MG KMeans initialization emits unnecessary warnings
Nobody has claimed this yet.
- Dominant language
- Cuda
- Stars
- 854
- Forks
- 236
- Avg merge
- 3d 3h
- Merged PRs (30d)
- 62
Description
The multi-GPU KMeans initialization and more specifically the reclustering step of the KMeans|| initialization method emits unnecessary/confusing warnings when used in MNMG mode (RAFT communicator).
This is caused by the fact that the reclustering step of the KMeans|| initialization method calls the public MG KMeans API interface that handles SG/MG dispatch.
The warning :
Multi-GPU handle detected on single-GPU kmeans::fit() entry; falling back to single-GPU. Use cuvs::cluster::kmeans::mg::fit(...) for multi-GPU.
Contributor guide
First steps
- Read the whole issue, then the project's contributing guide.
- Comment on the issue to say you are picking it up — it saves two people doing the same work.
- Fork the repository and make your change on a branch.
- Open a pull request that references the issue number.
Research direction
Start in cpp/src/cluster/detail/kmeans_mg_distributed_init.cuh at the reclustering call around line 444, where KMeans|| initialization uses the public MG KMeans API interface. Trace the SG/MG dispatch in that path and verify MNMG initialization no longer emits the single-GPU fallback warning while retaining the correct multi-GPU behavior.
Written by the indexing model from the issue text.
Assessment
- Tech stack
- cpp, machine-learning
- Domain
- distributed-systems, machine-learning
- Issue type
- Bug
- Difficulty
- 3/5
- Estimated time
- 1-2 days
- Activity status
- Quiet
- Clarity
- Mostly clear
- Newbie friendliness
- 56/100