The two pre-scan steps that were still CPU-bound in a GPU build now run where the projection already is. - FindBeamCenterFromBackground: the per-iteration binning pass and the two clip rounds run on the device (BeamCenterBackgroundGPU); the fit itself stays on the host. Each cell is summed in the host's order (pixel order within the host's row blocks, blocks in order), and the per-pixel cell/derivative formula is shared (BackgroundBand.h). The angles come from BackgroundAtan2 (IEEE ops only) instead of atan2f, and both translation units are compiled without FMA contraction, so host and device give the same bits: 0 of 6.5 M pixels in a different cell, identical walks on the three in-house rotation sets. With glibc/CUDA atan2f and default contraction ~30 pixels per 16 Mpx sweep changed cell and the fitted centre moved by up to 0.05 px. - ShadowFinder::GetMask: the whole mask (pooling, ring medians, components, morphology, hole fill, arm search) runs on the device from ShadowAccumulatorGPU's projection (ShadowMaskGPU), so the 360 MB projection no longer comes back; the mean projection is divided on the device too (same bits). The two small fits over rings and sectors (BlockedOutTo, HarmonicFit) are shared with the host path in ShadowFinderInternal.h. Integers, comparisons, sorts and components are exact; the polarization trig, the Poisson log and the arm-search azimuth are not, so a pixel at a threshold can differ. The one-time change against the previous CPU arithmetic (BackgroundAtan2, no contraction), measured on the myoglobin, cytochrome C and thaumatin rotation sets: ring centre moves 0.002-0.045 px (fit sigma 0.75-1.2 px), beam-centre capture 0.01-0.04 px; beam-stop mask differs on 31 / 144 / 53 pixels of 259k / 144k / 198k (25 of the myoglobin ones are GPU-vs-CPU arithmetic in the mask, the rest follow the centre); hot-pixel mask identical. Spot width, integration radii, bandwidth, beam-centre arbitration, indexing, space group, cell, resolution and the merged statistics table are identical; only the error model moves in its 4th digit. CPU build: the same centres and decisions. Timing (GPU, box at load 30-38): ring walk 0.54 -> 0.23-0.27 s, mask 1.24-1.44 -> 0.18-0.22 s, beam-centre capture walk 1.1-1.3 -> 0.31-0.35 s. Tests: ShadowFinder_DeviceMaskMatchesHost, BeamCenterFromBackground_DeviceMatchesHost (bit-exact), plus [ShadowFinder], [BeamCenter], [HotPixelFinder]. Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01K5K8jvPPbmCrbqnWkddTuB
336 lines
17 KiB
C++
336 lines
17 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>
|
|
|
|
#include <cmath>
|
|
#include <cstdint>
|
|
#include <random>
|
|
#include <vector>
|
|
|
|
#include "../image_analysis/geom_refinement/BeamCenterFromBackground.h"
|
|
#include "../common/DetectorSetup.h"
|
|
#include "../common/JFJochMath.h"
|
|
|
|
namespace {
|
|
|
|
// Solvent and air scatter: a decaying continuum with the water ring on it. The ring is where the
|
|
// leverage comes from - the continuum here is a pure exponential, on which g' is proportional to g
|
|
// and a shift and an amplitude are the same thing - so `ring` is how much there is to fit.
|
|
float background(float two_theta_rad, float ring) {
|
|
const float ring_two_theta = 0.3239f; // ~3.1 A at 1 A
|
|
const float t = (two_theta_rad - ring_two_theta) / 0.035f;
|
|
return 140.0f * std::exp(-two_theta_rad / 0.25f) + ring * std::exp(-0.5f * t * t);
|
|
}
|
|
|
|
// The projection the pre-scan hands over: the mean of a few tens of frames, laid out about
|
|
// geom_true, NAN where the detector has nothing. `shadow_sector` multiplies one sextant, the way a
|
|
// holder arm or a cryostream does.
|
|
std::vector<float> SynthesiseProjection(const DiffractionExperiment &experiment,
|
|
const PixelMask &mask,
|
|
const DiffractionGeometry &geom_true,
|
|
float ring, float shadow_sector) {
|
|
const auto W = static_cast<int>(experiment.GetXPixelsNumConv());
|
|
const auto H = static_cast<int>(experiment.GetYPixelsNumConv());
|
|
const auto &pixel_mask = mask.GetMask(experiment);
|
|
|
|
std::vector<float> mean(static_cast<size_t>(W) * H, NAN);
|
|
std::mt19937 rng(20260812);
|
|
std::normal_distribution<float> gauss(0.0f, 1.0f);
|
|
constexpr float FRAMES = 60.0f; // the mean of this many frames, so the noise is that far down
|
|
|
|
for (int y = 0; y < H; y++) {
|
|
for (int x = 0; x < W; x++) {
|
|
const size_t i = static_cast<size_t>(y) * W + x;
|
|
if (pixel_mask[i] != 0)
|
|
continue;
|
|
float value = background(geom_true.TwoTheta_rad(static_cast<float>(x), static_cast<float>(y)), ring);
|
|
const float phi = geom_true.Phi_rad(static_cast<float>(x), static_cast<float>(y));
|
|
if (phi > 0.0f && phi < static_cast<float>(PI) / 3.0f)
|
|
value *= shadow_sector;
|
|
mean[i] = value + gauss(rng) * std::sqrt(value / FRAMES);
|
|
}
|
|
}
|
|
return mean;
|
|
}
|
|
|
|
DiffractionExperiment TestExperiment() {
|
|
DiffractionExperiment x(DetJF4M());
|
|
x.IncidentEnergy_keV(WVL_1A_IN_KEV).DetectorDistance_mm(100.0f);
|
|
// The band the estimator fits, 12-2.2 A, has to be on the detector, so start from its centre.
|
|
x.BeamX_pxl(static_cast<float>(x.GetXPixelsNumConv()) / 2.0f)
|
|
.BeamY_pxl(static_cast<float>(x.GetYPixelsNumConv()) / 2.0f);
|
|
return x;
|
|
}
|
|
|
|
DiffractionGeometry OffsetBy(const DiffractionGeometry &geom, float dx, float dy) {
|
|
DiffractionGeometry out = geom;
|
|
out.BeamX_pxl(geom.GetBeamX_pxl() + dx).BeamY_pxl(geom.GetBeamY_pxl() + dy);
|
|
return out;
|
|
}
|
|
|
|
} // namespace
|
|
|
|
// The measurement: the background is isotropic in 2-theta about the beam, so a centre that is off
|
|
// shifts each azimuthal sector's radial profile by a different amount, and the shifts give the
|
|
// centre back. Nothing here is indexed, so this is what a de-novo run has to start from - and the
|
|
// estimate has to arrive, because a routine that quietly returns "not measurable" is
|
|
// indistinguishable from a careful refusal in every log line and every merging statistic.
|
|
TEST_CASE("BeamCenterFromBackground_RecoversAnInjectedOffset", "[BeamCenter]") {
|
|
DiffractionExperiment x = TestExperiment();
|
|
PixelMask pixel_mask(x);
|
|
|
|
const DiffractionGeometry geom_true = OffsetBy(x.GetDiffractionGeometry(), 3.0f, -2.5f);
|
|
const auto projection = SynthesiseProjection(x, pixel_mask, geom_true, 60.0f, 1.0f);
|
|
const auto estimate = FindBeamCenterFromBackground(x, pixel_mask, projection);
|
|
|
|
REQUIRE(estimate.has_value());
|
|
CHECK(estimate->beam_x_pxl == Catch::Approx(geom_true.GetBeamX_pxl()).margin(0.5));
|
|
CHECK(estimate->beam_y_pxl == Catch::Approx(geom_true.GetBeamY_pxl()).margin(0.5));
|
|
// And it has to say so precisely enough to be used: the caller commits at 1 px.
|
|
CHECK(estimate->sigma_pxl < 1.0f);
|
|
}
|
|
|
|
// The sigma is the only thing standing between a bad background and a wrong geometry, so it has to
|
|
// grow when the ring it is fitting does not. With the ring at 1.4% of the continuum the centre is
|
|
// still found, but the fit says it is an order of magnitude less sure of it.
|
|
TEST_CASE("BeamCenterFromBackground_SigmaTracksTheLeverage", "[BeamCenter]") {
|
|
DiffractionExperiment x = TestExperiment();
|
|
PixelMask pixel_mask(x);
|
|
|
|
const DiffractionGeometry geom_true = OffsetBy(x.GetDiffractionGeometry(), 3.0f, -2.5f);
|
|
const auto strong = FindBeamCenterFromBackground(
|
|
x, pixel_mask, SynthesiseProjection(x, pixel_mask, geom_true, 60.0f, 1.0f));
|
|
const auto weak = FindBeamCenterFromBackground(
|
|
x, pixel_mask, SynthesiseProjection(x, pixel_mask, geom_true, 2.0f, 1.0f));
|
|
|
|
REQUIRE(strong.has_value());
|
|
REQUIRE(weak.has_value());
|
|
CHECK(weak->beam_x_pxl == Catch::Approx(geom_true.GetBeamX_pxl()).margin(1.0));
|
|
CHECK(weak->beam_y_pxl == Catch::Approx(geom_true.GetBeamY_pxl()).margin(1.0));
|
|
CHECK(weak->sigma_pxl > 5.0f * strong->sigma_pxl);
|
|
}
|
|
|
|
// A holder arm or a cryostream is multiplicative and azimuthal, and a sector that is simply darker
|
|
// looks exactly like a sector whose profile has moved. The per-sector amplitude is what tells them
|
|
// apart: without it half a sextant of shadow reads as tens of pixels of centre error.
|
|
TEST_CASE("BeamCenterFromBackground_AnAzimuthalShadowIsNotACentreError", "[BeamCenter]") {
|
|
DiffractionExperiment x = TestExperiment();
|
|
PixelMask pixel_mask(x);
|
|
|
|
const DiffractionGeometry geom_true = x.GetDiffractionGeometry(); // the centre is already right
|
|
const auto projection = SynthesiseProjection(x, pixel_mask, geom_true, 60.0f, 0.5f);
|
|
const auto estimate = FindBeamCenterFromBackground(x, pixel_mask, projection);
|
|
|
|
REQUIRE(estimate.has_value());
|
|
CHECK(estimate->beam_x_pxl == Catch::Approx(geom_true.GetBeamX_pxl()).margin(1.0));
|
|
CHECK(estimate->beam_y_pxl == Catch::Approx(geom_true.GetBeamY_pxl()).margin(1.0));
|
|
}
|
|
|
|
// The same, with the detector tilted. Every pixel's 2-theta and azimuth, and the derivative of
|
|
// 2-theta with respect to the centre that the fit is built on, go through the detector rotation
|
|
// matrix, so the tilt is not a detail of the geometry here - it is in the Jacobian.
|
|
TEST_CASE("BeamCenterFromBackground_SurvivesADetectorTilt", "[BeamCenter]") {
|
|
DiffractionExperiment x = TestExperiment();
|
|
x.PoniRot1_rad(0.005f).PoniRot2_rad(-0.003f);
|
|
PixelMask pixel_mask(x);
|
|
|
|
const DiffractionGeometry geom_true = OffsetBy(x.GetDiffractionGeometry(), 3.0f, -2.5f);
|
|
const auto [direct_x, direct_y] = geom_true.GetDirectBeam_pxl();
|
|
REQUIRE(std::hypot(direct_x - geom_true.GetBeamX_pxl(), direct_y - geom_true.GetBeamY_pxl()) > 5.0f);
|
|
|
|
const auto projection = SynthesiseProjection(x, pixel_mask, geom_true, 60.0f, 1.0f);
|
|
const auto estimate = FindBeamCenterFromBackground(x, pixel_mask, projection);
|
|
|
|
REQUIRE(estimate.has_value());
|
|
CHECK(estimate->beam_x_pxl == Catch::Approx(geom_true.GetBeamX_pxl()).margin(0.5));
|
|
CHECK(estimate->beam_y_pxl == Catch::Approx(geom_true.GetBeamY_pxl()).margin(0.5));
|
|
CHECK(estimate->sigma_pxl < 1.0f);
|
|
}
|
|
|
|
// A centre hundreds of pixels from the detector's middle - a header that names the middle when the
|
|
// detector was raised - is the case the fit exists for, and it is the case an iteration cap decides.
|
|
// The shift a sector's regression can report is bounded by the width of the features it reads, so
|
|
// the walk advances by a bounded distance per iteration however far it still has to go: the count
|
|
// is a travel budget, and too small a one leaves the fit part way there, still walking, reporting
|
|
// the precision of its last step as though it had arrived.
|
|
TEST_CASE("BeamCenterFromBackground_RecoversACentreFarFromTheDetectorMiddle", "[BeamCenter]") {
|
|
DiffractionExperiment x = TestExperiment();
|
|
PixelMask pixel_mask(x);
|
|
|
|
const DiffractionGeometry geom_true = OffsetBy(x.GetDiffractionGeometry(), 40.0f, 300.0f);
|
|
const auto projection = SynthesiseProjection(x, pixel_mask, geom_true, 60.0f, 1.0f);
|
|
const auto estimate = FindBeamCenterFromBackground(x, pixel_mask, projection);
|
|
|
|
REQUIRE(estimate.has_value());
|
|
CHECK(estimate->beam_x_pxl == Catch::Approx(geom_true.GetBeamX_pxl()).margin(2.0));
|
|
CHECK(estimate->beam_y_pxl == Catch::Approx(geom_true.GetBeamY_pxl()).margin(2.0));
|
|
}
|
|
|
|
// The fit is accumulated over row blocks so that a machine with more cores gives the same answer as
|
|
// one with fewer: the blocks are a property of the image and the fold runs over them in order.
|
|
TEST_CASE("BeamCenterFromBackground_DoesNotDependOnTheThreadCount", "[BeamCenter]") {
|
|
DiffractionExperiment x = TestExperiment();
|
|
PixelMask pixel_mask(x);
|
|
|
|
const DiffractionGeometry geom_true = OffsetBy(x.GetDiffractionGeometry(), 3.0f, -2.5f);
|
|
const auto projection = SynthesiseProjection(x, pixel_mask, geom_true, 60.0f, 1.0f);
|
|
const auto one = FindBeamCenterFromBackground(x, pixel_mask, projection, 1);
|
|
const auto many = FindBeamCenterFromBackground(x, pixel_mask, projection, 8);
|
|
|
|
REQUIRE(one.has_value());
|
|
REQUIRE(many.has_value());
|
|
CHECK(one->beam_x_pxl == many->beam_x_pxl);
|
|
CHECK(one->beam_y_pxl == many->beam_y_pxl);
|
|
CHECK(one->sigma_pxl == many->sigma_pxl);
|
|
}
|
|
|
|
// FindBeamCenter is the walk with an FFT capture in front of it: the capture chooses the basin and
|
|
// the walk finishes inside it and reports what it knows the answer to. The tests below are about
|
|
// the composition - that the walk really is seeded where the capture landed, that a decline falls
|
|
// back the way it is documented to, and that the beam stop is taken out of the image the capture
|
|
// scores.
|
|
|
|
namespace {
|
|
|
|
// A projection that is NAN everywhere except one exactly centrosymmetric patch. The capture has an
|
|
// unambiguous answer on it (a correlation coefficient cannot exceed 1, and the patch reaches 1 at
|
|
// its own centre) while the ring fit has nothing at all: no ring of the 12-2.2 A band is covered
|
|
// by a patch that small, from any start.
|
|
std::vector<float> SymmetricPatchOnly(const DiffractionExperiment &experiment, float cx, float cy,
|
|
int half_size) {
|
|
const auto W = static_cast<int>(experiment.GetXPixelsNumConv());
|
|
const auto H = static_cast<int>(experiment.GetYPixelsNumConv());
|
|
std::vector<float> mean(static_cast<size_t>(W) * H, NAN);
|
|
for (int y = static_cast<int>(cy) - half_size; y <= static_cast<int>(cy) + half_size; y++)
|
|
for (int x = static_cast<int>(cx) - half_size; x <= static_cast<int>(cx) + half_size; x++) {
|
|
// Keyed on the canonical member of the +- pair about (cx, cy), so the patch is exactly
|
|
// its own mirror there.
|
|
int64_t dx = std::lround(2.0f * (static_cast<float>(x) - cx));
|
|
int64_t dy = std::lround(2.0f * (static_cast<float>(y) - cy));
|
|
if (dy < 0 || (dy == 0 && dx < 0)) {
|
|
dx = -dx;
|
|
dy = -dy;
|
|
}
|
|
uint64_t k = static_cast<uint64_t>(dx) * 0x9e3779b97f4a7c15ULL
|
|
+ static_cast<uint64_t>(dy) * 0xc2b2ae3d27d4eb4fULL;
|
|
k ^= k >> 31;
|
|
k *= 0xbf58476d1ce4e5b9ULL;
|
|
k ^= k >> 29;
|
|
mean[static_cast<size_t>(y) * W + x] = 100.0f + static_cast<float>(k % 4096u) * 0.25f;
|
|
}
|
|
return mean;
|
|
}
|
|
|
|
} // namespace
|
|
|
|
// The case the composition exists for. With the centre in the file destroyed - the value a
|
|
// beamline writes when it has nothing to write - the walk has no ring of the fitted band covered
|
|
// from where it starts, so it declines and says nothing at all. The capture does not care how
|
|
// wrong the file is: it scores every centre on the detector in one transform set, and the walk
|
|
// then finishes from there with a real fit sigma, low enough that a caller would commit it.
|
|
TEST_CASE("FindBeamCenter_AnswersWhereTheWalkDeclines", "[BeamCenter]") {
|
|
DiffractionExperiment x = TestExperiment();
|
|
PixelMask pixel_mask(x);
|
|
|
|
const DiffractionGeometry geom_true = x.GetDiffractionGeometry();
|
|
const auto projection = SynthesiseProjection(x, pixel_mask, geom_true, 60.0f, 1.0f);
|
|
x.BeamX_pxl(0.0f).BeamY_pxl(0.0f); // the destroyed header
|
|
|
|
CHECK(!FindBeamCenterFromBackground(x, pixel_mask, projection).has_value());
|
|
|
|
const auto composed = FindBeamCenter(x, pixel_mask, projection);
|
|
REQUIRE(composed.has_value());
|
|
CHECK(composed->beam_x_pxl == Catch::Approx(geom_true.GetBeamX_pxl()).margin(1.0));
|
|
CHECK(composed->beam_y_pxl == Catch::Approx(geom_true.GetBeamY_pxl()).margin(1.0));
|
|
CHECK(composed->sigma_pxl < 1.0f);
|
|
}
|
|
|
|
// The end of the fallback chain. Where the walk has nothing to fit from either start - neither from
|
|
// the capture nor from the centre in the file - the capture stands alone, and it is disclosed as a
|
|
// capture: BEAM_CENTER_CAPTURE_SIGMA_PXL sits above every ceiling a caller commits geometry on, so
|
|
// the answer reaches the consumers that only need a hypothesis to test and is refused by the two
|
|
// that would change the geometry with it.
|
|
TEST_CASE("FindBeamCenter_FallsBackToTheCaptureAlone", "[BeamCenter]") {
|
|
DiffractionExperiment x = TestExperiment();
|
|
PixelMask pixel_mask(x);
|
|
|
|
const float cx = 700.5f, cy = 640.0f;
|
|
const auto projection = SymmetricPatchOnly(x, cx, cy, 90);
|
|
|
|
CHECK(!FindBeamCenterFromBackground(x, pixel_mask, projection).has_value());
|
|
|
|
const auto composed = FindBeamCenter(x, pixel_mask, projection);
|
|
REQUIRE(composed.has_value());
|
|
CHECK(composed->beam_x_pxl == cx);
|
|
CHECK(composed->beam_y_pxl == cy);
|
|
CHECK(composed->sigma_pxl == BEAM_CENTER_CAPTURE_SIGMA_PXL);
|
|
}
|
|
|
|
// The beam stop is blanked out of the image the capture scores. A large one-sided opaque region is
|
|
// centrosymmetric about ITS own centre, and where it is big enough that preference beats the
|
|
// background's - measured on a real umbra over 9.6 % of the detector, the capture landed 48 px out
|
|
// and masking it put it back to 1.1 px. The mask that reaches the capture already has the stop in
|
|
// it, so this costs nothing but has to keep working.
|
|
TEST_CASE("FindBeamCenter_BlanksTheBeamStopOutOfTheCapture", "[BeamCenter]") {
|
|
DiffractionExperiment x = TestExperiment();
|
|
PixelMask pixel_mask(x);
|
|
|
|
const DiffractionGeometry geom_true = x.GetDiffractionGeometry();
|
|
auto projection = SynthesiseProjection(x, pixel_mask, geom_true, 60.0f, 1.0f);
|
|
|
|
// An umbra on one side of the beam: opaque, so the pixels under it carry no background.
|
|
const auto W = static_cast<int>(x.GetXPixelsNumConv());
|
|
const auto H = static_cast<int>(x.GetYPixelsNumConv());
|
|
const float sx = geom_true.GetBeamX_pxl() + 330.0f, sy = geom_true.GetBeamY_pxl() + 70.0f;
|
|
std::vector<uint32_t> stop(static_cast<size_t>(W) * H, 0);
|
|
for (int y = 0; y < H; y++)
|
|
for (int x_pxl = 0; x_pxl < W; x_pxl++)
|
|
if (std::hypot(x_pxl - sx, y - sy) < 300.0f) {
|
|
stop[static_cast<size_t>(y) * W + x_pxl] = 1;
|
|
projection[static_cast<size_t>(y) * W + x_pxl] = 0.0f;
|
|
}
|
|
|
|
const auto unmasked = BeamCenterFFTScore(W, H, projection);
|
|
REQUIRE(!unmasked.point.empty());
|
|
|
|
pixel_mask.LoadBeamStopMask(x, stop);
|
|
BeamCenterFFTResult capture;
|
|
const auto composed = FindBeamCenter(x, pixel_mask, projection, 0, &capture);
|
|
REQUIRE(composed.has_value());
|
|
REQUIRE(!capture.point.empty());
|
|
|
|
const float masked_error = std::hypot(capture.point[0].beam_x_pxl - geom_true.GetBeamX_pxl(),
|
|
capture.point[0].beam_y_pxl - geom_true.GetBeamY_pxl());
|
|
const float unmasked_error = std::hypot(unmasked.point[0].beam_x_pxl - geom_true.GetBeamX_pxl(),
|
|
unmasked.point[0].beam_y_pxl - geom_true.GetBeamY_pxl());
|
|
CHECK(masked_error < 12.0f);
|
|
CHECK(masked_error <= unmasked_error);
|
|
}
|
|
|
|
#ifdef JFJOCH_USE_CUDA
|
|
#include "../common/CUDAWrapper.h"
|
|
|
|
// With a GPU the passes over the pixels run on it. The cell a pixel lands in is computed from IEEE
|
|
// operations alone and without contraction on both sides, and each cell is summed in the host's order,
|
|
// so the walk ends on the same bits - a tilted detector and an offset centre included.
|
|
TEST_CASE("BeamCenterFromBackground_DeviceMatchesHost", "[BeamCenter]") {
|
|
if (get_gpu_count() == 0)
|
|
SKIP("no GPU");
|
|
DiffractionExperiment x = TestExperiment();
|
|
x.PoniRot1_rad(0.005f).PoniRot2_rad(-0.003f);
|
|
PixelMask pixel_mask(x);
|
|
|
|
const DiffractionGeometry geom_true = OffsetBy(x.GetDiffractionGeometry(), 7.0f, -4.5f);
|
|
const auto projection = SynthesiseProjection(x, pixel_mask, geom_true, 60.0f, 0.5f);
|
|
const auto host = FindBeamCenterFromBackground(x, pixel_mask, projection, 0, {}, /*allow_device=*/false);
|
|
const auto device = FindBeamCenterFromBackground(x, pixel_mask, projection, 0, {}, /*allow_device=*/true);
|
|
|
|
REQUIRE(host.has_value());
|
|
REQUIRE(device.has_value());
|
|
CHECK(device->beam_x_pxl == host->beam_x_pxl);
|
|
CHECK(device->beam_y_pxl == host->beam_y_pxl);
|
|
CHECK(device->sigma_pxl == host->sigma_pxl);
|
|
}
|
|
#endif
|