiGLU/QKV resize templates and the shipped `mul` hook): - Keep the per-input `in_meta`/`params` UBOs and the shared `out_meta` UBO alive via `own_uniform_buffer` (previously released after build); collect their handles plus each `add_dispatch` index. - Register an idempotent `add_tensor_resize_hook` on every input id: from `cur_dims` recompute live out dims (`set_cur_dims(out_id, ...)` to cascade to consumers + fix the delegate-output shape), rebuild `out_meta` + each input's `in_meta`/`params` and `wgpuQueueWriteBuffer` them, and rewrite each dispatch's `workgroup_count_x` via `compute_1d_workgroup_count`. On a static graph `cur_dims == dims`, so the hook rewrites identical values (no behavior change). Applied identically to both the `xplat/` and `fbcode/` mirrors (byte-identical). Co-authored-with: Claude Code. ghstack-source-id: 406388769 @exported-using-ghexport Differential Revision: [D113319596](https://our.internmc.facebook.com/intern/diff/D113319596/)
JCNTH
added a commit
that referenced
this pull request
Jul 24, 2026
…match) Pull Request resolved: #21232 Under dynamic shapes `cat` produced a wrong (scrambled + tail-garbage) output. `cat_impl` baked every shape-dependent quantity - each input `in_meta` (strides/numel), the shared `out_meta` (strides), the per-input `off_k`, and each dispatch's `workgroup_count` - from the MAX (upper-bound) build shape and registered NO resize hook, then released the UBOs. Under a smaller live shape the kernel then decoded coords with max strides, scattered with max out-strides, and looped over the max numel. This is the `[dynamic]` output-mismatch behind `test_inception_v3`, `test_squeezenet1_1`, `test_densenet161` (all channel-cat on spatial-dynamic feature maps; the concat dim is static so `off_k` was already correct - the defect is the stale spatial strides/numel). Fix (mirrors the WebGPUGraph SwiGLU/QKV resize templates and the shipped `mul` hook): - Keep the per-input `in_meta`/`params` UBOs and the shared `out_meta` UBO alive via `own_uniform_buffer` (previously released after build); collect their handles plus each `add_dispatch` index. - Register an idempotent `add_tensor_resize_hook` on every input id: from `cur_dims` recompute live out dims (`set_cur_dims(out_id, ...)` to cascade to consumers + fix the delegate-output shape), rebuild `out_meta` + each input's `in_meta`/`params` and `wgpuQueueWriteBuffer` them, and rewrite each dispatch's `workgroup_count_x` via `compute_1d_workgroup_count`. On a static graph `cur_dims == dims`, so the hook rewrites identical values (no behavior change). Applied identically to both the `xplat/` and `fbcode/` mirrors (byte-identical). Co-authored-with: Claude Code. ghstack-source-id: 406388769 @exported-using-ghexport Differential Revision: [D113319596](https://our.internmc.facebook.com/intern/diff/D113319596/)
JCNTH
added a commit
that referenced
this pull request
Jul 24, 2026
…match) Pull Request resolved: #21232 Under dynamic shapes `cat` produced a wrong (scrambled + tail-garbage) output. `cat_impl` baked every shape-dependent quantity - each input `in_meta` (strides/numel), the shared `out_meta` (strides), the per-input `off_k`, and each dispatch's `workgroup_count` - from the MAX (upper-bound) build shape and registered NO resize hook, then released the UBOs. Under a smaller live shape the kernel then decoded coords with max strides, scattered with max out-strides, and looped over the max numel. This is the `[dynamic]` output-mismatch behind `test_inception_v3`, `test_squeezenet1_1`, `test_densenet161` (all channel-cat on spatial-dynamic feature maps; the concat dim is static so `off_k` was already correct - the defect is the stale spatial strides/numel). Fix (mirrors the WebGPUGraph SwiGLU/QKV resize templates and the shipped `mul` hook): - Keep the per-input `in_meta`/`params` UBOs and the shared `out_meta` UBO alive via `own_uniform_buffer` (previously released after build); collect their handles plus each `add_dispatch` index. - Register an idempotent `add_tensor_resize_hook` on every input id: from `cur_dims` recompute live out dims (`set_cur_dims(out_id, ...)` to cascade to consumers + fix the delegate-output shape), rebuild `out_meta` + each input's `in_meta`/`params` and `wgpuQueueWriteBuffer` them, and rewrite each dispatch's `workgroup_count_x` via `compute_1d_workgroup_count`. On a static graph `cur_dims == dims`, so the hook rewrites identical values (no behavior change). Applied identically to both the `xplat/` and `fbcode/` mirrors (byte-identical). Co-authored-with: Claude Code. ghstack-source-id: 406388769 @exported-using-ghexport Differential Revision: [D113319596](https://our.internmc.facebook.com/intern/diff/D113319596/)
JCNTH
added a commit
that referenced
this pull request
Jul 24, 2026
…match) Pull Request resolved: #21232 Under dynamic shapes `cat` produced a wrong (scrambled + tail-garbage) output. `cat_impl` baked every shape-dependent quantity - each input `in_meta` (strides/numel), the shared `out_meta` (strides), the per-input `off_k`, and each dispatch's `workgroup_count` - from the MAX (upper-bound) build shape and registered NO resize hook, then released the UBOs. Under a smaller live shape the kernel then decoded coords with max strides, scattered with max out-strides, and looped over the max numel. This is the `[dynamic]` output-mismatch behind `test_inception_v3`, `test_squeezenet1_1`, `test_densenet161` (all channel-cat on spatial-dynamic feature maps; the concat dim is static so `off_k` was already correct - the defect is the stale spatial strides/numel). Fix (mirrors the WebGPUGraph SwiGLU/QKV resize templates and the shipped `mul` hook): - Keep the per-input `in_meta`/`params` UBOs and the shared `out_meta` UBO alive via `own_uniform_buffer` (previously released after build); collect their handles plus each `add_dispatch` index. - Register an idempotent `add_tensor_resize_hook` on every input id: from `cur_dims` recompute live out dims (`set_cur_dims(out_id, ...)` to cascade to consumers + fix the delegate-output shape), rebuild `out_meta` + each input's `in_meta`/`params` and `wgpuQueueWriteBuffer` them, and rewrite each dispatch's `workgroup_count_x` via `compute_1d_workgroup_count`. On a static graph `cur_dims == dims`, so the hook rewrites identical values (no behavior change). Applied identically to both the `xplat/` and `fbcode/` mirrors (byte-identical). Co-authored-with: Claude Code. ghstack-source-id: 406388769 @exported-using-ghexport Differential Revision: [D113319596](https://our.internmc.facebook.com/intern/diff/D113319596/)
JCNTH
added a commit
that referenced
this pull request
Jul 24, 2026
…match) Pull Request resolved: #21232 Under dynamic shapes `cat` produced a wrong (scrambled + tail-garbage) output. `cat_impl` baked every shape-dependent quantity - each input `in_meta` (strides/numel), the shared `out_meta` (strides), the per-input `off_k`, and each dispatch's `workgroup_count` - from the MAX (upper-bound) build shape and registered NO resize hook, then released the UBOs. Under a smaller live shape the kernel then decoded coords with max strides, scattered with max out-strides, and looped over the max numel. This is the `[dynamic]` output-mismatch behind `test_inception_v3`, `test_squeezenet1_1`, `test_densenet161` (all channel-cat on spatial-dynamic feature maps; the concat dim is static so `off_k` was already correct - the defect is the stale spatial strides/numel). Fix (mirrors the WebGPUGraph SwiGLU/QKV resize templates and the shipped `mul` hook): - Keep the per-input `in_meta`/`params` UBOs and the shared `out_meta` UBO alive via `own_uniform_buffer` (previously released after build); collect their handles plus each `add_dispatch` index. - Register an idempotent `add_tensor_resize_hook` on every input id: from `cur_dims` recompute live out dims (`set_cur_dims(out_id, ...)` to cascade to consumers + fix the delegate-output shape), rebuild `out_meta` + each input's `in_meta`/`params` and `wgpuQueueWriteBuffer` them, and rewrite each dispatch's `workgroup_count_x` via `compute_1d_workgroup_count`. On a static graph `cur_dims == dims`, so the hook rewrites identical values (no behavior change). Applied identically to both the `xplat/` and `fbcode/` mirrors (byte-identical). Co-authored-with: Claude Code. ghstack-source-id: 406388769 @exported-using-ghexport Differential Revision: [D113319596](https://our.internmc.facebook.com/intern/diff/D113319596/)
JCNTH
added a commit
that referenced
this pull request
Jul 24, 2026
…match) Pull Request resolved: #21232 Under dynamic shapes `cat` produced a wrong (scrambled + tail-garbage) output. `cat_impl` baked every shape-dependent quantity - each input `in_meta` (strides/numel), the shared `out_meta` (strides), the per-input `off_k`, and each dispatch's `workgroup_count` - from the MAX (upper-bound) build shape and registered NO resize hook, then released the UBOs. Under a smaller live shape the kernel then decoded coords with max strides, scattered with max out-strides, and looped over the max numel. This is the `[dynamic]` output-mismatch behind `test_inception_v3`, `test_squeezenet1_1`, `test_densenet161` (all channel-cat on spatial-dynamic feature maps; the concat dim is static so `off_k` was already correct - the defect is the stale spatial strides/numel). Fix (mirrors the WebGPUGraph SwiGLU/QKV resize templates and the shipped `mul` hook): - Keep the per-input `in_meta`/`params` UBOs and the shared `out_meta` UBO alive via `own_uniform_buffer` (previously released after build); collect their handles plus each `add_dispatch` index. - Register an idempotent `add_tensor_resize_hook` on every input id: from `cur_dims` recompute live out dims (`set_cur_dims(out_id, ...)` to cascade to consumers + fix the delegate-output shape), rebuild `out_meta` + each input's `in_meta`/`params` and `wgpuQueueWriteBuffer` them, and rewrite each dispatch's `workgroup_count_x` via `compute_1d_workgroup_count`. On a static graph `cur_dims == dims`, so the hook rewrites identical values (no behavior change). Applied identically to both the `xplat/` and `fbcode/` mirrors (byte-identical). Co-authored-with: Claude Code. ghstack-source-id: 406388769 @exported-using-ghexport Differential Revision: [D113319596](https://our.internmc.facebook.com/intern/diff/D113319596/)
JCNTH
added a commit
that referenced
this pull request
Jul 24, 2026
…match) Pull Request resolved: #21232 Under dynamic shapes `cat` produced a wrong (scrambled + tail-garbage) output. `cat_impl` baked every shape-dependent quantity - each input `in_meta` (strides/numel), the shared `out_meta` (strides), the per-input `off_k`, and each dispatch's `workgroup_count` - from the MAX (upper-bound) build shape and registered NO resize hook, then released the UBOs. Under a smaller live shape the kernel then decoded coords with max strides, scattered with max out-strides, and looped over the max numel. This is the `[dynamic]` output-mismatch behind `test_inception_v3`, `test_squeezenet1_1`, `test_densenet161` (all channel-cat on spatial-dynamic feature maps; the concat dim is static so `off_k` was already correct - the defect is the stale spatial strides/numel). Fix (mirrors the WebGPUGraph SwiGLU/QKV resize templates and the shipped `mul` hook): - Keep the per-input `in_meta`/`params` UBOs and the shared `out_meta` UBO alive via `own_uniform_buffer` (previously released after build); collect their handles plus each `add_dispatch` index. - Register an idempotent `add_tensor_resize_hook` on every input id: from `cur_dims` recompute live out dims (`set_cur_dims(out_id, ...)` to cascade to consumers + fix the delegate-output shape), rebuild `out_meta` + each input's `in_meta`/`params` and `wgpuQueueWriteBuffer` them, and rewrite each dispatch's `workgroup_count_x` via `compute_1d_workgroup_count`. On a static graph `cur_dims == dims`, so the hook rewrites identical values (no behavior change). Applied identically to both the `xplat/` and `fbcode/` mirrors (byte-identical). Co-authored-with: Claude Code. ghstack-source-id: 406388769 @exported-using-ghexport Differential Revision: [D113319596](https://our.internmc.facebook.com/intern/diff/D113319596/)
JCNTH
added a commit
that referenced
this pull request
Jul 24, 2026
…match) Pull Request resolved: #21232 Under dynamic shapes `cat` produced a wrong (scrambled + tail-garbage) output. `cat_impl` baked every shape-dependent quantity - each input `in_meta` (strides/numel), the shared `out_meta` (strides), the per-input `off_k`, and each dispatch's `workgroup_count` - from the MAX (upper-bound) build shape and registered NO resize hook, then released the UBOs. Under a smaller live shape the kernel then decoded coords with max strides, scattered with max out-strides, and looped over the max numel. This is the `[dynamic]` output-mismatch behind `test_inception_v3`, `test_squeezenet1_1`, `test_densenet161` (all channel-cat on spatial-dynamic feature maps; the concat dim is static so `off_k` was already correct - the defect is the stale spatial strides/numel). Fix (mirrors the WebGPUGraph SwiGLU/QKV resize templates and the shipped `mul` hook): - Keep the per-input `in_meta`/`params` UBOs and the shared `out_meta` UBO alive via `own_uniform_buffer` (previously released after build); collect their handles plus each `add_dispatch` index. - Register an idempotent `add_tensor_resize_hook` on every input id: from `cur_dims` recompute live out dims (`set_cur_dims(out_id, ...)` to cascade to consumers + fix the delegate-output shape), rebuild `out_meta` + each input's `in_meta`/`params` and `wgpuQueueWriteBuffer` them, and rewrite each dispatch's `workgroup_count_x` via `compute_1d_workgroup_count`. On a static graph `cur_dims == dims`, so the hook rewrites identical values (no behavior change). Applied identically to both the `xplat/` and `fbcode/` mirrors (byte-identical). Co-authored-with: Claude Code. ghstack-source-id: 406388769 @exported-using-ghexport Differential Revision: [D113319596](https://our.internmc.facebook.com/intern/diff/D113319596/)
JCNTH
added a commit
that referenced
this pull request
Jul 24, 2026
…match) Pull Request resolved: #21232 Under dynamic shapes `cat` produced a wrong (scrambled + tail-garbage) output. `cat_impl` baked every shape-dependent quantity - each input `in_meta` (strides/numel), the shared `out_meta` (strides), the per-input `off_k`, and each dispatch's `workgroup_count` - from the MAX (upper-bound) build shape and registered NO resize hook, then released the UBOs. Under a smaller live shape the kernel then decoded coords with max strides, scattered with max out-strides, and looped over the max numel. This is the `[dynamic]` output-mismatch behind `test_inception_v3`, `test_squeezenet1_1`, `test_densenet161` (all channel-cat on spatial-dynamic feature maps; the concat dim is static so `off_k` was already correct - the defect is the stale spatial strides/numel). Fix (mirrors the WebGPUGraph SwiGLU/QKV resize templates and the shipped `mul` hook): - Keep the per-input `in_meta`/`params` UBOs and the shared `out_meta` UBO alive via `own_uniform_buffer` (previously released after build); collect their handles plus each `add_dispatch` index. - Register an idempotent `add_tensor_resize_hook` on every input id: from `cur_dims` recompute live out dims (`set_cur_dims(out_id, ...)` to cascade to consumers + fix the delegate-output shape), rebuild `out_meta` + each input's `in_meta`/`params` and `wgpuQueueWriteBuffer` them, and rewrite each dispatch's `workgroup_count_x` via `compute_1d_workgroup_count`. On a static graph `cur_dims == dims`, so the hook rewrites identical values (no behavior change). Applied identically to both the `xplat/` and `fbcode/` mirrors (byte-identical). Co-authored-with: Claude Code. ghstack-source-id: 406388769 @exported-using-ghexport Differential Revision: [D113319596](https://our.internmc.facebook.com/intern/diff/D113319596/)
JCNTH
added a commit
that referenced
this pull request
Jul 24, 2026
…match) Pull Request resolved: #21232 Under dynamic shapes `cat` produced a wrong (scrambled + tail-garbage) output. `cat_impl` baked every shape-dependent quantity - each input `in_meta` (strides/numel), the shared `out_meta` (strides), the per-input `off_k`, and each dispatch's `workgroup_count` - from the MAX (upper-bound) build shape and registered NO resize hook, then released the UBOs. Under a smaller live shape the kernel then decoded coords with max strides, scattered with max out-strides, and looped over the max numel. This is the `[dynamic]` output-mismatch behind `test_inception_v3`, `test_squeezenet1_1`, `test_densenet161` (all channel-cat on spatial-dynamic feature maps; the concat dim is static so `off_k` was already correct - the defect is the stale spatial strides/numel). Fix (mirrors the WebGPUGraph SwiGLU/QKV resize templates and the shipped `mul` hook): - Keep the per-input `in_meta`/`params` UBOs and the shared `out_meta` UBO alive via `own_uniform_buffer` (previously released after build); collect their handles plus each `add_dispatch` index. - Register an idempotent `add_tensor_resize_hook` on every input id: from `cur_dims` recompute live out dims (`set_cur_dims(out_id, ...)` to cascade to consumers + fix the delegate-output shape), rebuild `out_meta` + each input's `in_meta`/`params` and `wgpuQueueWriteBuffer` them, and rewrite each dispatch's `workgroup_count_x` via `compute_1d_workgroup_count`. On a static graph `cur_dims == dims`, so the hook rewrites identical values (no behavior change). Applied identically to both the `xplat/` and `fbcode/` mirrors (byte-identical). Co-authored-with: Claude Code. ghstack-source-id: 406388769 @exported-using-ghexport Differential Revision: [D113319596](https://our.internmc.facebook.com/intern/diff/D113319596/)
JCNTH
added a commit
that referenced
this pull request
Jul 24, 2026
…match) Pull Request resolved: #21232 Under dynamic shapes `cat` produced a wrong (scrambled + tail-garbage) output. `cat_impl` baked every shape-dependent quantity - each input `in_meta` (strides/numel), the shared `out_meta` (strides), the per-input `off_k`, and each dispatch's `workgroup_count` - from the MAX (upper-bound) build shape and registered NO resize hook, then released the UBOs. Under a smaller live shape the kernel then decoded coords with max strides, scattered with max out-strides, and looped over the max numel. This is the `[dynamic]` output-mismatch behind `test_inception_v3`, `test_squeezenet1_1`, `test_densenet161` (all channel-cat on spatial-dynamic feature maps; the concat dim is static so `off_k` was already correct - the defect is the stale spatial strides/numel). Fix (mirrors the WebGPUGraph SwiGLU/QKV resize templates and the shipped `mul` hook): - Keep the per-input `in_meta`/`params` UBOs and the shared `out_meta` UBO alive via `own_uniform_buffer` (previously released after build); collect their handles plus each `add_dispatch` index. - Register an idempotent `add_tensor_resize_hook` on every input id: from `cur_dims` recompute live out dims (`set_cur_dims(out_id, ...)` to cascade to consumers + fix the delegate-output shape), rebuild `out_meta` + each input's `in_meta`/`params` and `wgpuQueueWriteBuffer` them, and rewrite each dispatch's `workgroup_count_x` via `compute_1d_workgroup_count`. On a static graph `cur_dims == dims`, so the hook rewrites identical values (no behavior change). Applied identically to both the `xplat/` and `fbcode/` mirrors (byte-identical). Co-authored-with: Claude Code. ghstack-source-id: 406388769 @exported-using-ghexport Differential Revision: [D113319596](https://our.internmc.facebook.com/intern/diff/D113319596/)
JCNTH
added a commit
that referenced
this pull request
Jul 24, 2026
…match) Pull Request resolved: #21232 Under dynamic shapes `cat` produced a wrong (scrambled + tail-garbage) output. `cat_impl` baked every shape-dependent quantity - each input `in_meta` (strides/numel), the shared `out_meta` (strides), the per-input `off_k`, and each dispatch's `workgroup_count` - from the MAX (upper-bound) build shape and registered NO resize hook, then released the UBOs. Under a smaller live shape the kernel then decoded coords with max strides, scattered with max out-strides, and looped over the max numel. This is the `[dynamic]` output-mismatch behind `test_inception_v3`, `test_squeezenet1_1`, `test_densenet161` (all channel-cat on spatial-dynamic feature maps; the concat dim is static so `off_k` was already correct - the defect is the stale spatial strides/numel). Fix (mirrors the WebGPUGraph SwiGLU/QKV resize templates and the shipped `mul` hook): - Keep the per-input `in_meta`/`params` UBOs and the shared `out_meta` UBO alive via `own_uniform_buffer` (previously released after build); collect their handles plus each `add_dispatch` index. - Register an idempotent `add_tensor_resize_hook` on every input id: from `cur_dims` recompute live out dims (`set_cur_dims(out_id, ...)` to cascade to consumers + fix the delegate-output shape), rebuild `out_meta` + each input's `in_meta`/`params` and `wgpuQueueWriteBuffer` them, and rewrite each dispatch's `workgroup_count_x` via `compute_1d_workgroup_count`. On a static graph `cur_dims == dims`, so the hook rewrites identical values (no behavior change). Applied identically to both the `xplat/` and `fbcode/` mirrors (byte-identical). Co-authored-with: Claude Code. ghstack-source-id: 406388769 @exported-using-ghexport Differential Revision: [D113319596](https://our.internmc.facebook.com/intern/diff/D113319596/)
JCNTH
added a commit
that referenced
this pull request
Jul 24, 2026
…match) Pull Request resolved: #21232 Under dynamic shapes `cat` produced a wrong (scrambled + tail-garbage) output. `cat_impl` baked every shape-dependent quantity - each input `in_meta` (strides/numel), the shared `out_meta` (strides), the per-input `off_k`, and each dispatch's `workgroup_count` - from the MAX (upper-bound) build shape and registered NO resize hook, then released the UBOs. Under a smaller live shape the kernel then decoded coords with max strides, scattered with max out-strides, and looped over the max numel. This is the `[dynamic]` output-mismatch behind `test_inception_v3`, `test_squeezenet1_1`, `test_densenet161` (all channel-cat on spatial-dynamic feature maps; the concat dim is static so `off_k` was already correct - the defect is the stale spatial strides/numel). Fix (mirrors the WebGPUGraph SwiGLU/QKV resize templates and the shipped `mul` hook): - Keep the per-input `in_meta`/`params` UBOs and the shared `out_meta` UBO alive via `own_uniform_buffer` (previously released after build); collect their handles plus each `add_dispatch` index. - Register an idempotent `add_tensor_resize_hook` on every input id: from `cur_dims` recompute live out dims (`set_cur_dims(out_id, ...)` to cascade to consumers + fix the delegate-output shape), rebuild `out_meta` + each input's `in_meta`/`params` and `wgpuQueueWriteBuffer` them, and rewrite each dispatch's `workgroup_count_x` via `compute_1d_workgroup_count`. On a static graph `cur_dims == dims`, so the hook rewrites identical values (no behavior change). Applied identically to both the `xplat/` and `fbcode/` mirrors (byte-identical). Co-authored-with: Claude Code. ghstack-source-id: 406388769 @exported-using-ghexport Differential Revision: [D113319596](https://our.internmc.facebook.com/intern/diff/D113319596/)
JCNTH
added a commit
that referenced
this pull request
Jul 24, 2026
…match) Pull Request resolved: #21232 Under dynamic shapes `cat` produced a wrong (scrambled + tail-garbage) output. `cat_impl` baked every shape-dependent quantity - each input `in_meta` (strides/numel), the shared `out_meta` (strides), the per-input `off_k`, and each dispatch's `workgroup_count` - from the MAX (upper-bound) build shape and registered NO resize hook, then released the UBOs. Under a smaller live shape the kernel then decoded coords with max strides, scattered with max out-strides, and looped over the max numel. This is the `[dynamic]` output-mismatch behind `test_inception_v3`, `test_squeezenet1_1`, `test_densenet161` (all channel-cat on spatial-dynamic feature maps; the concat dim is static so `off_k` was already correct - the defect is the stale spatial strides/numel). Fix (mirrors the WebGPUGraph SwiGLU/QKV resize templates and the shipped `mul` hook): - Keep the per-input `in_meta`/`params` UBOs and the shared `out_meta` UBO alive via `own_uniform_buffer` (previously released after build); collect their handles plus each `add_dispatch` index. - Register an idempotent `add_tensor_resize_hook` on every input id: from `cur_dims` recompute live out dims (`set_cur_dims(out_id, ...)` to cascade to consumers + fix the delegate-output shape), rebuild `out_meta` + each input's `in_meta`/`params` and `wgpuQueueWriteBuffer` them, and rewrite each dispatch's `workgroup_count_x` via `compute_1d_workgroup_count`. On a static graph `cur_dims == dims`, so the hook rewrites identical values (no behavior change). Applied identically to both the `xplat/` and `fbcode/` mirrors (byte-identical). Co-authored-with: Claude Code. ghstack-source-id: 406388769 @exported-using-ghexport Differential Revision: [D113319596](https://our.internmc.facebook.com/intern/diff/D113319596/)
JCNTH
added a commit
that referenced
this pull request
Jul 24, 2026
…match) Pull Request resolved: #21232 Under dynamic shapes `cat` produced a wrong (scrambled + tail-garbage) output. `cat_impl` baked every shape-dependent quantity - each input `in_meta` (strides/numel), the shared `out_meta` (strides), the per-input `off_k`, and each dispatch's `workgroup_count` - from the MAX (upper-bound) build shape and registered NO resize hook, then released the UBOs. Under a smaller live shape the kernel then decoded coords with max strides, scattered with max out-strides, and looped over the max numel. This is the `[dynamic]` output-mismatch behind `test_inception_v3`, `test_squeezenet1_1`, `test_densenet161` (all channel-cat on spatial-dynamic feature maps; the concat dim is static so `off_k` was already correct - the defect is the stale spatial strides/numel). Fix (mirrors the WebGPUGraph SwiGLU/QKV resize templates and the shipped `mul` hook): - Keep the per-input `in_meta`/`params` UBOs and the shared `out_meta` UBO alive via `own_uniform_buffer` (previously released after build); collect their handles plus each `add_dispatch` index. - Register an idempotent `add_tensor_resize_hook` on every input id: from `cur_dims` recompute live out dims (`set_cur_dims(out_id, ...)` to cascade to consumers + fix the delegate-output shape), rebuild `out_meta` + each input's `in_meta`/`params` and `wgpuQueueWriteBuffer` them, and rewrite each dispatch's `workgroup_count_x` via `compute_1d_workgroup_count`. On a static graph `cur_dims == dims`, so the hook rewrites identical values (no behavior change). Applied identically to both the `xplat/` and `fbcode/` mirrors (byte-identical). Co-authored-with: Claude Code. ghstack-source-id: 406388769 @exported-using-ghexport Differential Revision: [D113319596](https://our.internmc.facebook.com/intern/diff/D113319596/)
JCNTH
added a commit
that referenced
this pull request
Jul 24, 2026
…match) Pull Request resolved: #21232 Under dynamic shapes `cat` produced a wrong (scrambled + tail-garbage) output. `cat_impl` baked every shape-dependent quantity - each input `in_meta` (strides/numel), the shared `out_meta` (strides), the per-input `off_k`, and each dispatch's `workgroup_count` - from the MAX (upper-bound) build shape and registered NO resize hook, then released the UBOs. Under a smaller live shape the kernel then decoded coords with max strides, scattered with max out-strides, and looped over the max numel. This is the `[dynamic]` output-mismatch behind `test_inception_v3`, `test_squeezenet1_1`, `test_densenet161` (all channel-cat on spatial-dynamic feature maps; the concat dim is static so `off_k` was already correct - the defect is the stale spatial strides/numel). Fix (mirrors the WebGPUGraph SwiGLU/QKV resize templates and the shipped `mul` hook): - Keep the per-input `in_meta`/`params` UBOs and the shared `out_meta` UBO alive via `own_uniform_buffer` (previously released after build); collect their handles plus each `add_dispatch` index. - Register an idempotent `add_tensor_resize_hook` on every input id: from `cur_dims` recompute live out dims (`set_cur_dims(out_id, ...)` to cascade to consumers + fix the delegate-output shape), rebuild `out_meta` + each input's `in_meta`/`params` and `wgpuQueueWriteBuffer` them, and rewrite each dispatch's `workgroup_count_x` via `compute_1d_workgroup_count`. On a static graph `cur_dims == dims`, so the hook rewrites identical values (no behavior change). Applied identically to both the `xplat/` and `fbcode/` mirrors (byte-identical). Co-authored-with: Claude Code. ghstack-source-id: 406388769 @exported-using-ghexport Differential Revision: [D113319596](https://our.internmc.facebook.com/intern/diff/D113319596/)
JCNTH
added a commit
that referenced
this pull request
Jul 24, 2026
…match) Pull Request resolved: #21232 Under dynamic shapes `cat` produced a wrong (scrambled + tail-garbage) output. `cat_impl` baked every shape-dependent quantity - each input `in_meta` (strides/numel), the shared `out_meta` (strides), the per-input `off_k`, and each dispatch's `workgroup_count` - from the MAX (upper-bound) build shape and registered NO resize hook, then released the UBOs. Under a smaller live shape the kernel then decoded coords with max strides, scattered with max out-strides, and looped over the max numel. This is the `[dynamic]` output-mismatch behind `test_inception_v3`, `test_squeezenet1_1`, `test_densenet161` (all channel-cat on spatial-dynamic feature maps; the concat dim is static so `off_k` was already correct - the defect is the stale spatial strides/numel). Fix (mirrors the WebGPUGraph SwiGLU/QKV resize templates and the shipped `mul` hook): - Keep the per-input `in_meta`/`params` UBOs and the shared `out_meta` UBO alive via `own_uniform_buffer` (previously released after build); collect their handles plus each `add_dispatch` index. - Register an idempotent `add_tensor_resize_hook` on every input id: from `cur_dims` recompute live out dims (`set_cur_dims(out_id, ...)` to cascade to consumers + fix the delegate-output shape), rebuild `out_meta` + each input's `in_meta`/`params` and `wgpuQueueWriteBuffer` them, and rewrite each dispatch's `workgroup_count_x` via `compute_1d_workgroup_count`. On a static graph `cur_dims == dims`, so the hook rewrites identical values (no behavior change). Applied identically to both the `xplat/` and `fbcode/` mirrors (byte-identical). Co-authored-with: Claude Code. ghstack-source-id: 406388769 @exported-using-ghexport Differential Revision: [D113319596](https://our.internmc.facebook.com/intern/diff/D113319596/)
JCNTH
added a commit
that referenced
this pull request
Jul 24, 2026
…match) Pull Request resolved: #21232 Under dynamic shapes `cat` produced a wrong (scrambled + tail-garbage) output. `cat_impl` baked every shape-dependent quantity - each input `in_meta` (strides/numel), the shared `out_meta` (strides), the per-input `off_k`, and each dispatch's `workgroup_count` - from the MAX (upper-bound) build shape and registered NO resize hook, then released the UBOs. Under a smaller live shape the kernel then decoded coords with max strides, scattered with max out-strides, and looped over the max numel. This is the `[dynamic]` output-mismatch behind `test_inception_v3`, `test_squeezenet1_1`, `test_densenet161` (all channel-cat on spatial-dynamic feature maps; the concat dim is static so `off_k` was already correct - the defect is the stale spatial strides/numel). Fix (mirrors the WebGPUGraph SwiGLU/QKV resize templates and the shipped `mul` hook): - Keep the per-input `in_meta`/`params` UBOs and the shared `out_meta` UBO alive via `own_uniform_buffer` (previously released after build); collect their handles plus each `add_dispatch` index. - Register an idempotent `add_tensor_resize_hook` on every input id: from `cur_dims` recompute live out dims (`set_cur_dims(out_id, ...)` to cascade to consumers + fix the delegate-output shape), rebuild `out_meta` + each input's `in_meta`/`params` and `wgpuQueueWriteBuffer` them, and rewrite each dispatch's `workgroup_count_x` via `compute_1d_workgroup_count`. On a static graph `cur_dims == dims`, so the hook rewrites identical values (no behavior change). Applied identically to both the `xplat/` and `fbcode/` mirrors (byte-identical). Co-authored-with: Claude Code. ghstack-source-id: 406388769 @exported-using-ghexport Differential Revision: [D113319596](https://our.internmc.facebook.com/intern/diff/D113319596/)
JCNTH
added a commit
that referenced
this pull request
Jul 24, 2026
…match) Pull Request resolved: #21232 Under dynamic shapes `cat` produced a wrong (scrambled + tail-garbage) output. `cat_impl` baked every shape-dependent quantity - each input `in_meta` (strides/numel), the shared `out_meta` (strides), the per-input `off_k`, and each dispatch's `workgroup_count` - from the MAX (upper-bound) build shape and registered NO resize hook, then released the UBOs. Under a smaller live shape the kernel then decoded coords with max strides, scattered with max out-strides, and looped over the max numel. This is the `[dynamic]` output-mismatch behind `test_inception_v3`, `test_squeezenet1_1`, `test_densenet161` (all channel-cat on spatial-dynamic feature maps; the concat dim is static so `off_k` was already correct - the defect is the stale spatial strides/numel). Fix (mirrors the WebGPUGraph SwiGLU/QKV resize templates and the shipped `mul` hook): - Keep the per-input `in_meta`/`params` UBOs and the shared `out_meta` UBO alive via `own_uniform_buffer` (previously released after build); collect their handles plus each `add_dispatch` index. - Register an idempotent `add_tensor_resize_hook` on every input id: from `cur_dims` recompute live out dims (`set_cur_dims(out_id, ...)` to cascade to consumers + fix the delegate-output shape), rebuild `out_meta` + each input's `in_meta`/`params` and `wgpuQueueWriteBuffer` them, and rewrite each dispatch's `workgroup_count_x` via `compute_1d_workgroup_count`. On a static graph `cur_dims == dims`, so the hook rewrites identical values (no behavior change). Applied identically to both the `xplat/` and `fbcode/` mirrors (byte-identical). Co-authored-with: Claude Code. ghstack-source-id: 406388769 @exported-using-ghexport Differential Revision: [D113319596](https://our.internmc.facebook.com/intern/diff/D113319596/)
JCNTH
added a commit
that referenced
this pull request
Jul 24, 2026
…match) Pull Request resolved: #21232 Under dynamic shapes `cat` produced a wrong (scrambled + tail-garbage) output. `cat_impl` baked every shape-dependent quantity - each input `in_meta` (strides/numel), the shared `out_meta` (strides), the per-input `off_k`, and each dispatch's `workgroup_count` - from the MAX (upper-bound) build shape and registered NO resize hook, then released the UBOs. Under a smaller live shape the kernel then decoded coords with max strides, scattered with max out-strides, and looped over the max numel. This is the `[dynamic]` output-mismatch behind `test_inception_v3`, `test_squeezenet1_1`, `test_densenet161` (all channel-cat on spatial-dynamic feature maps; the concat dim is static so `off_k` was already correct - the defect is the stale spatial strides/numel). Fix (mirrors the WebGPUGraph SwiGLU/QKV resize templates and the shipped `mul` hook): - Keep the per-input `in_meta`/`params` UBOs and the shared `out_meta` UBO alive via `own_uniform_buffer` (previously released after build); collect their handles plus each `add_dispatch` index. - Register an idempotent `add_tensor_resize_hook` on every input id: from `cur_dims` recompute live out dims (`set_cur_dims(out_id, ...)` to cascade to consumers + fix the delegate-output shape), rebuild `out_meta` + each input's `in_meta`/`params` and `wgpuQueueWriteBuffer` them, and rewrite each dispatch's `workgroup_count_x` via `compute_1d_workgroup_count`. On a static graph `cur_dims == dims`, so the hook rewrites identical values (no behavior change). Applied identically to both the `xplat/` and `fbcode/` mirrors (byte-identical). Co-authored-with: Claude Code. ghstack-source-id: 406388769 @exported-using-ghexport Differential Revision: [D113319596](https://our.internmc.facebook.com/intern/diff/D113319596/)
JCNTH
added a commit
that referenced
this pull request
Jul 24, 2026
…match) Pull Request resolved: #21232 Under dynamic shapes `cat` produced a wrong (scrambled + tail-garbage) output. `cat_impl` baked every shape-dependent quantity - each input `in_meta` (strides/numel), the shared `out_meta` (strides), the per-input `off_k`, and each dispatch's `workgroup_count` - from the MAX (upper-bound) build shape and registered NO resize hook, then released the UBOs. Under a smaller live shape the kernel then decoded coords with max strides, scattered with max out-strides, and looped over the max numel. This is the `[dynamic]` output-mismatch behind `test_inception_v3`, `test_squeezenet1_1`, `test_densenet161` (all channel-cat on spatial-dynamic feature maps; the concat dim is static so `off_k` was already correct - the defect is the stale spatial strides/numel). Fix (mirrors the WebGPUGraph SwiGLU/QKV resize templates and the shipped `mul` hook): - Keep the per-input `in_meta`/`params` UBOs and the shared `out_meta` UBO alive via `own_uniform_buffer` (previously released after build); collect their handles plus each `add_dispatch` index. - Register an idempotent `add_tensor_resize_hook` on every input id: from `cur_dims` recompute live out dims (`set_cur_dims(out_id, ...)` to cascade to consumers + fix the delegate-output shape), rebuild `out_meta` + each input's `in_meta`/`params` and `wgpuQueueWriteBuffer` them, and rewrite each dispatch's `workgroup_count_x` via `compute_1d_workgroup_count`. On a static graph `cur_dims == dims`, so the hook rewrites identical values (no behavior change). Applied identically to both the `xplat/` and `fbcode/` mirrors (byte-identical). Co-authored-with: Claude Code. ghstack-source-id: 406388769 @exported-using-ghexport Differential Revision: [D113319596](https://our.internmc.facebook.com/intern/diff/D113319596/)
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
2 participants
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.
Stack from ghstack (oldest at bottom):
Under dynamic shapes
catproduced a wrong (scrambled + tail-garbage) output.cat_implbaked every shape-dependent quantity - each inputin_meta(strides/numel), the sharedout_meta(strides), the per-inputoff_k, and each dispatch'sworkgroup_count- from the MAX (upper-bound) build shape and registered NO resize hook, then released the UBOs. Under a smaller live shape the kernel then decoded coords with max strides, scattered with max out-strides, and looped over the max numel. This is the[dynamic]output-mismatch behindtest_inception_v3,test_squeezenet1_1,test_densenet161(all channel-cat on spatial-dynamic feature maps; the concat dim is static sooff_kwas already correct - the defect is the stale spatial strides/numel).Fix (mirrors the WebGPUGraph SwiGLU/QKV resize templates and the shipped
mulhook):in_meta/paramsUBOs and the sharedout_metaUBO alive viaown_uniform_buffer(previously released after build); collect their handles plus eachadd_dispatchindex.add_tensor_resize_hookon every input id: fromcur_dimsrecompute live out dims (set_cur_dims(out_id, ...)to cascade to consumers + fix the delegate-output shape), rebuildout_meta+ each input'sin_meta/paramsandwgpuQueueWriteBufferthem, and rewrite each dispatch'sworkgroup_count_xviacompute_1d_workgroup_count. On a static graphcur_dims == dims, so the hook rewrites identical values (no behavior change).Applied identically to both the
xplat/andfbcode/mirrors (byte-identical).Co-authored-with: Claude Code.
@exported-using-ghexport
Differential Revision: D113319596
Differential Revision: D113319596