Add an example for 2D type 3 on GPU

examples/cuda/example2d3many.cpp

@@ -0,0 +1,131 @@
+/* This is an example of performing 2d3many
+   in double precision.
+*/
+
+#include <complex>
+#include <iomanip>
+#include <iostream>
+#include <math.h>
+#include <random>
+
+#include <cufinufft.h>
+#include <cufinufft/utils.h>
+
+#include <cuda_runtime.h>
+
+// FIXME: This isn't actually public, though maybe it should be?
+using cufinufft::utils::infnorm;
+
+int main(int argc, char *argv[])
+/*
+ * example code for 2D Type 3 transformation.
+ *
+ * To compile the code:
+ * nvcc example2d3many.cpp -o example2d3many loc/to/cufinufft/lib-static/libcufinufft.a
+ * -I/loc/to/cufinufft/include -lcudart -lcufft -lnvToolsExt
+ *
+ * or
+ * export LD_LIBRARY_PATH=$LD_LIBRARY_PATH:~/loc/to/cufinufft/lib
+ * nvcc example2d3many.cpp -example2d3many -L/loc/to/cufinufft/lib/
+ * -I/loc/to/cufinufft/include -lcufinufft
+ *
+ *
+ */
+{
+  std::cout << std::scientific << std::setprecision(3);
+
+  int ier;
+  int M            = 10;
+  int N            = 20;
+  int ntransf      = 4;
+  int maxbatchsize = 4;
+  int iflag        = 1;
+  double tol       = 1e-6;
+
+  double *x, *y, *s, *t;
+  std::complex<double> *c, *fk;
+  cudaMallocHost(&x, M * sizeof(double));
+  cudaMallocHost(&y, M * sizeof(double));
+  cudaMallocHost(&s, N * sizeof(double));
+  cudaMallocHost(&t, N * sizeof(double));
+  cudaMallocHost(&c, M * ntransf * sizeof(std::complex<double>));
+  cudaMallocHost(&fk, N * ntransf * sizeof(std::complex<double>));
+
+  double *d_x, *d_y, *d_s, *d_t;
+  cuDoubleComplex *d_c, *d_fk;
+  cudaMalloc(&d_x, M * sizeof(double));
+  cudaMalloc(&d_y, M * sizeof(double));
+  cudaMalloc(&d_c, M * ntransf * sizeof(cuDoubleComplex));
+  cudaMalloc(&d_s, N * sizeof(double));
+  cudaMalloc(&d_t, N * sizeof(double));
+  cudaMalloc(&d_fk, N * ntransf * sizeof(cuDoubleComplex));
+
+  std::default_random_engine eng(1);
+  std::uniform_real_distribution<double> distr(-1, 1);
+
+  for (int i = 0; i < M; i++) {
+    x[i] = M_PI * distr(eng);
+    y[i] = M_PI * distr(eng);
+  }
+
+  for (int i = 0; i < N; i++) {
+    s[i] = distr(eng);
+    t[i] = distr(eng);
+  }
+
+  for (int i = 0; i < M * ntransf; i++) {
+    c[i].real(distr(eng));
+    c[i].imag(distr(eng));
+  }
+
+  cudaMemcpy(d_x, x, M * sizeof(double), cudaMemcpyHostToDevice);
+  cudaMemcpy(d_y, y, M * sizeof(double), cudaMemcpyHostToDevice);
+  cudaMemcpy(d_s, s, N * sizeof(double), cudaMemcpyHostToDevice);
+  cudaMemcpy(d_t, t, N * sizeof(double), cudaMemcpyHostToDevice);
+  cudaMemcpy(d_c, c, M * ntransf * sizeof(cuDoubleComplex), cudaMemcpyHostToDevice);
+
+  cufinufft_plan dplan;
+
+  int dim           = 2;
+  int64_t nmodes[3] = {N, 1, 1};
+  int type          = 3;
+
+  cufinufft_opts opts;
+  cufinufft_default_opts(&opts);
+
+  ier = cufinufft_makeplan(type, dim, nmodes, iflag, ntransf, tol, &dplan, &opts);
+
+  ier = cufinufft_setpts(dplan, M, d_x, d_y, NULL, N, d_s, d_t, NULL);
+
+  ier = cufinufft_execute(dplan, d_c, d_fk);
+
+  ier = cufinufft_destroy(dplan);
+
+  cudaMemcpy(fk, d_fk, N * ntransf * sizeof(cuDoubleComplex), cudaMemcpyDeviceToHost);
+
+  std::cout << std::endl << "Accuracy check:" << std::endl;
+  std::complex<double> *fkstart;
+  std::complex<double> *cstart;
+  for (int tr = 0; tr < ntransf; tr++) {
+    fkstart = fk + tr * N;
+    cstart  = c + tr * M;
+    int jt  = N / 2; // check arbitrary choice of one targ pt
+    std::complex<double> J(0, iflag * 1);
+    std::complex<double> fkt(0, 0);
+    for (int m = 0; m < M; m++) fkt += cstart[m] * exp(J * (s[jt] * x[m] + t[jt] * y[m]));
+
+    printf("[gpu %3d] one targ: rel err in c[%d] is %.3g\n", tr, jt,
+           abs(fkstart[jt] - fkt) / infnorm(N, fkstart));
+  }
+
+  cudaFreeHost(x);
+  cudaFreeHost(y);
+  cudaFreeHost(c);
+  cudaFreeHost(fk);
+
+  cudaFree(d_x);
+  cudaFree(d_y);
+  cudaFree(d_c);
+  cudaFree(d_fk);
+  return 0;
+}