Drop unused global_idx from logit kernel, run CUDA build whenever the CUDA source is changed - #5
Merged
Merged
Conversation
nvcc warned:
logit.cu(10): warning #177-D: variable "global_idx" was declared but
never referenced
It is dead rather than a missed offset. global_idx exists in the other kernels
to apply stepSize at the caller's index, and logit_kernel takes no stepSize --
its sibling sigmoid_kernel, same signature, correctly declares no such
variable. The data pointer the executor passes is already offset for the
chunk, so idx alone is right.
logit.cu was the only kernel with the problem. The CUDA build now compiles with
no nvcc warnings at all. Verified on a GPU: logit matches the CPU path to
8.9e-16 over 200k values, and 0, 1, out-of-domain inputs, nan and inf all
return nan.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
stub.yml is the only job that compiles the CUDA backend, and its path filter covered src/bindings, src/wrappers, include/fastdist, CMakeLists.txt and the stub itself -- but not src/cuda. A pull request touching only a .cu file got no CUDA compile check, so a broken kernel could land unnoticed. The preceding commit is exactly that shape. Adds src/cuda/** to the filter. The job needs no GPU: nvcc compiles the kernels without one, and the job only builds, imports, and checks that the *_cuda bindings exist. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
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
1 participant
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Drop unused global_idx from the logit kernel and run the CUDA build whenever the CUDA source changes