leonarski_fandClaude Opus 5.5
18835f93f8
The two other points between merges where the engine's MergeAndStats arrays (~330 B per ASU group and ~55 B per full; ~3 GiB per P1 merge on a 400 A cell) sat beside other work: - the all-observation arm merged in series (the CPU path, or a card too small for the arm beside), whose arrays were held through that arm's SearchSpaceGroup; - after the answer's merge, through the analyses (a twin-law or ledger merge refills them). -N 10 --no-gpu (the macOS path's thread count): myob_x10sa 7.73 -> 7.28, cytc_x10sa 7.96 -> 7.56, 4nwv (timing benchmark) 20.34 -> 20.07 GiB peak RSS. Outputs byte-identical (p.hkl, MTZ column data, maps, placed model, report fields) at -N 10 on the CPU path and on the GPU path for myob_x10sa, cytc_x10sa, 4nwv, 8qaw. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01PXc1zD1yxHqaR1DGdCTHL4
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.