make aoti_torch_empty_strided support creating incontiguous tensor #15228
+373
−44
Conversation
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
This diff modifies the `aoti_torch_empty_strided` function to support the creation of incontiguous tensors. To achieve it, this diff: 1. update the way to calculate the memory size by using both tensor size and the stride 2. skip stride check in ETensor by adding and checking cmake macro `USE_CUDA_BACKEND` when building with CUDA backend support. we will soon bring the ETensor check back for every backend after migrating to use slimtensor. Differential Revision: [D84938258](https://our.internmc.facebook.com/intern/diff/D84938258/) [ghstack-poisoned]
Gasoonjia
requested review from
kirklandsign,
larryliu0820 and
shoumikhin
as code owners
🔗 Helpful Links🧪 See artifacts and rendered test results at hud.pytorch.org/pr/pytorch/executorch/15228
Note: Links to docs will display an error until the docs builds have been completed. ❗ 1 Active SEVsThere are 1 currently active SEVs. If your PR is affected, please view them below: ❌ 2 New Failures, 3 Unrelated FailuresAs of commit a1c83d3 with merge base 9b744b4 ( NEW FAILURES - The following jobs have failed:
BROKEN TRUNK - The following jobs failed but were present on the merge base:👉 Rebase onto the `viable/strict` branch to avoid these failures
This comment was automatically generated by Dr. CI and updates every 15 minutes. |
Gasoonjia
added a commit
that referenced
this pull request
Oct 17, 2025
This diff modifies the `aoti_torch_empty_strided` function to support the creation of incontiguous tensors. To achieve it, this diff: 1. update the way to calculate the memory size by using both tensor size and the stride 2. skip stride check in ETensor by adding and checking cmake macro `USE_CUDA_BACKEND` when building with CUDA backend support. we will soon bring the ETensor check back for every backend after migrating to use slimtensor. Differential Revision: [D84938258](https://our.internmc.facebook.com/intern/diff/D84938258/) ghstack-source-id: 317035163 Pull Request resolved: #15228
meta-cla
bot
added
the
CLA Signed
This PR needs a
|
…s tensor" This diff modifies the `aoti_torch_empty_strided` function to support the creation of incontiguous tensors. To achieve it, this diff: 1. update the way to calculate the memory size by using both tensor size and the stride 2. skip stride check in ETensor by adding and checking cmake macro `USE_CUDA_BACKEND` when building with CUDA backend support. we will soon bring the ETensor check back for every backend after migrating to use slimtensor. Differential Revision: [D84938258](https://our.internmc.facebook.com/intern/diff/D84938258/) [ghstack-poisoned]
Gasoonjia
added a commit
that referenced
this pull request
Oct 17, 2025
Pull Request resolved: #15228 This diff modifies the `aoti_torch_empty_strided` function to support the creation of incontiguous tensors. To achieve it, this diff: 1. update the way to calculate the memory size by using both tensor size and the stride 2. skip stride check in ETensor by adding and checking cmake macro `USE_CUDA_BACKEND` when building with CUDA backend support. we will soon bring the ETensor check back for every backend after migrating to use slimtensor. ghstack-source-id: 317086685 @exported-using-ghexport Differential Revision: [D84938258](https://our.internmc.facebook.com/intern/diff/D84938258/)
msluszniak
reviewed
Oct 20, 2025
Comment on lines
234
to
237
| int64_t stride_i = (strides_ptr != nullptr) ? strides_ptr[i] : 0; | ||
| if (strides_ptr == nullptr) { | ||
| // Calculate contiguous stride if not provided | ||
| stride_i = 1; |
There was a problem hiding this comment.
Suggested change
| int64_t stride_i = (strides_ptr != nullptr) ? strides_ptr[i] : 0; | |
| if (strides_ptr == nullptr) { | |
| // Calculate contiguous stride if not provided | |
| stride_i = 1; | |
| int64_t stride_i = (strides_ptr != nullptr) ? strides_ptr[i] : 1; | |
| if (strides_ptr == nullptr) { | |
| // Calculate contiguous stride if not provided |
…s tensor" This diff modifies the `aoti_torch_empty_strided` function to support the creation of incontiguous tensors. To achieve it, this diff: 1. update the way to calculate the memory size by using both tensor size and the stride 2. skip stride check in ETensor by adding and checking cmake macro `USE_CUDA_BACKEND` when building with CUDA backend support. we will soon bring the ETensor check back for every backend after migrating to use slimtensor. Differential Revision: [D84938258](https://our.internmc.facebook.com/intern/diff/D84938258/) [ghstack-poisoned]
…s tensor" This diff modifies the `aoti_torch_empty_strided` function to support the creation of incontiguous tensors. To achieve it, this diff: 1. update the way to calculate the memory size by using both tensor size and the stride 2. skip stride check in ETensor by adding and checking cmake macro `USE_CUDA_BACKEND` when building with CUDA backend support. we will soon bring the ETensor check back for every backend after migrating to use slimtensor. Differential Revision: [D84938258](https://our.internmc.facebook.com/intern/diff/D84938258/) [ghstack-poisoned]
…s tensor" This diff modifies the `aoti_torch_empty_strided` function to support the creation of incontiguous tensors. To achieve it, this diff: 1. update the way to calculate the memory size by using both tensor size and the stride 2. skip stride check in ETensor by adding and checking cmake macro `USE_CUDA_BACKEND` when building with CUDA backend support. we will soon bring the ETensor check back for every backend after migrating to use slimtensor. Differential Revision: [D84938258](https://our.internmc.facebook.com/intern/diff/D84938258/) [ghstack-poisoned]
Gasoonjia
added a commit
that referenced
this pull request
Oct 21, 2025
Pull Request resolved: #15228 This diff modifies the `aoti_torch_empty_strided` function to support the creation of incontiguous tensors. To achieve it, this diff: 1. update the way to calculate the memory size by using both tensor size and the stride 2. skip stride check in ETensor by adding and checking cmake macro `USE_CUDA_BACKEND` when building with CUDA backend support. we will soon bring the ETensor check back for every backend after migrating to use slimtensor. ghstack-source-id: 317525743 @exported-using-ghexport Differential Revision: [D84938258](https://our.internmc.facebook.com/intern/diff/D84938258/)
extension/tensor/tensor_ptr.cpp
Outdated
| // Skip stride calculation and incontiguous tensor check for CUDA backend since | ||
| // AOTI-CUDA handles both contiguous and incontiguous tensors. This will be | ||
| // removed after SlimTensor migration. | ||
| #ifndef USE_CUDA_BACKEND |
There was a problem hiding this comment.
This is not needed anymore?
There was a problem hiding this comment.
you are absolutely right. Removed.
…s tensor" This diff modifies the `aoti_torch_empty_strided` function to support the creation of incontiguous tensors. To achieve it, this diff: 1. update the way to calculate the memory size by using both tensor size and the stride 2. skip stride check in ETensor by adding and checking cmake macro `USE_CUDA_BACKEND` when building with CUDA backend support. we will soon bring the ETensor check back for every backend after migrating to use slimtensor. Differential Revision: [D84938258](https://our.internmc.facebook.com/intern/diff/D84938258/) [ghstack-poisoned]
…s tensor" This diff modifies the `aoti_torch_empty_strided` function to support the creation of incontiguous tensors. To achieve it, this diff: 1. update the way to calculate the memory size by using both tensor size and the stride 2. skip stride check in ETensor by adding and checking cmake macro `USE_CUDA_BACKEND` when building with CUDA backend support. we will soon bring the ETensor check back for every backend after migrating to use slimtensor. Differential Revision: [D84938258](https://our.internmc.facebook.com/intern/diff/D84938258/) [ghstack-poisoned]
…s tensor" This diff modifies the `aoti_torch_empty_strided` function to support the creation of incontiguous tensors. To achieve it, this diff: 1. update the way to calculate the memory size by using both tensor size and the stride 2. skip stride check in ETensor by adding and checking cmake macro `USE_CUDA_BACKEND` when building with CUDA backend support. we will soon bring the ETensor check back for every backend after migrating to use slimtensor. Differential Revision: [D84938258](https://our.internmc.facebook.com/intern/diff/D84938258/) [ghstack-poisoned]
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
4 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):
This diff modifies the
aoti_torch_empty_stridedfunction to support the creation of incontiguous tensors.To achieve it, this diff:
USE_CUDA_BACKENDwhen building with CUDA backend support.we will soon bring the ETensor check back for every backend after migrating to use slimtensor.
Differential Revision: D84938258