From ca0a6dd343b7aee88075ab7704678c8d2059b564 Mon Sep 17 00:00:00 2001
From: Filip Leonarski <filip.leonarski@psi.ch>
Date: Mon, 26 Oct 2026 07:57:30 +0100
Subject: [PATCH] Refinement: each slot's likelihood fit on all of the run's
 threads

Bitwise (the bins are fitted independently: the same bits at any thread count). The weight-trial slots Refit
in lockstep while the GPU idles, each with its 1/slots share of -N for the ML bins; the bins now go to the
shared worker pool with -N tasks per slot, so the threads of slots that are not fitting help the ones that are.
After the Newton searches (previous commit) the ML fit is 0.1-0.5 s summed over a run, so the wall gain is
within noise: 5reo 1.63-1.75 -> 1.72, 5f6m 2.60 -> 2.57, 9yl4 7.34 -> 7.28, 7d98 3.78 -> 3.78, 5f6m MD start
4.82 -> 4.76 s. Refined CIF md5 unchanged on 5reo, 5f6m, 9yl4, 7d98, 5f6m MD start (GPU) and 6o2h (CPU).

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_015CeEGgpKdfkCTJb3YNMJcu
(cherry picked from commit 56815f5390cc01e8da5d420fff73cdf4394c385c)
---
 image_analysis/structure_refinement/Refinement.cpp | 7 ++++++-
 1 file changed, 6 insertions(+), 1 deletion(-)

diff --git a/image_analysis/structure_refinement/Refinement.cpp b/image_analysis/structure_refinement/Refinement.cpp
index 60846728d..760b19f60 100644
--- a/image_analysis/structure_refinement/Refinement.cpp
+++ b/image_analysis/structure_refinement/Refinement.cpp
@@ -118,6 +118,9 @@ struct Context {
     std::unique_ptr<ModelScaleGPUPool> scale_gpu;
 #endif
     size_t nthreads = 1;
+    // the likelihood fit's: all of the run's, also in a trial's slot - the slots refit at the same time, the GPU
+    // idles meanwhile, and the bins they queue on the shared worker pool are fitted by whichever thread is free
+    size_t ml_threads = 1;
     int device = -1;            // the engine's GPU
     size_t engine_bytes = 0;    // and what it holds there
     XrayRowModel rows;
@@ -260,7 +263,7 @@ struct Context {
         }
         {
             const RefinementStage s4("refit.ml");
-            EstimateLikelihoodParameters(refl, fmodel, rows, nthreads);
+            EstimateLikelihoodParameters(refl, fmodel, rows, ml_threads);
         }
         engine->SetRowModel(rows);
         return ComputeRFactors(refl, fmodel);
@@ -967,6 +970,7 @@ std::unique_ptr<OwnSlot> MakeOwnSlot(const gemmi::Structure &before_topology, co
     ctx.d_min = main.d_min;
     ctx.refl = main.refl;
     ctx.nthreads = nthreads;
+    ctx.ml_threads = settings.nthreads;
     ctx.cancel = main.cancel;
     ctx.b_ceiling = main.b_ceiling;
     ctx.mask = std::make_unique<BulkSolventMask>(ctx.cell, *ctx.sg, ctx.d_min, nthreads);
@@ -1158,6 +1162,7 @@ RefinementResult RefineModel(const gemmi::Structure &placed, const std::vector<M
     ctx.cell = cell;
     ctx.sg = &sg;
     ctx.nthreads = settings.nthreads;
+    ctx.ml_threads = settings.nthreads;
     ctx.cancel = settings.cancel;
     ctx.atoms = AtomList(ctx.Model());
     result.atoms = static_cast<int>(ctx.atoms.size());