temporary addition of C++ test code

horner_test.cc

+// compile with "g++ -O3 -ffast-math -std=c++17 -march=native -Isrc horner_test.cc"
+
+#include "ducc0/math/horner_kernel.h"
+#include "ducc0/infra/timers.h"
+#include <iostream>
+
+using namespace ducc0;
+using namespace std;
+
+
+// you can use any kernel defined on [-1; 1] here
+double es_kernel(size_t w, double x)
+  {
+  auto beta = 2.3*w;
+  return exp(beta*(sqrt(1-x*x)-1));
+  }
+
+
+int main()
+  {
+  constexpr size_t W=12; // kernel support in pixels
+  constexpr size_t D=12; // degree of approximating polynomials
+  using FLT=double;
+
+  size_t Neval = 1000000000; // we will do a billion function evaluations altogether 
+  size_t Ncall = Neval/W;
+  FLT delta = 2./W/double(Ncall); // small offsets between individual calls to prevent the compiler from optimizing too much
+
+  HornerKernel<W,D,FLT> hk([](double x){return es_kernel(W,x);});
+  double sum=0;
+  SimpleTimer timer;
+  for (size_t i=0; i<Ncall; ++i)
+    {
+    FLT p0 = -FLT(1)+i*delta;  // position of first sample
+    auto res = hk.eval(p0);
+
+    for (size_t i=0; i<W; ++i)
+      sum +=res[i]; // needed to avoid over-optimization
+    }
+  cout << "HornerKernel: " << Neval/timer() << " function approximations per second" << endl;
+  cout << sum << endl;
+
+  HornerKernelFlexible<FLT> hk2(W,D,[](double x){return es_kernel(W,x);});
+  sum=0;
+  timer.reset();
+  for (size_t i=0; i<Ncall; ++i)
+    {
+    FLT p0 = -FLT(1)+i*delta;  // position of first sample
+    auto res = hk2.eval(p0);
+
+    for (size_t i=0; i<W; ++i)
+      sum +=res[i]; // needed to avoid over-optimization
+    }
+  cout << "HornerKernelFlexible: " << Neval/timer() << " function approximations per second" << endl;
+  cout << sum << endl;
+  }