De-macroize FFT

include/finufft/defs.h

@@ -18,6 +18,7 @@
 // public header gives access to f_opts, f_spread_opts, f_plan...
 // (and clobbers FINUFFT* macros; watch out!)
 #include <finufft.h>
+#include <memory>
 
 // --------------- Private data types for compilation in either prec ---------
 // Devnote: must match those in relevant prec of public finufft.h interface!
@@ -89,7 +90,7 @@
 inline constexpr BIGINT MAX_NU_PTS = BIGINT(1e14);
 
 // -------------- Math consts (not in math.h) and useful math macros ----------
-#include <math.h>
+#include <cmath>
 
 // either-precision unit imaginary number...
 #define IMA (CPX(0.0, 1.0))
@@ -118,11 +119,11 @@
 // These macros should probably be replaced by modern C++ std lib or random123.
 // (RAND_MAX is in stdlib.h)
 #include <stdlib.h>
-static inline FLT rand01() { return FLT(rand()) / FLT(RAND_MAX); }
+static inline FLT rand01 [[maybe_unused]] () { return FLT(rand()) / FLT(RAND_MAX); }
 // unif[-1,1]:
-static inline FLT randm11() { return 2 * rand01() - FLT(1); }
+static inline FLT randm11 [[maybe_unused]] () { return 2 * rand01() - FLT(1); }
 // complex unif[-1,1] for Re and Im:
-static inline CPX crandm11() { return randm11() + IMA * randm11(); }
+static inline CPX crandm11 [[maybe_unused]] () { return randm11() + IMA * randm11(); }
 
 // Thread-safe seed-carrying versions of above (x is ptr to seed)...
 // MR: we have to leave those as macros for now, as "rand_r" is deprecated
@@ -134,11 +135,17 @@
 // complex unif[-1,1] for Re and Im:
 #define crandm11r(x) (randm11r(x) + IMA * randm11r(x))
 #else
-static inline FLT rand01r(unsigned int *x) { return FLT(rand_r(x)) / FLT(RAND_MAX); }
+static inline FLT rand01r [[maybe_unused]] (unsigned int *x) {
+  return FLT(rand_r(x)) / FLT(RAND_MAX);
+}
 // unif[-1,1]:
-static inline FLT randm11r(unsigned int *x) { return 2 * rand01r(x) - FLT(1); }
+static inline FLT randm11r [[maybe_unused]] (unsigned int *x) {
+  return 2 * rand01r(x) - FLT(1);
+}
 // complex unif[-1,1] for Re and Im:
-static inline CPX crandm11r(unsigned int *x) { return randm11r(x) + IMA * randm11r(x); }
+static inline CPX crandm11r [[maybe_unused]] (unsigned int *x) {
+  return randm11r(x) + IMA * randm11r(x);
+}
 #endif
 
 // ----- OpenMP macros which also work when omp not present -----
@@ -147,16 +154,24 @@
 #ifdef _OPENMP
 #include <omp.h>
 // point to actual omp utils
-static inline int MY_OMP_GET_NUM_THREADS() { return omp_get_num_threads(); }
-static inline int MY_OMP_GET_MAX_THREADS() { return omp_get_max_threads(); }
-static inline int MY_OMP_GET_THREAD_NUM() { return omp_get_thread_num(); }
-static inline void MY_OMP_SET_NUM_THREADS(int x) { omp_set_num_threads(x); }
+static inline int MY_OMP_GET_NUM_THREADS [[maybe_unused]] () {
+  return omp_get_num_threads();
+}
+static inline int MY_OMP_GET_MAX_THREADS [[maybe_unused]] () {
+  return omp_get_max_threads();
+}
+static inline int MY_OMP_GET_THREAD_NUM [[maybe_unused]] () {
+  return omp_get_thread_num();
+}
+static inline void MY_OMP_SET_NUM_THREADS [[maybe_unused]] (int x) {
+  omp_set_num_threads(x);
+}
 #else
 // non-omp safe dummy versions of omp utils...
-static inline int MY_OMP_GET_NUM_THREADS() { return 1; }
-static inline int MY_OMP_GET_MAX_THREADS() { return 1; }
-static inline int MY_OMP_GET_THREAD_NUM() { return 0; }
-static inline void MY_OMP_SET_NUM_THREADS(int) {}
+static inline int MY_OMP_GET_NUM_THREADS [[maybe_unused]] () { return 1; }
+static inline int MY_OMP_GET_MAX_THREADS [[maybe_unused]] () { return 1; }
+static inline int MY_OMP_GET_THREAD_NUM [[maybe_unused]] () { return 0; }
+static inline void MY_OMP_SET_NUM_THREADS [[maybe_unused]] (int) {}
 #endif
 
 // Prec-switching name macros (respond to SINGLE), used in lib & test sources
@@ -212,16 +227,23 @@
   T X3, C3, D3, h3, gam3; // z
 };
 
-typedef struct FINUFFT_PLAN_S { // the main plan object, fully C++
-
-  int type;                     // transform type (Rokhlin naming): 1,2 or 3
-  int dim;                      // overall dimension: 1,2 or 3
+struct FINUFFT_PLAN_S { // the main plan object, fully C++
+  // These default and delete specifications just state the obvious,
+  // but are here to silence compiler warnings.
+  FINUFFT_PLAN_S() = default;
+  // Copy construction and assignent are already deleted implicitly
+  // because of the unique_ptr member.
+  FINUFFT_PLAN_S(const FINUFFT_PLAN_S &)            = delete;
+  FINUFFT_PLAN_S &operator=(const FINUFFT_PLAN_S &) = delete;
+
+  int type;            // transform type (Rokhlin naming): 1,2 or 3
+  int dim;             // overall dimension: 1,2 or 3
   int ntrans;          // how many transforms to do at once (vector or "many" mode)
   BIGINT nj;           // num of NU pts in type 1,2 (for type 3, num input x pts)
   BIGINT nk;           // number of NU freq pts (type 3 only)
   FLT tol;             // relative user tolerance
   int batchSize;       // # strength vectors to group together for FFTW, etc
   int nbatch;          // how many batches done to cover all ntrans vectors
 
   BIGINT ms;           // number of modes in x (1) dir (historical CMCL name) = N1
   BIGINT mt;           // number of modes in y (2) direction = N2
@@ -233,7 +255,7 @@
   BIGINT nf3;          // " z (3)
   BIGINT nf;           // total # fine grid points (product of the above three)
 
   int fftSign;         // sign in exponential for NUFFT defn, guaranteed to be +-1
 
   FLT *phiHat1;        // FT of kernel in t1,2, on x-axis mode grid
   FLT *phiHat2;        // " y-axis.
@@ -243,7 +265,7 @@
                        // & act on. Usually the largest working array
 
   BIGINT *sortIndices; // precomputed NU pt permutation, speeds spread/interp
   bool didSort;        // whether binsorting used (false: identity perm used)
 
   FLT *X, *Y, *Z;      // for t1,2: ptr to user-supplied NU pts (no new allocs).
                        // for t3: allocated as "primed" (scaled) src pts x'_j, etc
@@ -254,16 +276,13 @@
   CPX *deconv;              // reciprocal of kernel FT, phase, all output NU pts
   CPX *CpBatch;             // working array of prephased strengths
   FLT *Sp, *Tp, *Up;        // internal primed targs (s'_k, etc), allocated
   type3params<FLT> t3P;     // groups together type 3 shift, scale, phase, parameters
   FINUFFT_PLAN innerT2plan; // ptr used for type 2 in step 2 of type 3
 
   // other internal structs; each is C-compatible of course
-#ifndef FINUFFT_USE_DUCC0
-  FFTW_PLAN fftwPlan;
-#endif
+  std::unique_ptr<Finufft_FFT_plan<FLT>> fftPlan;
   finufft_opts opts; // this and spopts could be made ptrs
   finufft_spread_opts spopts;
-
-} FINUFFT_PLAN_S;
+};
 
 #endif // DEFS_H

include/finufft/fft.h

@@ -1,18 +1,189 @@
 #ifndef FINUFFT_INCLUDE_FINUFFT_FFT_H
 #define FINUFFT_INCLUDE_FINUFFT_FFT_H
 
+#include <vector>
+
 #ifdef FINUFFT_USE_DUCC0
-#include "ducc0/fft/fftnd_impl.h"
-// temporary hacks to allow compilation of tests that assume FFTW is used
-static inline void FFTW_FORGET_WISDOM() {}
-static inline void FFTW_CLEANUP() {}
-static inline void FFTW_CLEANUP_THREADS() {}
+#include <complex>
+
+template<typename T> class Finufft_FFT_plan {
+public:
+  [[maybe_unused]] Finufft_FFT_plan(void (*)(void *) = nullptr,
+                                    void (*)(void *) = nullptr, void * = nullptr) {}
+  [[maybe_unused]] void plan(const std::vector<int> & /*dims*/, size_t /*batchSize*/,
+                             std::complex<T> * /*ptr*/, int /*sign*/, int /*options*/,
+                             int /*nthreads*/) {}
+  [[maybe_unused]] static std::complex<T> *alloc_complex(size_t N) {
+    return new std::complex<T>[N];
+  }
+  [[maybe_unused]] static void free(std::complex<T> *ptr) { delete[] ptr; }
+
+  [[maybe_unused]] static void forget_wisdom() {}
+  [[maybe_unused]] static void cleanup() {}
+  [[maybe_unused]] static void cleanup_threads() {}
+};
+
 #else
-#include "fftw_defs.h"
+
+//clang-format off
+#include <complex>
+#include <fftw3.h> // (after complex) needed so can typedef FFTW_CPX
+//clang-format on
+#include <mutex>
+
+template<typename T> class Finufft_FFT_plan {};
+
+template<> struct Finufft_FFT_plan<float> {
+private:
+  static std::mutex &mut() {
+    static std::mutex mut_;
+    return mut_;
+  }
+  fftwf_plan plan_;
+
+  void (*fftw_lock_fun)(void *);   // Function ptr that locks the FFTW planner
+  void (*fftw_unlock_fun)(void *); // Function ptr that unlocks the FFTW planner
+  void *lock_data;
+  void lock() { fftw_lock_fun ? fftw_lock_fun(lock_data) : mut().lock(); }
+  void unlock() { fftw_lock_fun ? fftw_unlock_fun(lock_data) : mut().unlock(); }
+
+public:
+  [[maybe_unused]] Finufft_FFT_plan(void (*fftw_lock_fun_)(void *)   = nullptr,
+                                    void (*fftw_unlock_fun_)(void *) = nullptr,
+                                    void *lock_data_                 = nullptr)
+      : plan_(nullptr), fftw_lock_fun(fftw_lock_fun_), fftw_unlock_fun(fftw_unlock_fun_),
+        lock_data(lock_data_) {
+    lock();
+#ifdef _OPENMP
+    static bool initialized = false;
+    if (!initialized) {
+      fftwf_init_threads();
+      initialized = true;
+    }
+#endif
+    unlock();
+  }
+  [[maybe_unused]] ~Finufft_FFT_plan() {
+    lock();
+    fftwf_destroy_plan(plan_);
+    unlock();
+  }
+
+  void plan
+      [[maybe_unused]] (const std::vector<int> &dims, size_t batchSize,
+                        std::complex<float> *ptr, int sign, int options, int nthreads) {
+    uint64_t nf = 1;
+    for (auto i : dims) nf *= i;
+    lock();
+#ifdef _OPENMP
+    fftwf_plan_with_nthreads(nthreads);
+#endif
+    plan_ = fftwf_plan_many_dft(int(dims.size()), dims.data(), int(batchSize),
+                                reinterpret_cast<fftwf_complex *>(ptr), nullptr, 1,
+                                int(nf), reinterpret_cast<fftwf_complex *>(ptr), nullptr,
+                                1, int(nf), sign, unsigned(options));
+    unlock();
+  }
+  static std::complex<float> *alloc_complex [[maybe_unused]] (size_t N) {
+    return reinterpret_cast<std::complex<float> *>(fftwf_alloc_complex(N));
+  }
+  static void free [[maybe_unused]] (std::complex<float> *ptr) {
+    if (ptr) fftwf_free(reinterpret_cast<fftwf_complex *>(ptr));
+  }
+  void execute [[maybe_unused]] () { fftwf_execute(plan_); }
+
+  static void forget_wisdom [[maybe_unused]] () { fftwf_forget_wisdom(); }
+  static void cleanup [[maybe_unused]] () { fftwf_cleanup(); }
+  static void cleanup_threads [[maybe_unused]] () {
+#ifdef _OPENMP
+    fftwf_cleanup_threads();
 #endif
+  }
+};
+
+template<> struct Finufft_FFT_plan<double> {
+private:
+  static std::mutex &mut() {
+    static std::mutex mut_;
+    return mut_;
+  }
+  fftw_plan plan_;
+
+  void (*fftw_lock_fun)(void *);   // Function ptr that locks the FFTW planner
+  void (*fftw_unlock_fun)(void *); // Function ptr that unlocks the FFTW planner
+  void *lock_data;
+  void lock() { fftw_lock_fun ? fftw_lock_fun(lock_data) : mut().lock(); }
+  void unlock() { fftw_lock_fun ? fftw_unlock_fun(lock_data) : mut().unlock(); }
+
+public:
+  [[maybe_unused]] Finufft_FFT_plan(void (*fftw_lock_fun_)(void *)   = nullptr,
+                                    void (*fftw_unlock_fun_)(void *) = nullptr,
+                                    void *lock_data_                 = nullptr)
+      : plan_(nullptr), fftw_lock_fun(fftw_lock_fun_), fftw_unlock_fun(fftw_unlock_fun_),
+        lock_data(lock_data_) {
+    lock();
+#ifdef _OPENMP
+    static bool initialized = false;
+    if (!initialized) {
+      fftw_init_threads();
+      initialized = true;
+    }
+#endif
+    unlock();
+  }
+  [[maybe_unused]] ~Finufft_FFT_plan() {
+    lock();
+    fftw_destroy_plan(plan_);
+    unlock();
+  }
+
+  void plan
+      [[maybe_unused]] (const std::vector<int> &dims, size_t batchSize,
+                        std::complex<double> *ptr, int sign, int options, int nthreads) {
+    uint64_t nf = 1;
+    for (auto i : dims) nf *= i;
+    lock();
+#ifdef _OPENMP
+    fftw_plan_with_nthreads(nthreads);
+#endif
+    plan_ = fftw_plan_many_dft(int(dims.size()), dims.data(), int(batchSize),
+                               reinterpret_cast<fftw_complex *>(ptr), nullptr, 1, int(nf),
+                               reinterpret_cast<fftw_complex *>(ptr), nullptr, 1, int(nf),
+                               sign, unsigned(options));
+    unlock();
+  }
+  static std::complex<double> *alloc_complex [[maybe_unused]] (size_t N) {
+    return reinterpret_cast<std::complex<double> *>(fftw_alloc_complex(N));
+  }
+  static void free [[maybe_unused]] (std::complex<double> *ptr) {
+    fftw_free(reinterpret_cast<fftw_complex *>(ptr));
+  }
+  void execute [[maybe_unused]] () { fftw_execute(plan_); }
+
+  static void forget_wisdom [[maybe_unused]] () { fftw_forget_wisdom(); }
+  static void cleanup [[maybe_unused]] () { fftw_cleanup(); }
+  static void cleanup_threads [[maybe_unused]] () {
+#ifdef _OPENMP
+    fftw_cleanup_threads();
+#endif
+  }
+};
+
+#endif
+
 #include <finufft/defs.h>
 
-int *gridsize_for_fft(FINUFFT_PLAN p);
+static inline void finufft_fft_forget_wisdom [[maybe_unused]] () {
+  Finufft_FFT_plan<FLT>::forget_wisdom();
+}
+static inline void finufft_fft_cleanup [[maybe_unused]] () {
+  Finufft_FFT_plan<FLT>::cleanup();
+}
+static inline void finufft_fft_cleanup_threads [[maybe_unused]] () {
+  Finufft_FFT_plan<FLT>::cleanup_threads();
+}
+
+std::vector<int> gridsize_for_fft(FINUFFT_PLAN p);
 void do_fft(FINUFFT_PLAN p);
 
 #endif // FINUFFT_INCLUDE_FINUFFT_FFT_H

perftest/guru_timing_test.cpp

@@ -145,9 +145,9 @@ int main(int argc, char *argv[])
   }
 
   // Andrea found the following are needed to get reliable independent timings:
-  FFTW_CLEANUP();
-  FFTW_CLEANUP_THREADS();
-  FFTW_FORGET_WISDOM();
+  finufft_fft_cleanup();
+  finufft_fft_cleanup_threads();
+  finufft_fft_forget_wisdom();
   // std::this_thread::sleep_for(std::chrono::seconds(1));
   sleep(tsleep);