leonarski_f and Claude Opus 5.5 03dee2dda7
AzIntEngineGPU: four pixels per thread, one shared add per ring run ...
The standalone GPU azimuthal integration did three shared-memory atomics per
pixel on the same few ring addresses, which is what it was limited by. It now
reads four pixels per thread as vector loads and keeps a running total per ring,
flushed when the ring changes - the scheme the adaptive finder's ring pass
(reduce_rings_shared) already uses. The npix % 4 leftovers are done one at a time.
Used wherever the fused adaptive engine is not (fixed-threshold spot finding, the
broker's non-adaptive path). Measured on a 16 Mpx sweep (1800 frames,
--no-adaptive-spots, RTX 5080): 843 -> 295 us per call (min 621 -> 196 us).
Not bit-identical, and the old kernel was not either: float atomics arrive in any
order, so two runs of the OLD kernel already differ by up to 1.7e-6 relative in
the per-frame profile; new vs old differs by up to 1.9e-6, the same order. Per-ring
pixel counts are identical. Default rugnux runs do not reach this kernel (p.mtz
md5 unchanged on three sets); on the fixed-threshold path p_unmerged.mtz is
md5-identical to the old kernel's. New test: GPU vs CPU engine on a pixel count
that is not a multiple of four, with masked and saturated pixels.
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com >
Claude-Session: https://claude.ai/code/session_01K5K8jvPPbmCrbqnWkddTuB
2026-10-03 11:43:12 +02:00
2026-08-13 17:03:10 +02:00
2026-09-29 15:57:32 +02:00
2026-07-19 09:39:28 +02:00
2026-04-09 13:30:47 +02:00
2024-10-05 13:14:49 +02:00
2024-10-05 13:14:49 +02:00
2025-05-05 19:32:22 +02:00
2026-09-29 15:57:32 +02:00
2026-10-03 11:43:12 +02:00
2026-09-02 21:17:31 +02:00
2026-09-09 07:25:13 +02:00
2026-04-29 09:50:50 +02:00
2026-09-29 15:57:32 +02:00
2026-09-29 15:57:32 +02:00
2026-09-15 17:09:31 +02:00
2026-09-15 17:09:31 +02:00
2026-09-15 17:09:31 +02:00
2026-09-02 21:17:31 +02:00
2026-09-29 15:57:32 +02:00
2026-09-29 15:57:32 +02:00
2026-09-29 15:57:32 +02:00
2026-09-29 15:57:32 +02:00
2026-09-29 15:57:32 +02:00
2026-09-29 15:57:32 +02:00
2026-08-13 17:03:10 +02:00
2026-08-13 17:03:10 +02:00
2026-09-29 15:57:32 +02:00
2026-09-02 21:17:31 +02:00
2026-09-29 15:57:32 +02:00
2026-05-28 18:48:35 +02:00
2025-05-05 19:32:22 +02:00
2026-09-29 15:57:32 +02:00
2025-06-10 18:14:04 +02:00
2026-09-02 21:17:31 +02:00
2026-09-29 15:57:32 +02:00
2026-09-22 06:48:37 +02:00
2026-09-29 15:57:32 +02:00
2025-11-19 09:40:50 +01:00
2026-04-29 09:50:50 +02:00
2026-08-25 08:21:39 +02:00
2026-09-02 21:17:31 +02:00
2026-09-22 06:48:37 +02:00
2026-09-02 21:17:31 +02:00
2026-09-02 21:17:31 +02:00
2026-08-13 17:03:10 +02:00
2026-05-08 11:39:51 +02:00
2024-11-22 21:25:20 +01:00
2025-10-20 20:43:44 +02:00
2025-04-14 11:52:06 +02:00
2025-11-09 12:42:27 +01:00
2025-09-08 20:28:59 +02:00
2025-09-08 20:28:59 +02:00
2025-03-24 12:16:33 +01:00
2026-09-22 06:48:37 +02:00
2026-09-29 15:57:32 +02:00
2025-06-18 15:19:18 +02:00
2026-09-02 21:17:31 +02:00
2026-09-09 07:25:13 +02:00
2025-12-12 21:24:20 +01:00
2026-02-01 13:29:33 +01:00
2026-09-29 15:57:32 +02:00
2026-03-03 22:24:44 +01:00
2025-11-19 09:40:50 +01:00
2026-08-25 08:21:39 +02:00
2026-09-29 15:57:32 +02:00
2026-08-13 17:03:10 +02:00
2026-09-29 15:57:32 +02:00
2025-05-05 19:32:22 +02:00
2024-11-22 21:25:20 +01:00
2026-08-25 08:21:39 +02:00
2026-10-01 21:44:44 +02:00
2026-06-15 20:24:15 +02:00
2026-09-29 15:57:32 +02:00
2026-06-23 20:29:49 +02:00
2026-07-19 09:39:28 +02:00
2026-09-15 17:09:31 +02:00
2024-11-22 21:25:20 +01:00
2026-09-29 15:57:32 +02:00
2026-06-23 20:29:49 +02:00
2026-09-29 15:57:32 +02:00
2026-09-09 07:25:13 +02:00
2026-09-29 15:57:32 +02:00
2026-09-29 15:57:32 +02:00
2024-11-22 21:25:20 +01:00
2026-09-29 15:57:32 +02:00
2026-09-29 15:57:32 +02:00
2026-09-29 15:57:32 +02:00
2026-09-22 06:48:37 +02:00
2026-09-29 15:57:32 +02:00
2024-12-04 20:58:51 +01:00
2025-03-02 13:15:28 +01:00
2026-06-08 08:30:35 +02:00
2026-08-25 08:21:39 +02:00
2026-09-29 15:57:32 +02:00
2025-11-09 12:42:27 +01:00
2026-08-13 17:03:10 +02:00
2026-09-29 15:57:32 +02:00
2026-09-29 15:57:32 +02:00
2026-08-27 22:16:54 +02:00
2024-11-22 21:25:20 +01:00
2025-11-09 12:42:27 +01:00
2025-10-01 22:54:40 +02:00
2026-09-22 06:48:37 +02:00
2026-09-02 21:17:31 +02:00
2026-09-29 15:57:32 +02:00
2026-09-22 06:48:37 +02:00
2026-09-02 21:17:31 +02:00
2026-06-23 20:29:49 +02:00
2026-06-23 20:29:49 +02:00
2026-06-23 20:29:49 +02:00
2026-09-29 15:57:32 +02:00
2026-09-29 15:57:32 +02:00
2026-09-29 15:57:32 +02:00
2026-09-29 15:57:32 +02:00
2026-09-22 06:48:37 +02:00
2026-09-15 17:09:31 +02:00
2026-09-29 15:57:32 +02:00
2026-09-29 15:57:32 +02:00
2026-09-09 07:25:13 +02:00
2026-09-09 07:25:13 +02:00
2026-08-13 17:03:10 +02:00
2026-09-29 15:57:32 +02:00
2026-03-02 15:57:12 +01:00
2025-11-09 12:42:27 +01:00
2026-09-02 21:17:31 +02:00
2026-09-29 15:57:32 +02:00
2026-08-13 17:03:10 +02:00
2026-07-11 07:19:11 +02:00
2026-06-23 20:29:49 +02:00
2026-04-16 11:59:59 +02:00
2026-06-23 20:29:49 +02:00
2025-11-02 13:45:57 +01:00
2025-12-12 21:24:20 +01:00
2026-09-22 06:48:37 +02:00
2026-09-22 06:48:37 +02:00
2026-05-28 18:48:35 +02:00
2026-09-29 15:57:32 +02:00
2026-09-29 15:57:32 +02:00
2026-08-13 17:03:10 +02:00
2026-09-02 21:17:31 +02:00
2026-09-15 17:09:31 +02:00
2026-03-26 20:50:33 +01:00
2024-11-22 21:25:20 +01:00
2026-09-29 15:57:32 +02:00
leonarski_fandClaude Opus 5.5
03dee2dda7