426   float *d_a, *d_b, *d_c;  in runStream()  local
434   PetscCallCUDA(cudaMalloc((void **)&d_b, sizeof(float) * N));  in runStream()
445   set_array<<<dimGrid, dimBlock>>>(d_b, .5f, N);  in runStream()
471     STREAM_Scale<<<dimGrid, dimBlock>>>(d_b, d_c, scalar, N);  in runStream()
480     STREAM_Scale_Optimized<<<dimGrid, dimBlock>>>(d_b, d_c, scalar, N);  in runStream()
490     STREAM_Add<<<dimGrid, dimBlock>>>(d_a, d_b, d_c, N);  in runStream()
504     STREAM_Add_Optimized<<<dimGrid, dimBlock>>>(d_a, d_b, d_c, N);  in runStream()
513     STREAM_Triad<<<dimGrid, dimBlock>>>(d_b, d_c, d_a, scalar, N);  in runStream()
522     STREAM_Triad_Optimized<<<dimGrid, dimBlock>>>(d_b, d_c, d_a, scalar, N);  in runStream()
555     set_array<<<dimGrid, dimBlock>>>(d_b, .5f, N);  in runStream()
569     set_array<<<dimGrid, dimBlock>>>(d_b, .5f, N);  in runStream()
583     set_array<<<dimGrid, dimBlock>>>(d_b, .5f, N);  in runStream()
586     STREAM_Scale<<<dimGrid, dimBlock>>>(d_b, d_c, scalar, N);  in runStream()
587     PetscCallCUDA(cudaMemcpy(h_b, d_b, sizeof(float) * N, cudaMemcpyDeviceToHost));  in runStream()
597     set_array<<<dimGrid, dimBlock>>>(d_b, .5f, N);  in runStream()
600     STREAM_Add<<<dimGrid, dimBlock>>>(d_a, d_b, d_c, N);  in runStream()
602     PetscCallCUDA(cudaMemcpy(h_b, d_b, sizeof(float) * N, cudaMemcpyDeviceToHost));  in runStream()
612     set_array<<<dimGrid, dimBlock>>>(d_b, .5f, N);  in runStream()
615     STREAM_Triad<<<dimGrid, dimBlock>>>(d_b, d_c, d_a, scalar, N);  in runStream()
617     PetscCallCUDA(cudaMemcpy(h_b, d_b, sizeof(float) * N, cudaMemcpyDeviceToHost));  in runStream()
634   PetscCallCUDA(cudaFree(d_b));  in runStream()
641   double *d_a, *d_b, *d_c;  in runStreamDouble()  local
649   PetscCallCUDA(cudaMalloc((void **)&d_b, sizeof(double) * N));  in runStreamDouble()
660   set_array_double<<<dimGrid, dimBlock>>>(d_b, .5, N);  in runStreamDouble()
691     STREAM_Scale_double<<<dimGrid, dimBlock>>>(d_b, d_c, scalar, N);  in runStreamDouble()
700     STREAM_Scale_Optimized_double<<<dimGrid, dimBlock>>>(d_b, d_c, scalar, N);  in runStreamDouble()
709     STREAM_Add_double<<<dimGrid, dimBlock>>>(d_a, d_b, d_c, N);  in runStreamDouble()
718     STREAM_Add_Optimized_double<<<dimGrid, dimBlock>>>(d_a, d_b, d_c, N);  in runStreamDouble()
727     STREAM_Triad_double<<<dimGrid, dimBlock>>>(d_b, d_c, d_a, scalar, N);  in runStreamDouble()
736     STREAM_Triad_Optimized_double<<<dimGrid, dimBlock>>>(d_b, d_c, d_a, scalar, N);  in runStreamDouble()
769     set_array_double<<<dimGrid, dimBlock>>>(d_b, .5, N);  in runStreamDouble()
783     set_array_double<<<dimGrid, dimBlock>>>(d_b, .5, N);  in runStreamDouble()
796     set_array_double<<<dimGrid, dimBlock>>>(d_b, .5, N);  in runStreamDouble()
799     STREAM_Scale_double<<<dimGrid, dimBlock>>>(d_b, d_c, scalar, N);  in runStreamDouble()
800     PetscCallCUDA(cudaMemcpy(h_b, d_b, sizeof(double) * N, cudaMemcpyDeviceToHost));  in runStreamDouble()
810     set_array_double<<<dimGrid, dimBlock>>>(d_b, .5, N);  in runStreamDouble()
813     STREAM_Add_double<<<dimGrid, dimBlock>>>(d_a, d_b, d_c, N);  in runStreamDouble()
815     PetscCallCUDA(cudaMemcpy(h_b, d_b, sizeof(double) * N, cudaMemcpyDeviceToHost));  in runStreamDouble()
825     set_array_double<<<dimGrid, dimBlock>>>(d_b, .5, N);  in runStreamDouble()
828     STREAM_Triad_double<<<dimGrid, dimBlock>>>(d_b, d_c, d_a, scalar, N);  in runStreamDouble()
830     PetscCallCUDA(cudaMemcpy(h_b, d_b, sizeof(double) * N, cudaMemcpyDeviceToHost));  in runStreamDouble()
847   PetscCallCUDA(cudaFree(d_b));  in runStreamDouble()