Loading cuda_fft.cu +19 −8 Original line number Diff line number Diff line Loading @@ -52,7 +52,7 @@ void cuda_fft( // Descriptor creation /* cudaLibXtDesc *fftwgrid_g; cudaLibXtDesc *fftwgrid_g2; Loading @@ -64,7 +64,7 @@ void cuda_fft( status = cufftXtMalloc(plan, &fftwgrid_g2, CUFFT_XT_FORMAT_INPLACE); if (status != CUFFT_SUCCESS) {printf("!!! cufftXtMalloc 2 ERROR %d !!!\n", status);} cudaDeviceSynchronize(); */ Loading Loading @@ -92,6 +92,15 @@ void cuda_fft( } cudaLibXtDesc *fftwgrid_g; cudaLibXtDesc *fftwgrid_g2; status = cufftXtMalloc(plan, &fftwgrid_g, CUFFT_XT_FORMAT_INPLACE); if (status != CUFFT_SUCCESS) {printf("!!! cufftXtMalloc ERROR %d !!!\n", status);} status = cufftXtMalloc(plan, &fftwgrid_g2, CUFFT_XT_FORMAT_INPLACE); if (status != CUFFT_SUCCESS) {printf("!!! cufftXtMalloc 2 ERROR %d !!!\n", status);} cudaDeviceSynchronize(); mmm = cudaStreamSynchronize(stream); Loading @@ -102,7 +111,6 @@ void cuda_fft( cudaDeviceSynchronize(); status = cufftXtExecDescriptor(plan, fftwgrid_g, fftwgrid_g, CUFFT_INVERSE); if (status != CUFFT_SUCCESS) {printf("!!! cufftXtExecDescriptor ERROR %d !!!\n", status);} Loading @@ -126,6 +134,9 @@ void cuda_fft( if (status != CUFFT_SUCCESS) {printf("!!! cufftXtMemcpy dth fftwgrid ERROR %d !!!\n", status);} cufftXtFree(fftwgrid_g); cufftXtFree(fftwgrid_g2); for (int iv=0; iv<yaxis; iv++) { for (int iu=0; iu<xaxis; iu++) Loading @@ -139,12 +150,12 @@ void cuda_fft( } status = cufftXtFree(fftwgrid_g); if (status != CUFFT_SUCCESS) {printf("!!! cufftXtFree ERROR %d !!!\n", status);} // status = cufftXtFree(fftwgrid_g); // if (status != CUFFT_SUCCESS) {printf("!!! cufftXtFree ERROR %d !!!\n", status);} status = cufftXtFree(fftwgrid_g2); if (status != CUFFT_SUCCESS) {printf("!!! cufftXtFree ERROR %d !!!\n", status);} // status = cufftXtFree(fftwgrid_g2); // if (status != CUFFT_SUCCESS) {printf("!!! cufftXtFree ERROR %d !!!\n", status);} status = cufftDestroy(plan); Loading Loading
cuda_fft.cu +19 −8 Original line number Diff line number Diff line Loading @@ -52,7 +52,7 @@ void cuda_fft( // Descriptor creation /* cudaLibXtDesc *fftwgrid_g; cudaLibXtDesc *fftwgrid_g2; Loading @@ -64,7 +64,7 @@ void cuda_fft( status = cufftXtMalloc(plan, &fftwgrid_g2, CUFFT_XT_FORMAT_INPLACE); if (status != CUFFT_SUCCESS) {printf("!!! cufftXtMalloc 2 ERROR %d !!!\n", status);} cudaDeviceSynchronize(); */ Loading Loading @@ -92,6 +92,15 @@ void cuda_fft( } cudaLibXtDesc *fftwgrid_g; cudaLibXtDesc *fftwgrid_g2; status = cufftXtMalloc(plan, &fftwgrid_g, CUFFT_XT_FORMAT_INPLACE); if (status != CUFFT_SUCCESS) {printf("!!! cufftXtMalloc ERROR %d !!!\n", status);} status = cufftXtMalloc(plan, &fftwgrid_g2, CUFFT_XT_FORMAT_INPLACE); if (status != CUFFT_SUCCESS) {printf("!!! cufftXtMalloc 2 ERROR %d !!!\n", status);} cudaDeviceSynchronize(); mmm = cudaStreamSynchronize(stream); Loading @@ -102,7 +111,6 @@ void cuda_fft( cudaDeviceSynchronize(); status = cufftXtExecDescriptor(plan, fftwgrid_g, fftwgrid_g, CUFFT_INVERSE); if (status != CUFFT_SUCCESS) {printf("!!! cufftXtExecDescriptor ERROR %d !!!\n", status);} Loading @@ -126,6 +134,9 @@ void cuda_fft( if (status != CUFFT_SUCCESS) {printf("!!! cufftXtMemcpy dth fftwgrid ERROR %d !!!\n", status);} cufftXtFree(fftwgrid_g); cufftXtFree(fftwgrid_g2); for (int iv=0; iv<yaxis; iv++) { for (int iu=0; iu<xaxis; iu++) Loading @@ -139,12 +150,12 @@ void cuda_fft( } status = cufftXtFree(fftwgrid_g); if (status != CUFFT_SUCCESS) {printf("!!! cufftXtFree ERROR %d !!!\n", status);} // status = cufftXtFree(fftwgrid_g); // if (status != CUFFT_SUCCESS) {printf("!!! cufftXtFree ERROR %d !!!\n", status);} status = cufftXtFree(fftwgrid_g2); if (status != CUFFT_SUCCESS) {printf("!!! cufftXtFree ERROR %d !!!\n", status);} // status = cufftXtFree(fftwgrid_g2); // if (status != CUFFT_SUCCESS) {printf("!!! cufftXtFree ERROR %d !!!\n", status);} status = cufftDestroy(plan); Loading