2 Commits

Author	SHA1	Message	Date
kferjaoui	3ed773e520	Add multi-stream ClusterFinderCUDA with batched processing ... - Wrap per-stream CUDA resources (device buffers, stream handle) in StreamContext struct; ClusterFinderCUDA owns a vector of n_streams contexts with independent pedestal arrays - Split ClusterFinderCUDA.cuh into clusterfinder_kernel.cuh (device kernel) and ClusterFinderCUDA.hpp (host RAII wrapper) - Add find_clusters_batched(): processes N frames round-robin across streams, returns per-frame cluster vectors. - Update ClusterFinderCUDA.test.cu - Update Makefile for new file layout.	2026-04-23 11:26:29 +02:00
kferjaoui	69151de3c7	Add in-kernel pedestal update, disable quadrant test ... - Non-photon pixels now update pedestal (push_fast equivalent) directly in the kernel, no atomics needed - Commented out quadrant significance test (c2): absent from sequential CPU code, was producing GPU-only clusters. - Added d_pd_sum to device allocations and host upload Build (sm_89): 46 registers, 0 spills, 100% occupancy. Verified on 256x256 Jungfrau data, 5000 frames, nSigma=5.0: CPU 8428 vs GPU 8471 clusters, 99.8% match 0.63 ms/frame CPU vs 0.04 ms/frame GPU (~16x)	2026-04-13 11:28:03 +02:00