Conversation
Adds optional CUDA support via Triton and PyTorch for the two
compute-heavy pipeline steps: PQ encoding (encoder.transform) and
cluster label assignment (clusterer.predict).
Both methods accept a `device` parameter ('auto'|'gpu'|'cpu').
The default 'auto' uses the GPU when available and falls back to
CPU transparently. All existing tests pass (102/102).
Benchmarked on 20M real molecules (K=100K, RTX 4070 Ti 16 GB):
- PQ Transform: 7.3 s GPU vs 45.3 s CPU (6.2x)
- Cluster Assignment: 29.9 s GPU vs ~879 s CPU (29.4x)
- Combined for 9.6B: ~5 h GPU vs ~123 h CPU (~25x)
New files:
chelombus/clustering/_gpu_predict.py — Triton kernel
scripts/benchmark_gpu_predict.py — GPU vs CPU benchmark
scripts/cluster_smiles.py — end-to-end pipeline script
data/10M_smiles.txt.gz — test SMILES (85 MB gzip)
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Adds optional CUDA support via Triton and PyTorch for the two compute-heavy pipeline steps: PQ encoding (encoder.transform) and cluster label assignment (clusterer.predict).
Both methods accept a
deviceparameter ('auto'|'gpu'|'cpu'). The default 'auto' uses the GPU when available and falls back to CPU transparentlyBenchmarked on 20M real molecules (K=100K, RTX 4070 Ti 16 GB):