install codegen header to torch/include #1405
Open
+346
−231
Conversation
This file contains 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
guangyey
force-pushed
the
guangyey/codegen
branch
15 times, most recently
from
ccff7d5 to
8620737
Compare
guangyey
force-pushed
the
guangyey/codegen
branch
from
8620737 to
c3e1df3
Compare
|
@guangyey : this does not seem to work for me, I still don't get headers installed to |
guangyey
force-pushed
the
guangyey/codegen
branch
8 times, most recently
from
d6aa4e3 to
242ad4e
Compare
guangyey
commented
Feb 26, 2025
dvrogozh
added a commit
to dvrogozh/pytorch
that referenced
this pull request
Feb 26, 2025
Commit extends existing CUDA test to cover XPU SyclExtension case for the same feature - `py_limited_api`. NOTE: THE CHANGE CAN NOT BE MERGED AS IS Change requires update of the commit pin for torch-xpu-ops. Requires: intel/torch-xpu-ops#1405 Signed-off-by: Dmitry Rogozhkin <[email protected]>
@guangyey : I see this resolved now after last changes. |
dvrogozh
reviewed
Feb 26, 2025
| @@ -1,89 +1,95 @@ | |||
| if(Codegen_GPU_cmake_included) | |||
| if(Codegen_XPU_cmake_included) | |||
There was a problem hiding this comment.
That's quite a change. I worry we might see more issues with torch code being updated and 2 things getting out of sync:
- torch-xpu-ops Codegen.cmake and related scripts with torch side versions of the same
- ops code in native/xpu folder which you've modified for some include files
Any chance we can start bringing pieces of this code into torch codebase itself? For example, any chance we can stop having Codegen.cmake here on torch-xpu-ops side?
Note: don't consider above as a request to do that in this PR. I am just trying to discuss.
guangyey
force-pushed
the
guangyey/codegen
branch
from
242ad4e to
0b53f33
Compare
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.
Motivation
This PR addresses a code generation issue related to XPU. Currently, there are two separate codegen paths for XPU:
The corresponding build directories are:
build/aten/src/ATen(for stock PyTorch)build/xpu/ATen(for torch-xpu-ops)However, in the torch-xpu-ops codegen, we mistakenly omitted installing XPU op headers from
build/xpu/ATen/opstobuild/aten/src/ATen/ops. This PR fixes the issue and also removes some unnecessary code for better maintainability.Solution
We copy the codegen from torch-xpu-ops to stock PyTorch
Additional Context
Fix pytorch/pytorch#145902