Merge pull request #1 from cwuelker/master

dmlux · web-flow · commit 25d4d51cfe41 · 2018-09-03T15:59:29.000+02:00
Small changes
diff --git a/include/PFSOFTlib_headers/fftw_wrapper.hpp b/include/PFSOFTlib_headers/fftw_wrapper.hpp
@@ -28,8 +28,8 @@ PFSOFT_BEGIN
 extern "C"
 {
     /*- FFTW FUNCTIONS -*/
-    void uzl_fftw_layer_wise_DFT2_grid3D (int cols, int rows, int lays, double* arr);
-    void uzl_fftw_layer_wise_IDFT2_grid3D(int cols, int rows, int lays, double* arr);
+    void uzl_fftw_layer_wise_DFT2_grid3D (int cols, int rows, int lays, double* arr, int threads);
+    void uzl_fftw_layer_wise_IDFT2_grid3D(int cols, int rows, int lays, double* arr, int threads);
 }
 
 PFSOFT_END
diff --git a/include/PFSOFTlib_headers/grid3D_cx.hpp b/include/PFSOFTlib_headers/grid3D_cx.hpp
@@ -78,8 +78,8 @@ grid3D< complex< T >, if_pod_type< T > >
     inline       complex< pod_type >&           operator()(const size_t& row, const size_t& col, const size_t& lay);
     inline const complex< pod_type >&           operator()(const size_t& row, const size_t& col, const size_t& lay) const;
     
-    inline       void                           layer_wise_DFT2(const complex< double >& scale = complex< pod_type >(1, 0));
-    inline       void                           layer_wise_IDFT2(const complex< double >& scale = complex< pod_type >(1, 0));
+    inline       void                           layer_wise_DFT2(const complex< double >& scale = complex< pod_type >(1, 0), int threads = 1);
+    inline       void                           layer_wise_IDFT2(const complex< double >& scale = complex< pod_type >(1, 0), int threads = 1);
 };
 
 
@@ -306,7 +306,7 @@ const complex< T >& grid3D< complex< T >, if_pod_type< T > >::operator()(const s
 
 template< typename T >
 inline
-void grid3D< complex< T >, if_pod_type< T > >::layer_wise_DFT2(const complex< double >& scale)
+void grid3D< complex< T >, if_pod_type< T > >::layer_wise_DFT2(const complex< double >& scale, int threads)
 {
     // declare variables
     size_t i;
@@ -336,7 +336,7 @@ void grid3D< complex< T >, if_pod_type< T > >::layer_wise_DFT2(const complex< do
     }
     
     // perform layerwise FFT2
-    uzl_fftw_layer_wise_DFT2_grid3D(cols, rows, lays, data);
+    uzl_fftw_layer_wise_DFT2_grid3D(cols, rows, lays, data, threads);
     
     // skip if scale is default
     if (scale.re != 1 || scale.im != 0)
@@ -367,7 +367,7 @@ void grid3D< complex< T >, if_pod_type< T > >::layer_wise_DFT2(const complex< do
 
 template< typename T >
 inline
-void grid3D< complex< T >, if_pod_type< T > >::layer_wise_IDFT2(const complex< double >& scale)
+void grid3D< complex< T >, if_pod_type< T > >::layer_wise_IDFT2(const complex< double >& scale, int threads)
 {
     // declare variables
     size_t i;
@@ -397,7 +397,7 @@ void grid3D< complex< T >, if_pod_type< T > >::layer_wise_IDFT2(const complex< d
     }
     
     // perform layerwise FFT2
-    uzl_fftw_layer_wise_IDFT2_grid3D(cols, rows, lays, data);
+    uzl_fftw_layer_wise_IDFT2_grid3D(cols, rows, lays, data, threads);
     
     // skip if scale is default
     if (scale.re != 1 || scale.im != 0)
diff --git a/src/fftw_wrapper.cpp b/src/fftw_wrapper.cpp
@@ -44,7 +44,7 @@ PFSOFT_BEGIN
 /*- Wrapper for needed library functions -*/
 extern "C"
 {
-    void uzl_fftw_layer_wise_DFT2_grid3D(int cols, int rows, int lays, double* arr)
+    void uzl_fftw_layer_wise_DFT2_grid3D(int cols, int rows, int lays, double* arr, int threads)
     {
         #ifdef _OPENMP
         
@@ -67,7 +67,7 @@ extern "C"
         }
         
         // execute plans in parallel
-        #pragma omp parallel for private(i) shared(lays, plans) schedule(dynamic)
+        #pragma omp parallel for private(i) shared(lays, plans) schedule(dynamic) num_threads(threads)
         for (i = 0; i < lays; ++i)
         {
             // execute FFT2 plan
@@ -108,7 +108,7 @@ extern "C"
         #endif
     }
     
-    void uzl_fftw_layer_wise_IDFT2_grid3D(int cols, int rows, int lays, double*  arr)
+    void uzl_fftw_layer_wise_IDFT2_grid3D(int cols, int rows, int lays, double* arr, int threads)
     {
         #ifdef _OPENMP
         
@@ -131,7 +131,7 @@ extern "C"
         }
         
         // execute plans in parallel
-        #pragma omp parallel for private(i) shared(lays, plans) schedule(dynamic)
+        #pragma omp parallel for private(i) shared(lays, plans) schedule(dynamic) num_threads(threads)
         for (i = 0; i < lays; ++i)
         {
             // execute FFT2 plan
@@ -173,4 +173,4 @@ extern "C"
     }
 }
 
-PFSOFT_END
+PFSOFT_END
diff --git a/src/fn_dsoft.cpp b/src/fn_dsoft.cpp
@@ -101,7 +101,7 @@ void DSOFT(grid3D< complex< double > > sample, DSOFTFourierCoefficients& fc, int
     /*****************************************************************
      ** FFT2 transform layers of sample grid for fixed k            **
      *****************************************************************/
-    sample.layer_wise_DFT2();
+    sample.layer_wise_DFT2(complex< double > (1.0, 0.0), threads);
     
     /*****************************************************************
      ** M = 0, M' = 0                                               **
diff --git a/src/fn_idsoft.cpp b/src/fn_idsoft.cpp
@@ -286,7 +286,7 @@ void IDSOFT(const DSOFTFourierCoefficients& fc, grid3D< complex< double > >& syn
     /*****************************************************************
      ** IFFT2 transform layers of input sample grid for fixed k     **
      *****************************************************************/
-    synthesis.layer_wise_IDFT2(complex< double > (1. / (4. * bandwidth * bandwidth), 0));
+    synthesis.layer_wise_IDFT2(complex< double > (1. / (4. * bandwidth * bandwidth), 0), threads);
 }
 
 PFSOFT_NAMESPACE_END

Original file line number	Diff line number	Diff line change
`@@ -28,8 +28,8 @@ PFSOFT_BEGIN`
`28`	`28`	`extern "C"`
`29`	`29`	`{`
`30`	`30`	`/- FFTW FUNCTIONS -/`
`31`		`- void uzl_fftw_layer_wise_DFT2_grid3D (int cols, int rows, int lays, double* arr);`
`32`		`- void uzl_fftw_layer_wise_IDFT2_grid3D(int cols, int rows, int lays, double* arr);`
	`31`	`+ void uzl_fftw_layer_wise_DFT2_grid3D (int cols, int rows, int lays, double* arr, int threads);`
	`32`	`+ void uzl_fftw_layer_wise_IDFT2_grid3D(int cols, int rows, int lays, double* arr, int threads);`
`33`	`33`	`}`
`34`	`34`
`35`	`35`	`PFSOFT_END`
Original file line number	Diff line number	Diff line change
`@@ -44,7 +44,7 @@ PFSOFT_BEGIN`
`44`	`44`	`/- Wrapper for needed library functions -/`
`45`	`45`	`extern "C"`
`46`	`46`	`{`
`47`		`- void uzl_fftw_layer_wise_DFT2_grid3D(int cols, int rows, int lays, double* arr)`
	`47`	`+ void uzl_fftw_layer_wise_DFT2_grid3D(int cols, int rows, int lays, double* arr, int threads)`
`48`	`48`	`{`
`49`	`49`	`#ifdef _OPENMP`
`50`	`50`
`@@ -67,7 +67,7 @@ extern "C"`
`67`	`67`	`}`
`68`	`68`
`69`	`69`	`// execute plans in parallel`
`70`		`- #pragma omp parallel for private(i) shared(lays, plans) schedule(dynamic)`
	`70`	`+ #pragma omp parallel for private(i) shared(lays, plans) schedule(dynamic) num_threads(threads)`
`71`	`71`	`for (i = 0; i < lays; ++i)`
`72`	`72`	`{`
`73`	`73`	`// execute FFT2 plan`
`@@ -108,7 +108,7 @@ extern "C"`
`108`	`108`	`#endif`
`109`	`109`	`}`
`110`	`110`
`111`		`- void uzl_fftw_layer_wise_IDFT2_grid3D(int cols, int rows, int lays, double* arr)`
	`111`	`+ void uzl_fftw_layer_wise_IDFT2_grid3D(int cols, int rows, int lays, double* arr, int threads)`
`112`	`112`	`{`
`113`	`113`	`#ifdef _OPENMP`
`114`	`114`
`@@ -131,7 +131,7 @@ extern "C"`
`131`	`131`	`}`
`132`	`132`
`133`	`133`	`// execute plans in parallel`
`134`		`- #pragma omp parallel for private(i) shared(lays, plans) schedule(dynamic)`
	`134`	`+ #pragma omp parallel for private(i) shared(lays, plans) schedule(dynamic) num_threads(threads)`
`135`	`135`	`for (i = 0; i < lays; ++i)`
`136`	`136`	`{`
`137`	`137`	`// execute FFT2 plan`
`@@ -173,4 +173,4 @@ extern "C"`
`173`	`173`	`}`
`174`	`174`	`}`
`175`	`175`
`176`		`-PFSOFT_END`
	`176`	`+PFSOFT_END`
Original file line number	Diff line number	Diff line change
`@@ -286,7 +286,7 @@ void IDSOFT(const DSOFTFourierCoefficients& fc, grid3D< complex< double > >& syn`
`286`	`286`	`/*****************************************************************`
`287`	`287`	` IFFT2 transform layers of input sample grid for fixed k `
`288`	`288`	`*****************************************************************/`
`289`		`- synthesis.layer_wise_IDFT2(complex< double > (1. / (4. * bandwidth * bandwidth), 0));`
	`289`	`+ synthesis.layer_wise_IDFT2(complex< double > (1. / (4. * bandwidth * bandwidth), 0), threads);`
`290`	`290`	`}`
`291`	`291`
`292`	`292`	`PFSOFT_NAMESPACE_END`