leonarski_fandClaude Opus 5.5
de69e6d0d7
The demoted arm is a probe pass, run on the run itself before the canonical pass (the promoted arm). Its first pass measures a spot budget and writes it onto the experiment, and nothing put it back, so the canonical pass started from the probe's budget instead of the full spot list - measured, 446 spots per image instead of 1000 on one in-house crystal (validation spots 22055 instead of 43582) and 58 instead of 1000 on 6yqf. The same held where the arm lost: the demoted canonical pass started from the budget the stopped promoted pass measured. The demoted arm now runs on a copy of the run and gives back only the geometry its post-refinement asks for (which the promoted arm reads, as before) and its first-pass memo; where it wins, the demoted canonical pass starts from the experiment the promoted one started from. The indexing-settings and flag juggling around the probe goes. GPU, gpulock, shared machine, on the 22 sets whose baseline runs the check: p.hkl and MTZ reflection data identical to the base binary on 20; lcystine_x10sa_20keV fails with a cuFFT out-of-memory on the shared card with both binaries (four attempts). Changed: 6yqf only (its canonical first pass now reads the full spot list): merge 3.23 -> 3.15 A, R_meas 0.654 -> 0.568, ISa 4.97 -> 5.36; rugnux model R_free 0.459 -> 0.504 at its own d_min, phenix.model_vs_data R_free at the common 3.23 A 0.443 -> 0.435 (0.443 -> 0.447 at full resolution). The check's decision is unchanged on all 22. Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01EizBKhTqrAYago9KJG3dA3
Jungfraujoch
Application to receive and analyse data from the PSI JUNGFRAU and EIGER detectors, together with
rugnux, its program for processing stored diffraction data - rotation and serial, macromolecular
and small-molecule - from images to merged reflections.
All documentation is now placed in docs/ subdirectory and for the current version hosted on Jungfraujoch Read The Docs page.