PutMaskOnGrid() takes a `gpu` flag (default on, as PutModelDensityOnGrid's): on macOS the shrink by the 0.8 A radius (ShrinkMask) runs on the Apple GPU (ShrinkMaskMetal.mm/.metal), one thread per point applying the CPU's rule - a masked point becomes solvent where a near offset is solvent, or a far offset is a solvent point with a masked near offset - read from the mask as it was, with the CPU's offsets; the mask and its copy handed to the GPU in place where page aligned. The atoms, the symmetrize and gemmi's island removal stay on the CPU. --no-gpu, no device, or a failure (one warning) keep the CPU's loop. ShrinkMaskMetal_BitIdenticalToCPU ([portable][Metal]): clustered atoms, 0.4 A mask grids of a triclinic P1 and a P212121 cell - every point the same bits. M6 Mac mini, rugnux --mode refine (open-arm sets' merged MTZ and refined model), on top of the two previous Metal commits: 9s02 184.2 -> 175.6 s, 9zlo 41.3 -> 39.0 s; the refined models byte-identical to the build before on both. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01UqXYdJDrM65gxsCn2PrbfK
81 lines
3.3 KiB
C++
81 lines
3.3 KiB
C++
// SPDX-FileCopyrightText: 2026 Filip Leonarski, Paul Scherrer Institute <filip.leonarski@psi.ch>
|
|
// SPDX-License-Identifier: GPL-3.0-only
|
|
|
|
#include <catch2/catch_all.hpp>
|
|
|
|
#ifdef __APPLE__
|
|
|
|
#include <cstring>
|
|
#include <random>
|
|
#include <vector>
|
|
|
|
#include <gemmi/model.hpp>
|
|
|
|
#include "../image_analysis/structure_refinement/ModelGrid.h"
|
|
#include "../image_analysis/structure_refinement/RefinementXray.h"
|
|
#include "../image_analysis/structure_refinement/ShrinkMaskMetal.h"
|
|
|
|
// The bulk-solvent mask shrunk on the Apple GPU (ShrinkMaskMetal.h) is the CPU's, bit for bit. A Mac without a
|
|
// Metal device fails rather than skips, as in ModelDensityMetalTest.cpp.
|
|
|
|
namespace {
|
|
// Atoms in clusters over the cell - so that the mask has an edge to shrink everywhere and boxes wrap at every face.
|
|
gemmi::Model clustered_atoms(const gemmi::UnitCell &cell) {
|
|
std::mt19937 rng(29);
|
|
std::uniform_real_distribution<double> frac(0.0, 1.0);
|
|
std::normal_distribution<double> spread(0.0, 2.5);
|
|
const gemmi::El elements[] = {gemmi::El::C, gemmi::El::N, gemmi::El::O, gemmi::El::S};
|
|
gemmi::Model model(1);
|
|
gemmi::Chain chain("A");
|
|
gemmi::Residue res;
|
|
res.name = "UNK";
|
|
res.seqid = gemmi::SeqId(1, ' ');
|
|
for (int c = 0; c < 40; c++) {
|
|
const gemmi::Position centre = cell.orthogonalize(gemmi::Fractional(frac(rng), frac(rng), frac(rng)));
|
|
for (int i = 0; i < 50; i++) {
|
|
gemmi::Atom a;
|
|
a.name = "X";
|
|
a.element = elements[i % 4];
|
|
a.pos = centre + gemmi::Position(spread(rng), spread(rng), spread(rng));
|
|
a.b_iso = 20.0f;
|
|
a.occ = 1.0f;
|
|
res.atoms.push_back(a);
|
|
}
|
|
}
|
|
chain.residues.push_back(res);
|
|
model.chains.push_back(chain);
|
|
return model;
|
|
}
|
|
|
|
void check_same_mask(const gemmi::UnitCell &cell, const char *group, double d_min) {
|
|
const gemmi::SpaceGroup &sg = *gemmi::find_spacegroup_by_name(group);
|
|
gemmi::Grid<float> cpu = BulkSolventMask::MaskGrid(cell, sg, d_min);
|
|
cpu.data.resize(cpu.point_count());
|
|
gemmi::Grid<float> gpu = cpu;
|
|
const std::vector<size_t> leaders = OrbitLeaders(cpu, 4);
|
|
const gemmi::Model model = clustered_atoms(cell);
|
|
PutMaskOnGrid(cpu, model, leaders, 4, false);
|
|
const size_t before = ShrinkMaskMetalCalls();
|
|
PutMaskOnGrid(gpu, model, leaders, 4, true);
|
|
REQUIRE(ShrinkMaskMetalCalls() == before + 1); // the GPU shrank it
|
|
REQUIRE(gpu.data.size() == cpu.data.size());
|
|
size_t different = 0, solvent = 0;
|
|
for (size_t i = 0; i < cpu.data.size(); i++) {
|
|
different += std::memcmp(&cpu.data[i], &gpu.data[i], sizeof(float)) != 0;
|
|
solvent += cpu.data[i] >= 1.0f;
|
|
}
|
|
CAPTURE(group, solvent, cpu.data.size());
|
|
CHECK(solvent > cpu.data.size() / 10);
|
|
CHECK(solvent < cpu.data.size() * 9 / 10);
|
|
CHECK(different == 0);
|
|
}
|
|
}
|
|
|
|
TEST_CASE("ShrinkMaskMetal_BitIdenticalToCPU", "[portable][Metal]") {
|
|
// Triclinic, so that the offsets are not symmetric in the axes; then a group with symmetry.
|
|
check_same_mask(gemmi::UnitCell(41, 46, 52, 81, 96, 107), "P 1", 2.0);
|
|
check_same_mask(gemmi::UnitCell(50, 56, 63, 90, 90, 90), "P 21 21 21", 1.6);
|
|
}
|
|
|
|
#endif
|