- Notifications
You must be signed in to change notification settings - Fork26.3k
Fix TorchCPP Ext accelerator detection ROCm#170784
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to ourterms of service andprivacy statement. We’ll occasionally send you account related emails.
Already on GitHub?Sign in to your account
Open
thiagocrepaldi wants to merge1 commit intomainChoose a base branch
base:main
Could not load branches
Branch not found:{{ refName }}
Loading
Could not load tags
Nothing to show
Loading
Are you sure you want to change the base?
Some commits from the old base branch may be removed from the timeline, and old review comments may become outdated.
Open
+11 −3
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
pytorch-botbot commentedDec 18, 2025 • edited
Loading Uh oh!
There was an error while loading.Please reload this page.
edited
Uh oh!
There was an error while loading.Please reload this page.
🔗 Helpful Links🧪 See artifacts and rendered test results athud.pytorch.org/pr/170784
Note: Links to docs will display an error until the docs builds have been completed. ✅ No FailuresAs of commitc1549ae with merge base562a337 ( This comment was automatically generated by Dr. CI and updates every 15 minutes. |
linux-foundation-easyclabot commentedDec 18, 2025 • edited
Loading Uh oh!
There was an error while loading.Please reload this page.
edited
Uh oh!
There was an error while loading.Please reload this page.
|
86b7930 to56d59f7CompareCollaboratorAuthor
thiagocrepaldi commentedDec 18, 2025
@pytorchbot label "topic: not user facing" |
56d59f7 to4c23370Comparemalfet approved these changesDec 18, 2025
Uh oh!
There was an error while loading.Please reload this page.
Uh oh!
There was an error while loading.Please reload this page.
Uh oh!
There was an error while loading.Please reload this page.
4fa05bc to65f5c63ComparePreviously, _get_build_directory only checked for CUDA and defaulted to'cpu' for all other cases, causing build directory conflicts whenextensions were built for different GPU accelerators (ROCm/HIP, SYCL/XPU).This fix extends accelerator detection to properly handle:- ROCm/HIP: Checks torch.version.hip first (highest priority)- CUDA: Uses torch.backends.cuda.is_built() and torch.version.cuda- SYCL/XPU: Checks torch.xpu._is_compiled() and torch.version.xpu- CPU: Default fallbackThe build directory now includes the correct accelerator prefix(e.g., py310_rocm542, py310_cu118, py310_xpu2024, py310_cpu),preventing conflicts between extensions built for differentaccelerators on the same system.Priority order: ROCm/HIP > CUDA > SYCL/XPU > CPU
65f5c63 toc1549aeCompareCollaboratorAuthor
thiagocrepaldi commentedDec 18, 2025
is it possible to target it to milestone 2.10 due in January? |
Sign up for freeto join this conversation on GitHub. Already have an account?Sign in to comment
Labels
ciflow/rocm-mi300Trigger "default" config CI on ROCm MI300 module: cpp-extensionsRelated to torch.utils.cpp_extension module: rocmAMD GPU support for Pytorch open sourcetopic: not user facingtopic category
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.
Uh oh!
There was an error while loading.Please reload this page.
Previously, _get_build_directory only checked for CUDA and defaulted to 'cpu' for all other cases, causing build directory conflicts when extensions were built for different GPU accelerators (ROCm/HIP, SYCL/XPU).
This fix extends accelerator detection to properly handle:
The build directory now includes the correct accelerator prefix (e.g., py310_rocm542, py310_cu118, py310_xpu2024, py310_cpu), preventing conflicts between extensions built for different accelerators on the same system.
Priority order: ROCm/HIP > CUDA > SYCL/XPU > CPU
Fixes#170778
cc@janeyx99@jeffdaily@sunway513@jithunnair-amd@pruthvistony@ROCmSupport@jataylo@hongxiayang@naromero77amd@pragupta@jerrymannil@xinyazhang