Repository navigation
Load Triton only when the CUDA profiler marker is used - #27
Open
huangxinV587 wants to merge 1 commit into
Open
huangxinV587 wants to merge 1 commit into
huangxinV587 wants to merge 1 commit into
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
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Importing tests.kernelkit eagerly imports Triton and constructs a CUDA profiler marker, preventing packaging/correctness helpers from importing in an Ascend environment without Triton.
Create the marker on first use and cache the JIT object. CUDA marker calls still require Triton and launch the same one-program kernel; repeated calls reuse the marker.
Validation: three subprocess tests passed using real CPU PyTorch and explicit Triton import rejection: package import without Triton, an explicit missing dependency when the CUDA marker is invoked, and one JIT construction with two marker launches using a fixture. Before the fix, the two missing-dependency cases failed. CUDA profiling and NPU kernels were not run.