[FEA] Integrate balanced KMeans

[tfeher]

The approximate nearest neighbors algorithms (IVF-Flat #652, IVF-PQ) use a balanced k-means implementation. In #652, balanced kmeans is separate implementation from #605. These two implementations should be merged to avoid code duplications.

One should ensure that the changes are benchmarked, not to loose performance for the approximate nearest neighbors methods.

Tasks

• [x] move k-means benchmarks from cuML #795
• [x] add balanced k-means benchmark #795
• [x] define public API for balanced k-means #1113
• [x] reduce duplication in primitive functions (various PRs)

About reducing the duplication of primitive functions, the following issues came up earlier:

• [x] https://github.com/rapidsai/raft/pull/652#discussion_r915038999 copyRows
• [x] https://github.com/rapidsai/raft/pull/652#discussion_r915047790 #912
• [x] https://github.com/rapidsai/raft/pull/652#discussion_r915050104 #911
• [x] https://github.com/rapidsai/raft/pull/652#discussion_r915052244 #909
• [ ] compare minClusterAnddistanceComputeand _cuann_kmeans_predict[_core],
• [ ] replace _cuann_kmeans_update_centers[outdated] with reduce_cols_by_keyor reduce_rows_by_key,

[github-actions[bot]]

This issue has been labeled inactive-30d due to no recent activity in the past 30 days. Please close this issue if no further response or action is needed. Otherwise, please respond with a comment indicating any updates or changes to the original issue and/or confirm this issue still needs to be addressed. This issue will be labeled inactive-90d if there is no activity in the next 60 days.

[Nyrio]

move k-means benchmarks from cuML

Can be marked as done.

[tfeher]

Could you open a PR to remove https://github.com/rapidsai/cuml/blob/branch-22.12/cpp/bench/sg/kmeans.cu?

[Nyrio]

@tfeher Perhaps that should wait until we've merged https://github.com/rapidsai/cuml/pull/4713 that replaces the cuML implementation with the raft implementation?