KMeans centroid averaging on GPUs
Open
CUDA
enhancement
kmeans
performance
- Dominant language
- C++
- Stars
- 468
- Forks
- 96
- PR merge metrics
- No merged PRs in 30d
Description
If centroid averaging becomes a bottleneck and/or we want to stay on the GPUs 100% of the time we can research if averaging centroids from multiple GPUs would be better on one of the GPUs instead of the host (as is done now).
Contributor guide
Assessment
This issue has not been assessed yet.