Merge pull request #44360 from lecriste/para-magnetic-field-alpaka_1410

Parametrized magnetic field in alpaka

MagneticField/ParametrizedEngine/interface/ParabolicParametrizedMagneticField.h

@@ -0,0 +1,47 @@
+/**
+ Description: Utility function to calculate the Magnetic Field on the GPU. The Vec3 argument of the functions must support access to its components via (), note that e.g. Eigen::Matrix provides such an interface.
+*/
+
+#ifndef MagneticField_ParametrizedEngine_interface_ParabolicParametrizedMagneticField_h
+#define MagneticField_ParametrizedEngine_interface_ParabolicParametrizedMagneticField_h
+
+namespace magneticFieldParabolicPortable {
+
+  struct Parameters {
+    // These parameters are the best fit of 3.8T to the OAEParametrizedMagneticField parametrization.
+    // See MagneticField/ParametrizedEngine/src/ParabolicParametrizedMagneticField.cc
+    static constexpr float c1 = 3.8114;
+    static constexpr float b0 = -3.94991e-06;
+    static constexpr float b1 = 7.53701e-06;
+    static constexpr float a = 2.43878e-11;
+    static constexpr float max_radius2 = 13225.f;  // tracker radius
+    static constexpr float max_z = 280.f;          // tracker z
+  };
+
+  template <typename Vec3>
+  constexpr float Kr(Vec3 const& vec) {
+    return Parameters::a * (vec(0) * vec(0) + vec(1) * vec(1)) + 1.f;
+  }
+
+  template <typename Vec3>
+  constexpr float B0Z(Vec3 const& vec) {
+    return Parameters::b0 * vec(2) * vec(2) + Parameters::b1 * vec(2) + Parameters::c1;
+  }
+
+  template <typename Vec3>
+  constexpr bool isValid(Vec3 const& vec) {
+    return ((vec(0) * vec(0) + vec(1) * vec(1)) < Parameters::max_radius2 && fabs(vec(2)) < Parameters::max_z);
+  }
+
+  template <typename Vec3>
+  constexpr float magneticFieldAtPoint(Vec3 const& vec) {
+    if (isValid(vec)) {
+      return B0Z(vec) * Kr(vec);
+    } else {
+      return 0;
+    }
+  }
+
+}  // namespace magneticFieldParabolicPortable
+
+#endif

MagneticField/ParametrizedEngine/test/BuildFile.xml

@@ -0,0 +1,9 @@
+<bin file="alpaka/testParabolicParametrizedMagneticField.dev.cc">
+  <use name="alpaka"/>
+  <use name="eigen"/>
+  <use name="DataFormats/GeometryVector"/>
+  <use name="FWCore/Utilities"/>
+  <use name="HeterogeneousCore/AlpakaInterface"/>
+  <use name="MagneticField/ParametrizedEngine" source_only="1"/>
+  <flags ALPAKA_BACKENDS="1"/>
+</bin>

MagneticField/ParametrizedEngine/test/alpaka/testParabolicParametrizedMagneticField.dev.cc

@@ -0,0 +1,131 @@
+#include <iostream>
+#include <fstream>
+#include <Eigen/Core>
+#include <alpaka/alpaka.hpp>
+
+#include "DataFormats/GeometryVector/interface/GlobalPoint.h"
+#include "DataFormats/GeometryVector/interface/GlobalVector.h"
+#include "FWCore/Utilities/interface/FileInPath.h"
+#include "FWCore/Utilities/interface/Exception.h"
+#include "FWCore/Utilities/interface/stringize.h"
+#include "HeterogeneousCore/AlpakaInterface/interface/config.h"
+#include "HeterogeneousCore/AlpakaInterface/interface/memory.h"
+#include "HeterogeneousCore/AlpakaInterface/interface/workdivision.h"
+#include "MagneticField/ParametrizedEngine/interface/ParabolicParametrizedMagneticField.h"
+
+using namespace edm;
+using namespace std;
+using namespace ALPAKA_ACCELERATOR_NAMESPACE;
+using namespace magneticFieldParabolicPortable;
+using Vector3f = Eigen::Matrix<float, 3, 1>;
+
+struct MagneticFieldKernel {
+  template <typename TAcc, typename T>
+  ALPAKA_FN_ACC void operator()(TAcc const& acc, T const* __restrict__ in, T* __restrict__ out, size_t size) const {
+    for (auto index : cms::alpakatools::uniform_elements(acc, size)) {
+      out[index](0) = 0;
+      out[index](1) = 0;
+      out[index](2) = magneticFieldAtPoint(in[index]);
+    }
+  }
+};
+
+int main() {
+  // get the list of devices on the current platform
+  auto const& devices = cms::alpakatools::devices<Platform>();
+  if (devices.empty()) {
+    std::cerr << "No devices available for the " EDM_STRINGIZE(ALPAKA_ACCELERATOR_NAMESPACE) " backend, "
+      "the test will be skipped.\n";
+    exit(EXIT_FAILURE);
+  }
+
+  ifstream file;
+  edm::FileInPath mydata("MagneticField/Engine/data/Regression/referenceField_160812_RII_3_8T.bin");
+  file.open(mydata.fullPath().c_str(), ios::binary);
+
+  int count = 0;
+  float px, py, pz;
+  float bx, by, bz;
+  vector<Vector3f> points;
+  vector<GlobalVector> referenceB_vec;
+
+  int numberOfPoints = 100;
+  points.reserve(numberOfPoints);
+  referenceB_vec.reserve(numberOfPoints);
+  do {
+    if (!(file.read((char*)&px, sizeof(float)) && file.read((char*)&py, sizeof(float)) &&
+          file.read((char*)&pz, sizeof(float)) && file.read((char*)&bx, sizeof(float)) &&
+          file.read((char*)&by, sizeof(float)) && file.read((char*)&bz, sizeof(float))))
+      break;
+
+    const auto point = Vector3f(px, py, pz);
+    if (!isValid(point))
+      continue;
+
+    points.push_back(Vector3f(px, py, pz));
+    referenceB_vec.push_back(GlobalVector(bx, by, bz));
+    count++;
+  } while (count < numberOfPoints);
+
+  const size_t size = points.size();
+  // allocate the input and output host buffer in pinned memory accessible by the Platform devices
+  auto points_host = cms::alpakatools::make_host_buffer<Vector3f[], Platform>(size);
+  auto field_host = cms::alpakatools::make_host_buffer<Vector3f[], Platform>(size);
+  // fill the input buffers, and the output buffer with zeros
+  for (size_t i = 0; i < size; ++i) {
+    points_host[i] = points[i];
+    field_host[i] = Vector3f::Zero();
+  }
+
+  float resolution = 0.2;
+  float maxdelta = 0.;
+  int fail = 0;
+
+  // run the test on each device
+  for (auto const& device : devices) {
+    auto queue = Queue(device);
+    // allocate input and output buffers on the device
+    auto points_dev = cms::alpakatools::make_device_buffer<Vector3f[]>(queue, size);
+    auto field_dev = cms::alpakatools::make_device_buffer<Vector3f[]>(queue, size);
+
+    // copy the input data to the device; the size is known from the buffer objects
+    alpaka::memcpy(queue, points_dev, points_host);
+    // fill the output buffer with zeros; the size is known from the buffer objects
+    alpaka::memset(queue, field_dev, 0.);
+
+    auto workDiv = cms::alpakatools::make_workdiv<Acc1D>(1, size);
+    alpaka::exec<Acc1D>(queue, workDiv, MagneticFieldKernel{}, points_dev.data(), field_dev.data(), size);
+
+    // copy the results from the device to the host
+    alpaka::memcpy(queue, field_host, field_dev);
+
+    // wait for the kernel and the potential copy to complete
+    alpaka::wait(queue);
+
+    // check the results
+    for (uint i = 0; i < points.size(); i++) {
+      const auto& point = points[i];
+      const auto& referenceB = referenceB_vec[i];
+      GlobalVector parametricB(field_host[i](0), field_host[i](1), field_host[i](2));
+      float delta = (referenceB - parametricB).mag();
+      if (delta > resolution) {
+        ++fail;
+        if (delta > maxdelta)
+          maxdelta = delta;
+        if (fail < 10) {
+          const GlobalPoint gp(point(0), point(1), point(2));
+          cout << " Discrepancy at point  # " << i + 1 << ": " << gp << ", R " << gp.perp() << ", Phi " << gp.phi()
+               << ", delta: " << referenceB - parametricB << " " << delta << endl;
+          cout << " Reference: " << referenceB << ", Approximation: " << parametricB << endl;
+        } else if (fail == 10) {
+          cout << "..." << endl;
+        }
+      }
+    }
+
+    if (fail != 0) {
+      cout << "MF regression found: " << fail << " failures; max delta = " << maxdelta << endl;
+      exit(EXIT_FAILURE);
+    }
+  }
+}