WIP: Add bitwise left/right #15893

mergennachin · 2025-11-19T17:01:45Z

No description provided.

github-actions · 2025-11-19T17:02:27Z

This PR needs a `release notes:` label

If your change should be included in the release notes (i.e. would users of this library care about this change?), please use a label starting with release notes:. This helps us keep track and include your important work in the next release notes.

To add a label, you can comment to pytorchbot, for example
@pytorchbot label "release notes: none"

For more information, see
https://github.com/pytorch/pytorch/wiki/PyTorch-AutoLabel-Bot#why-categorize-for-release-notes-and-how-does-it-work.

mergennachin · 2025-11-19T17:02:31Z

#15673 (comment)

meta-codesync · 2025-11-19T17:08:50Z

@mergennachin has imported this pull request. If you are a Meta employee, you can view this in D87451202.

Copilot

Pull Request Overview

This PR adds support for bitwise left shift (<<) and right shift (>>) operations to the ExecutorTorch portable kernel library, following the existing pattern for bitwise operations (and, or, xor).

Key Changes:

Implementation of bitwise_left_shift and bitwise_right_shift operators with both Tensor and Scalar variants
Comprehensive test coverage for both operations including edge cases (zero shift, different data types)
Build system integration across Buck (bzl), CMake, and YAML configuration files

Reviewed Changes

Copilot reviewed 9 out of 9 changed files in this pull request and generated 9 comments.

Show a summary per file

File	Description
`shim_et/xplat/executorch/kernels/portable/op_registration_util.bzl`	Adds Buck build targets for the new shift operators (ordering needs correction)
`kernels/test/targets.bzl`	Registers test targets for both shift operations (correctly ordered)
`kernels/test/CMakeLists.txt`	Adds test source files to CMake build (correctly ordered)
`kernels/portable/functions.yaml`	Defines kernel function signatures (should use `- op:` instead of `- func:`)
`kernels/portable/cpu/pattern/bitwise_op.h`	Adds shift operation functors and templates (contains redundant code)
`kernels/portable/cpu/op_bitwise_left_shift.cpp`	Implements left shift operation using the bitwise pattern
`kernels/portable/cpu/op_bitwise_right_shift.cpp`	Implements right shift operation using the bitwise pattern
`kernels/test/op_bitwise_left_shift_test.cpp`	Provides comprehensive test coverage for left shift
`kernels/test/op_bitwise_right_shift_test.cpp`	Provides comprehensive test coverage for right shift

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

kernels/portable/cpu/pattern/bitwise_op.h

kernels/portable/cpu/op_bitwise_left_shift.cpp

Copilot · 2025-11-19T17:13:00Z

shim_et/xplat/executorch/kernels/portable/op_registration_util.bzl

+    op_target(
+        name = "op_bitwise_right_shift",
+        deps = [
+            ":scalar_utils",
+            "//executorch/kernels/portable/cpu/pattern:bitwise_op",
+            "//executorch/kernels/portable/cpu/util:broadcast_util",
+            "//executorch/kernels/portable/cpu/util:dtype_util",
+            "//executorch/kernels/portable/cpu/util:elementwise_util",
+        ],
+    ),


The op_bitwise_right_shift target is placed out of alphabetical order. It should be positioned between op_bitwise_or (line 357) and op_bitwise_xor (line 367) to maintain alphabetical ordering of the op targets.

Copilot · 2025-11-19T17:13:01Z

kernels/portable/functions.yaml

+- func: bitwise_left_shift.Tensor_out(Tensor self, Tensor other, *, Tensor(a!) out) -> Tensor(a!)
+  kernels:
+    - arg_meta: null
+      kernel_name: torch::executor::bitwise_left_shift_Tensor_out
+
+- func: bitwise_left_shift.Scalar_out(Tensor self, Scalar other, *, Tensor(a!) out) -> Tensor(a!)
+  kernels:
+    - arg_meta: null
+      kernel_name: torch::executor::bitwise_left_shift_Scalar_out
+
+- func: bitwise_right_shift.Tensor_out(Tensor self, Tensor other, *, Tensor(a!) out) -> Tensor(a!)
+  kernels:
+    - arg_meta: null
+      kernel_name: torch::executor::bitwise_right_shift_Tensor_out
+
+- func: bitwise_right_shift.Scalar_out(Tensor self, Scalar other, *, Tensor(a!) out) -> Tensor(a!)


The shift operations should use - op: instead of - func: to be consistent with the other bitwise operations (and, or, xor, not) which are ATen library operations. The - func: syntax is typically reserved for custom operations not in the ATen library.

Change to:

- op: bitwise_left_shift.Tensor_out kernels: - arg_meta: null kernel_name: torch::executor::bitwise_left_shift_Tensor_out

Suggested change

- func: bitwise_left_shift.Tensor_out(Tensor self, Tensor other, *, Tensor(a!) out) -> Tensor(a!)

kernels:

- arg_meta: null

kernel_name: torch::executor::bitwise_left_shift_Tensor_out

- func: bitwise_left_shift.Scalar_out(Tensor self, Scalar other, *, Tensor(a!) out) -> Tensor(a!)

kernels:

- arg_meta: null

kernel_name: torch::executor::bitwise_left_shift_Scalar_out

- func: bitwise_right_shift.Tensor_out(Tensor self, Tensor other, *, Tensor(a!) out) -> Tensor(a!)

kernels:

- arg_meta: null

kernel_name: torch::executor::bitwise_right_shift_Tensor_out

- func: bitwise_right_shift.Scalar_out(Tensor self, Scalar other, *, Tensor(a!) out) -> Tensor(a!)

- op: bitwise_left_shift.Tensor_out

kernels:

- arg_meta: null

kernel_name: torch::executor::bitwise_left_shift_Tensor_out

- op: bitwise_left_shift.Scalar_out

kernels:

- arg_meta: null

kernel_name: torch::executor::bitwise_left_shift_Scalar_out

- op: bitwise_right_shift.Tensor_out

kernels:

- arg_meta: null

kernel_name: torch::executor::bitwise_right_shift_Tensor_out

- op: bitwise_right_shift.Scalar_out

Copilot · 2025-11-19T17:13:01Z

kernels/portable/functions.yaml

+- func: bitwise_left_shift.Tensor_out(Tensor self, Tensor other, *, Tensor(a!) out) -> Tensor(a!)
+  kernels:
+    - arg_meta: null
+      kernel_name: torch::executor::bitwise_left_shift_Tensor_out
+
+- func: bitwise_left_shift.Scalar_out(Tensor self, Scalar other, *, Tensor(a!) out) -> Tensor(a!)
+  kernels:
+    - arg_meta: null
+      kernel_name: torch::executor::bitwise_left_shift_Scalar_out
+
+- func: bitwise_right_shift.Tensor_out(Tensor self, Tensor other, *, Tensor(a!) out) -> Tensor(a!)
+  kernels:
+    - arg_meta: null
+      kernel_name: torch::executor::bitwise_right_shift_Tensor_out
+
+- func: bitwise_right_shift.Scalar_out(Tensor self, Scalar other, *, Tensor(a!) out) -> Tensor(a!)


The shift operations should use - op: instead of - func: to be consistent with the other bitwise operations (and, or, xor, not) which are ATen library operations. The - func: syntax is typically reserved for custom operations not in the ATen library.

Change to:

- op: bitwise_right_shift.Tensor_out kernels: - arg_meta: null kernel_name: torch::executor::bitwise_right_shift_Tensor_out

Suggested change

- func: bitwise_left_shift.Tensor_out(Tensor self, Tensor other, *, Tensor(a!) out) -> Tensor(a!)

kernels:

- arg_meta: null

kernel_name: torch::executor::bitwise_left_shift_Tensor_out

- func: bitwise_left_shift.Scalar_out(Tensor self, Scalar other, *, Tensor(a!) out) -> Tensor(a!)

kernels:

- arg_meta: null

kernel_name: torch::executor::bitwise_left_shift_Scalar_out

- func: bitwise_right_shift.Tensor_out(Tensor self, Tensor other, *, Tensor(a!) out) -> Tensor(a!)

kernels:

- arg_meta: null

kernel_name: torch::executor::bitwise_right_shift_Tensor_out

- func: bitwise_right_shift.Scalar_out(Tensor self, Scalar other, *, Tensor(a!) out) -> Tensor(a!)

- op: bitwise_left_shift.Tensor_out

kernels:

- arg_meta: null

kernel_name: torch::executor::bitwise_left_shift_Tensor_out

- op: bitwise_left_shift.Scalar_out

kernels:

- arg_meta: null

kernel_name: torch::executor::bitwise_left_shift_Scalar_out

- op: bitwise_right_shift.Tensor_out

kernels:

- arg_meta: null

kernel_name: torch::executor::bitwise_right_shift_Tensor_out

- op: bitwise_right_shift.Scalar_out

Copilot · 2025-11-19T17:13:01Z

shim_et/xplat/executorch/kernels/portable/op_registration_util.bzl

+    op_target(
+        name = "op_bitwise_left_shift",
+        deps = [
+            ":scalar_utils",
+            "//executorch/kernels/portable/cpu/pattern:bitwise_op",
+            "//executorch/kernels/portable/cpu/util:broadcast_util",
+            "//executorch/kernels/portable/cpu/util:dtype_util",
+            "//executorch/kernels/portable/cpu/util:elementwise_util",
+        ],
+    ),


The op_bitwise_left_shift target is placed out of alphabetical order. It should be positioned between op_bitwise_and (line 339) and op_bitwise_not (line 349) to maintain alphabetical ordering of the op targets.

Copilot · 2025-11-19T17:13:02Z

kernels/portable/functions.yaml

+- func: bitwise_left_shift.Tensor_out(Tensor self, Tensor other, *, Tensor(a!) out) -> Tensor(a!)
+  kernels:
+    - arg_meta: null
+      kernel_name: torch::executor::bitwise_left_shift_Tensor_out
+
+- func: bitwise_left_shift.Scalar_out(Tensor self, Scalar other, *, Tensor(a!) out) -> Tensor(a!)
+  kernels:
+    - arg_meta: null
+      kernel_name: torch::executor::bitwise_left_shift_Scalar_out
+
+- func: bitwise_right_shift.Tensor_out(Tensor self, Tensor other, *, Tensor(a!) out) -> Tensor(a!)
+  kernels:
+    - arg_meta: null
+      kernel_name: torch::executor::bitwise_right_shift_Tensor_out
+
+- func: bitwise_right_shift.Scalar_out(Tensor self, Scalar other, *, Tensor(a!) out) -> Tensor(a!)


The shift operations should use - op: instead of - func: to be consistent with the other bitwise operations (and, or, xor, not) which are ATen library operations. The - func: syntax is typically reserved for custom operations not in the ATen library.

Change to:

- op: bitwise_left_shift.Scalar_out kernels: - arg_meta: null kernel_name: torch::executor::bitwise_left_shift_Scalar_out

Suggested change

- func: bitwise_left_shift.Tensor_out(Tensor self, Tensor other, *, Tensor(a!) out) -> Tensor(a!)

kernels:

- arg_meta: null

kernel_name: torch::executor::bitwise_left_shift_Tensor_out

- func: bitwise_left_shift.Scalar_out(Tensor self, Scalar other, *, Tensor(a!) out) -> Tensor(a!)

kernels:

- arg_meta: null

kernel_name: torch::executor::bitwise_left_shift_Scalar_out

- func: bitwise_right_shift.Tensor_out(Tensor self, Tensor other, *, Tensor(a!) out) -> Tensor(a!)

kernels:

- arg_meta: null

kernel_name: torch::executor::bitwise_right_shift_Tensor_out

- func: bitwise_right_shift.Scalar_out(Tensor self, Scalar other, *, Tensor(a!) out) -> Tensor(a!)

- op: bitwise_left_shift.Tensor_out

kernels:

- arg_meta: null

kernel_name: torch::executor::bitwise_left_shift_Tensor_out

- op: bitwise_left_shift.Scalar_out

kernels:

- arg_meta: null

kernel_name: torch::executor::bitwise_left_shift_Scalar_out

- op: bitwise_right_shift.Tensor_out

kernels:

- arg_meta: null

kernel_name: torch::executor::bitwise_right_shift_Tensor_out

- op: bitwise_right_shift.Scalar_out

Copilot · 2025-11-19T17:13:02Z

kernels/portable/functions.yaml

+- func: bitwise_left_shift.Tensor_out(Tensor self, Tensor other, *, Tensor(a!) out) -> Tensor(a!)
+  kernels:
+    - arg_meta: null
+      kernel_name: torch::executor::bitwise_left_shift_Tensor_out
+
+- func: bitwise_left_shift.Scalar_out(Tensor self, Scalar other, *, Tensor(a!) out) -> Tensor(a!)
+  kernels:
+    - arg_meta: null
+      kernel_name: torch::executor::bitwise_left_shift_Scalar_out
+
+- func: bitwise_right_shift.Tensor_out(Tensor self, Tensor other, *, Tensor(a!) out) -> Tensor(a!)
+  kernels:
+    - arg_meta: null
+      kernel_name: torch::executor::bitwise_right_shift_Tensor_out
+
+- func: bitwise_right_shift.Scalar_out(Tensor self, Scalar other, *, Tensor(a!) out) -> Tensor(a!)


The shift operations should use - op: instead of - func: to be consistent with the other bitwise operations (and, or, xor, not) which are ATen library operations. The - func: syntax is typically reserved for custom operations not in the ATen library.

Change to:

- op: bitwise_right_shift.Scalar_out kernels: - arg_meta: null kernel_name: torch::executor::bitwise_right_shift_Scalar_out

Suggested change

- func: bitwise_left_shift.Tensor_out(Tensor self, Tensor other, *, Tensor(a!) out) -> Tensor(a!)

kernels:

- arg_meta: null

kernel_name: torch::executor::bitwise_left_shift_Tensor_out

- func: bitwise_left_shift.Scalar_out(Tensor self, Scalar other, *, Tensor(a!) out) -> Tensor(a!)

kernels:

- arg_meta: null

kernel_name: torch::executor::bitwise_left_shift_Scalar_out

- func: bitwise_right_shift.Tensor_out(Tensor self, Tensor other, *, Tensor(a!) out) -> Tensor(a!)

kernels:

- arg_meta: null

kernel_name: torch::executor::bitwise_right_shift_Tensor_out

- func: bitwise_right_shift.Scalar_out(Tensor self, Scalar other, *, Tensor(a!) out) -> Tensor(a!)

- op: bitwise_left_shift.Tensor_out

kernels:

- arg_meta: null

kernel_name: torch::executor::bitwise_left_shift_Tensor_out

- op: bitwise_left_shift.Scalar_out

kernels:

- arg_meta: null

kernel_name: torch::executor::bitwise_left_shift_Scalar_out

- op: bitwise_right_shift.Tensor_out

kernels:

- arg_meta: null

kernel_name: torch::executor::bitwise_right_shift_Tensor_out

- op: bitwise_right_shift.Scalar_out

kernels/portable/cpu/pattern/bitwise_op.h

pytorch-bot · 2025-11-19T17:20:20Z

🔗 Helpful Links

🧪 See artifacts and rendered test results at hud.pytorch.org/pr/pytorch/executorch/15893

📄 Preview Python docs built from this PR

Note: Links to docs will display an error until the docs builds have been completed.

❌ 2 New Failures, 2 Unrelated Failures

As of commit 6861b0a with merge base 6e4e8fa ():

NEW FAILURES - The following jobs have failed:

Lint / lintrunner / linux-job (gh)
>>> Lint for kernels/test/op_bitwise_right_shift_test.cpp:
pull / test-static-llama-qnn-linux (stories_260k_bc) / linux-job (gh)
RuntimeError: Command docker exec -t 4c5de8ceb5b60027b7864dd984b603a3456a2685269cfb1c012f90bc5eb53f0e /exec failed with exit code 1

BROKEN TRUNK - The following jobs failed but were present on the merge base:

👉 Rebase onto the `viable/strict` branch to avoid these failures

pull / unittest / linux / linux-job (gh) (trunk failure)
exir/backend/test/test_lowered_backend_module.py::TestBackendAPI::test_emit_nested_lowered_backend_module
pull / unittest-editable / linux / linux-job (gh) (trunk failure)
exir/backend/test/test_lowered_backend_module.py::TestBackendAPI::test_emit_nested_lowered_backend_module

This comment was automatically generated by Dr. CI and updates every 15 minutes.

Co-authored-by: Copilot <[email protected]>

WIP: Add bitwise left/right

b5c8bc8

meta-cla bot added the CLA Signed This label is managed by the Facebook bot. Authors need to sign the CLA before a PR can be reviewed. label Nov 19, 2025

mergennachin requested a review from Copilot November 19, 2025 17:06

Copilot started reviewing on behalf of mergennachin November 19, 2025 17:07 View session

Copilot finished reviewing on behalf of mergennachin November 19, 2025 17:10

Copilot AI reviewed Nov 19, 2025

View reviewed changes

mergennachin and others added 3 commits November 19, 2025 14:08

Update kernels/portable/cpu/pattern/bitwise_op.h

f174649

Co-authored-by: Copilot <[email protected]>

Update kernels/portable/cpu/op_bitwise_left_shift.cpp

6d1b752

Co-authored-by: Copilot <[email protected]>

Update kernels/portable/cpu/pattern/bitwise_op.h

6861b0a

Co-authored-by: Copilot <[email protected]>

WIP: Add bitwise left/right #15893

Are you sure you want to change the base?

WIP: Add bitwise left/right #15893

Uh oh!

Conversation

mergennachin commented Nov 19, 2025

Uh oh!

github-actions bot commented Nov 19, 2025

This PR needs a release notes: label

Uh oh!

mergennachin commented Nov 19, 2025

Uh oh!

meta-codesync bot commented Nov 19, 2025

Uh oh!

Copilot AI left a comment

Choose a reason for hiding this comment

Pull Request Overview

Reviewed Changes

Uh oh!

Uh oh!

Uh oh!

Copilot AI Nov 19, 2025

Choose a reason for hiding this comment

Uh oh!

Copilot AI Nov 19, 2025

Choose a reason for hiding this comment

Uh oh!

Copilot AI Nov 19, 2025

Choose a reason for hiding this comment

Uh oh!

Copilot AI Nov 19, 2025

Choose a reason for hiding this comment

Uh oh!

Copilot AI Nov 19, 2025

Choose a reason for hiding this comment

Uh oh!

Copilot AI Nov 19, 2025

Choose a reason for hiding this comment

Uh oh!

Uh oh!

pytorch-bot bot commented Nov 19, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

🔗 Helpful Links

🧪 See artifacts and rendered test results at hud.pytorch.org/pr/pytorch/executorch/15893

❌ 2 New Failures, 2 Unrelated Failures

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

2 participants

This PR needs a `release notes:` label

pytorch-bot bot commented Nov 19, 2025 •

edited

Loading