more OpenMP

nifty_gridder.cc

@@ -308,6 +308,7 @@ template<typename T> pyarr_c<T> complex2hartley
 
   auto res = makearray<T>({nu,nv});
   auto grid2 = res.mutable_data();
+#pragma omp parallel for
   for (size_t u=0; u<nu; ++u)
     {
     size_t xu = (u==0) ? 0 : nu-u;
@@ -332,6 +333,7 @@ template<typename T> pyarr_c<complex<T>> hartley2complex
 
   auto res=makearray<complex<T>>({nu, nv});
   auto grid2 = res.mutable_data();
+#pragma omp parallel for
   for (size_t u=0; u<nu; ++u)
     {
     size_t xu = (u==0) ? 0 : nu-u;
@@ -356,6 +358,7 @@ template<typename T> void hartley2_2D(const pyarr_c<T> &in, pyarr_c<T> &out)
     {out.strides(0), out.strides(1)}, {0,1},
     in.data(), out.mutable_data(), T(1), 0);
   auto ptmp = out.mutable_data();
+#pragma omp parallel for
   for(size_t i=1; i<(nu+1)/2; ++i)
     for(size_t j=1; j<(nv+1)/2; ++j)
        {
@@ -384,6 +387,7 @@ vector<double> correction_factors (size_t n, size_t nval, size_t w)
   for (auto &v:psi)
     v = exp(beta*(sqrt(1-v*v)-1.));
   vector<double> res(nval);
+#pragma omp parallel for schedule(static)
   for (size_t k=0; k<nval; ++k)
     {
     double tmp=0;
@@ -490,6 +494,7 @@ template<typename T> class Baselines
 
       auto res=makearray<T2>({nvis});
       auto vis = res.mutable_data();
+#pragma omp parallel for
       for (size_t i=0; i<nvis; ++i)
         vis[i] = ms[offset(idx[i])];
       return res;
@@ -509,6 +514,7 @@ template<typename T> class Baselines
       auto ms = res.mutable_data();
       for (size_t i=0; i<nrows*nchan; ++i)
         ms[i] = T2(0);
+#pragma omp parallel for
       for (size_t i=0; i<nvis; ++i)
         ms[offset(idx[i])] = vis[i];
       return res;
@@ -528,6 +534,7 @@ template<typename T> class Baselines
       auto vis = vis_.data();
       auto ms = ms_.mutable_data();
 
+#pragma omp parallel for
       for (size_t i=0; i<nvis; ++i)
         ms[offset(idx[i])] += vis[i];
       return ms_;