diff --git a/docs/JFJOCH_BROKER.md b/docs/JFJOCH_BROKER.md index aff7cb939..a3b4dd1d5 100644 --- a/docs/JFJOCH_BROKER.md +++ b/docs/JFJOCH_BROKER.md @@ -398,11 +398,11 @@ many as the pre-pass measured (at most 1000, as `rugnux`), while the per-image a The indexed frames are also merged, as `rugnux --mode scale` merges stills, in memory - in the consensus cell of the frames indexed by the first merge (the cell of the dataset settings only picks the frames that are merged) and the space group of the dataset settings (else P1): a space group given for the collection is binding. -At half and at three quarters of the images the frames so far are merged beside the analysis, on 8 threads, and +At half of the images the frames so far are merged beside the analysis, on 8 threads, and the result is reported as a provisional step note: crystals, resolution, CC1/2, R_split, completeness and -⟨I/σ⟩. The first of these merges scales and post-refines every crystal so far and decides the indexing +⟨I/σ⟩. This merge scales and post-refines every crystal so far and decides the indexing ambiguity from the data; every crystal after it is placed in its indexing, scaled and post-refined against -the latest of them as it arrives. After the last frame only the error model, the merge, the resolution cut +it as it arrives. After the last frame only the error model, the merge, the resolution cut and the statistics are left to do, on every thread - a few seconds for a chip of 160,000 frames on the full JUNGFRAU 8M, against about half a minute for the same merge made from nothing. The merged data travel to the writer in the end message, as for rotation (`.mtz`), and the run's step "Merge" carries the diff --git a/docs/RUGNUX_CLI.md b/docs/RUGNUX_CLI.md index f03a83c15..2232a954d 100644 --- a/docs/RUGNUX_CLI.md +++ b/docs/RUGNUX_CLI.md @@ -310,7 +310,7 @@ serial stills; `fft` (GPU) / `fftw` (CPU) index **de novo** and suit strong rota | `--no-scaling-corrections` | rot3d: disable the default-on decay + absorption + modulation correction surfaces fitted on the fulls after scale-fulls | | `--relative-b[=deg]` | rot3d: fit a per-batch relative-*B* beyond the single decay slope over deg-degree batches, cross-validated (default 10° when bare; off otherwise) | | `--simple-stills` | Stills: treat every reflection as a full (*p* = 1, single-pass scale/merge) — disables the default-on physical partiality post-refinement | -| `--live-checkpoints [,...]` | `--mode scale`, stills (for developers): replay the stored frames in image order as the broker's live merge of a serial collection merges them - a merge at each fraction of the images (e.g. `0.5,0.75`), each crystal after the first checkpoint scaled, post-refined and placed in its indexing against the latest checkpoint's merge as it comes, and after the last frame one merge of all of them without a scaling or post-refinement over the whole run. The log gives each checkpoint's statistics and time, and the time of the merge after the last frame; the merge written is that last one. With several `_process.h5` files the frames are replayed file after file, in the order given, as one collection. Without it `--mode scale` merges cold, as the broker does when no checkpoint is reached Each fraction 0.01-0.99. | +| `--live-checkpoints [,...]` | `--mode scale`, stills (for developers): replay the stored frames in image order as the broker's live merge of a serial collection merges them - a merge at each fraction of the images (the broker uses `0.5`), each crystal after the first checkpoint scaled, post-refined and placed in its indexing against the latest checkpoint's merge as it comes, and after the last frame one merge of all of them without a scaling or post-refinement over the whole run. The log gives each checkpoint's statistics and time, and the time of the merge after the last frame; the merge written is that last one. With several `_process.h5` files the frames are replayed file after file, in the order given, as one collection. Without it `--mode scale` merges cold, as the broker does when no checkpoint is reached Each fraction 0.01-0.99. | | `--live-final-full` | With `--live-checkpoints`: the merge after the last frame scales and post-refines every crystal over the whole run, as a cold merge does, instead of merging the crystals as they were prepared against the checkpoints - what the broker's live merge would do with a full scaling pass at the end; its time is in the log beside the incremental one's. | | `--no-expected-variance-merge` | Stills: disable the default expected-variance merge weighting (which rebuilds each weak observation's signal variance at the reflection mean to de-bias the inverse-variance merge); restores observed-sigma weighting | | `--capture-uncertainty ` | Systematic sigma num·(1−captured fraction)·⟨I⟩: on rot3d fulls the share of the rocking curve not captured, on stills the share of each partial its partiality does not see (default: 1.0 for rotation, 0 otherwise) | diff --git a/receiver/StillsGPUAnalysis.h b/receiver/StillsGPUAnalysis.h index cbb2f3dc1..a033b95fd 100644 --- a/receiver/StillsGPUAnalysis.h +++ b/receiver/StillsGPUAnalysis.h @@ -51,7 +51,7 @@ public: constexpr static int64_t PREPASS_IMAGES = 500; // The most a pre-pass run on held frames takes: the last run stays short, so the frames arriving meanwhile flow constexpr static int64_t PREPASS_MAX_IMAGES = 4 * PREPASS_IMAGES; - constexpr static std::array CHECKPOINTS = {0.5, 0.75}; + constexpr static std::array CHECKPOINTS = {0.5}; constexpr static int MERGE_THREADS = 8; // The merge after the last frame scales and post-refines every crystal over the whole run (StillsLiveMerge::Final): // the cold merge's result at the cold merge's cost after the last frame, instead of the fixed-cost merge of the diff --git a/tests/JFJochReceiverRugnuxTest.cpp b/tests/JFJochReceiverRugnuxTest.cpp index bc3b7cc46..a4a9595d1 100644 --- a/tests/JFJochReceiverRugnuxTest.cpp +++ b/tests/JFJochReceiverRugnuxTest.cpp @@ -191,7 +191,7 @@ TEST_CASE("JFJochReceiverRugnux_Stills", "[JFJochReceiverRugnux]") { CHECK(live.state == RugnuxLiveState::Done); CHECK(live.cut.empty()); CHECK(HasNote(live, "Merge of 6 of 12 images (provisional)")); - CHECK(HasNote(live, "Merge of 9 of 12 images (provisional)")); + CHECK(!HasNote(live, "Merge of 9 of 12 images (provisional)")); CHECK(HasNote(live, "Merge of the collection")); // The report's checklist of the merge: completeness looked at, powder rings not (no pre-scan here) const auto triage_line = [&](const std::string &label) {