cuml icon indicating copy to clipboard operation
cuml copied to clipboard

DBSCAN utilize rbc eps_neighbors

Open mfoerste4 opened this issue 1 year ago • 2 comments

This PR enables rbc eps-neighbor computation via raft. The resulting adjacency matrix is sparse and allows to skip the implicit conversion.

Notes:

  • the 'algorithm'-parameter was added to the DBSCAN init signature to allow the user to choose (default is 'brute', 'rbc' is optional)
  • the memory management is still very conservative, assuming a dense adjacency matrix and therefore selecting comparably small batches
  • in case maximum row length of a batch is sufficiently small the CSR structure can be computed in a single pass

CC @tfeher

mfoerste4 avatar Jan 23 '24 23:01 mfoerste4

rerun tests

mfoerste4 avatar Jan 24 '24 10:01 mfoerste4

@tfeher , thanks for reviewing. I just added a small correction and re-triggered the CI.

mfoerste4 avatar Feb 22 '24 14:02 mfoerste4

/test

mfoerste4 avatar Mar 08 '24 17:03 mfoerste4

rerun tests

mfoerste4 avatar Mar 08 '24 21:03 mfoerste4

/merge

dantegd avatar Mar 11 '24 13:03 dantegd