feat(ClusterFinderCUDA): async submit_batch/collect API

Browse Source

- Eliminate the ~200–300 µs inter-batch idle gap by allowing two batches
to be in-flight simultaneously:
  - submit_batch() enqueues H2D+kernel+D2H without blocking
  - collect() syncs via cudaEventSynchronize (not
  cudaStreamSynchronize) so a queued second batch runs uninterrupted.

- Two ping-pong output slots (NUM_SLOTS=2) with per-slot pinned buffers
and cudaEventDisableTiming sync events.
- find_clusters_batched() keeps its direct implementation.

* Measured: 0.026 -> 0.022 ms/frame (~18%).

...

This commit is contained in:

kferjaoui

2026-05-28 16:23:37 +02:00

parent 4c66802980

commit 5922c73c07

5 changed files with 597 additions and 95 deletions

Show all changes Ignore whitespace when comparing lines Ignore changes in amount of whitespace Ignore changes in whitespace at EOL

Download Patch File Download Diff File Expand all files Collapse all files

python/tests/ClusterFinderCUDA.ipynb

+141 -37

File diff suppressed because one or more lines are too long