WIP: gpu support for TopK#42
Open
harz05 wants to merge 2 commits into
Open
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
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
This PR adds an initial GPU alpaka path for the TopK operator.
Approach: one thread per slice along the sorted axis. Each thread keeps a K-sized insertion-sorted buffer and selects its top-K in a single pass. largest/smallest, sorted, k and the strides are baked in at codegen, so the generated kernel has no attribute branches. Handles arbitrary axis (strided slices) and both largest and smallest; output is kept ordered to match the CPU
op, with smaller index winning on ties.
Test: added a TopK case to the alpaka test suite, reusing the existing TopK.onnx and its reference output.
WIP:
Currently working on implementin R-Topk, Warp Select for our use case