Live stills: 16 GPU analysis threads a card, whatever receiver_threads is
Build Packages / Create release (push) Successful in 8s
Build Packages / Build documentation (push) Successful in 49s
Build Packages / build:portable:macos-arm64:nocuda (push) Successful in 5m10s
Build Packages / build:rugnux:linux-aarch64:cuda (push) Successful in 11m6s
Build Packages / build:portable:linux-x86_64:nocuda (push) Successful in 16m20s
Build Packages / build:portable:linux-x86_64:cuda (push) Successful in 22m16s
Build Packages / build:portable:windows-x86_64:nocuda (push) Successful in 25m29s
Build Packages / build:jfjoch:ubuntu2204:nocuda (push) Successful in 26m0s
Build Packages / HDF5 consumer tests (DIALS, XDS) (push) Successful in 29m15s
Build Packages / build:portable:windows-x86_64:cuda (push) Successful in 29m36s
Build Packages / build:jfjoch:rocky8:nocuda (push) Successful in 31m38s
Build Packages / build:jfjoch:rocky9:nocuda (push) Successful in 32m36s
Build Packages / build:jfjoch:ubuntu2404:nocuda (push) Successful in 28m53s
Build Packages / Generate python client (push) Successful in 1m6s
Build Packages / build:jfjoch:rocky8:cuda-sls9 (push) Successful in 33m36s
Build Packages / build:jfjoch:rocky9:cuda-sls9 (push) Successful in 35m43s
Build Packages / build:jfjoch:ubuntu2204:cuda (push) Successful in 26m27s
Build Packages / build:jfjoch:rocky8:cuda (push) Successful in 33m30s
Build Packages / Unit tests (push) Successful in 1h0m47s
Build Packages / build:jfjoch:ubuntu2404:cuda (push) Successful in 30m37s
Build Packages / build:jfjoch:rocky9:cuda (push) Successful in 36m11s
Build Packages / Validation Battery (push) Successful in 32m4s

The rugnux live analysis of a serial collection took min(receiver_threads,
16 per card), so a site running 16 receiver threads on four cards analysed
on four threads a card. It now runs 16 a card (fewer where a card's memory
does not hold them), as offline rugnux does. Measured on a 10.7 Mpx stills
replay: 8 a card saturates, 16 is the fastest, 32 and 64 are 4-9 % slower.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01HVeDQrPqZ52Tp6rzzZGxf3
This commit is contained in:
leonarski_fandClaude Opus 5.5 committed 2026-11-19 16:07:41 +01:00
1 parent c9f0764233
commit d82fe5dfa6
2 files changed
+8 -3

No files matched your search

+3 -1
View File
@@ -230,7 +230,9 @@ goniometer and the rugnux live setting:
`receiver_threads` is the number of image threads. The Lite receiver uses at most 16 of them for a
collection at 100 Hz or slower or of fewer than 256 images (each thread holds a GPU analysis of its
own); the FPGA receiver takes four when it sums images on the host. The GPU FFT indexers and cuFFT
own); the FPGA receiver takes four when it sums images on the host. The rugnux live analysis of a serial
collection does not take its threads from `receiver_threads`: it runs 16 GPU analysis threads a card (fewer on a
card whose memory does not hold 16 of them). The GPU FFT indexers and cuFFT
plans the broker's own indexing uses are built once and kept for the life of the process, and the ones
built for earlier indexing settings are released when the settings change. `azim_int.force_cpu`
moves the azimuthal integration of the FPGA workflow from the card to the host, which a profile with
+5 -2
View File
@@ -6,6 +6,7 @@
#include <cstring>
#include <thread>
#include "../common/CUDAWrapper.h"
#include "../common/time_utc.h"
#include "../frame_serialize/CBORStream2Serializer.h"
#include "../rugnux/Rugnux.h"
@@ -42,8 +43,10 @@ JFJochReceiverRugnuxStills::JFJochReceiverRugnuxStills(const DiffractionExperime
prepass_images(in_prepass_images),
live_status(in_live_status),
indexing_thread_pool(in_indexing_thread_pool),
// As offline rugnux's per-image loop: 16 a card, at most the threads given
stills_nthreads(ImageWorkers(in_experiment.GetPixelsNum(), static_cast<int>(in_forward_nthreads))) {
// As offline rugnux's per-image loop: 16 a card (fewer where a card's memory does not hold them), whatever
// the receiver's own thread count
stills_nthreads(get_gpu_count() > 0 ? GPUWorkersPerCard(in_experiment.GetPixelsNum(), 16) * get_gpu_count()
: static_cast<int>(in_forward_nthreads)) {
StartThreads();
}