Loading Makefile +1 −1 Original line number Diff line number Diff line Loading @@ -92,7 +92,7 @@ OPT += -DGAUSS_HI_PRECISION #OPT += -DCUFFTMP # FULL NVIDIA GPU SUPPORT - Recommended for full NVIDIA GPU code execution #OPT += -DFULL_NVIDIA OPT += -DFULL_NVIDIA ifeq (FULL_NVIDIA,$(findstring FULL_NVIDIA,$(OPT))) OPT += -DCUDACC -DNCCL_REDUCE -DCUFFTMP endif Loading gridding_cpu.c +7 −1 Original line number Diff line number Diff line Loading @@ -181,6 +181,11 @@ void gridding_data() else stacking_target_array = grid; //for(int iii=0; iii<9000000; iii++)if(stacking_target_array[iii]!=0.0)printf("%f\n",stacking_target_array[iii]); //printf(">>>>>>>>>>>>>>>>>>>>>>>>\n"); //int ppp; //scanf("%d",&ppp); start = CPU_TIME_wt; //We have to call different GPUs per MPI task!!! [GL] Loading Loading @@ -218,6 +223,8 @@ void gridding_data() start = CPU_TIME_wt; if (isector == rank) for (long isum = 0; isum < size_of_grid; isum++) gridss[isum] += grid[isum]; //Force to use MPI_Reduce when -fopenmp is not active #ifdef _OPENMP Loading Loading @@ -246,7 +253,6 @@ void gridding_data() #endif timing_wt.reduce += CPU_TIME_wt - start; printf("=========================\n"); // Go to next sector memset ( gridss, 0, 2*param.num_w_planes*xaxis*yaxis * sizeof(double) ); Loading init.c +5 −0 Original line number Diff line number Diff line Loading @@ -92,6 +92,10 @@ void op_filename() { strcat(buf, outparam.fftfile_writedata2); strcpy(out.fftfile_writedata2, buf); #endif // CLAAAAAAAAAA strcpy(out.fftfile2,outparam.fftfile2); strcpy(out.fftfile3,outparam.fftfile3); /* strcpy(buf, num_buf); strcat(buf, outparam.fftfile2); strcpy(out.fftfile2, buf); Loading @@ -99,6 +103,7 @@ void op_filename() { strcpy(buf, num_buf); strcat(buf, outparam.fftfile3); strcpy(out.fftfile3, buf); */ strcpy(buf, num_buf); strcat(buf, outparam.logfile); Loading Loading
Makefile +1 −1 Original line number Diff line number Diff line Loading @@ -92,7 +92,7 @@ OPT += -DGAUSS_HI_PRECISION #OPT += -DCUFFTMP # FULL NVIDIA GPU SUPPORT - Recommended for full NVIDIA GPU code execution #OPT += -DFULL_NVIDIA OPT += -DFULL_NVIDIA ifeq (FULL_NVIDIA,$(findstring FULL_NVIDIA,$(OPT))) OPT += -DCUDACC -DNCCL_REDUCE -DCUFFTMP endif Loading
gridding_cpu.c +7 −1 Original line number Diff line number Diff line Loading @@ -181,6 +181,11 @@ void gridding_data() else stacking_target_array = grid; //for(int iii=0; iii<9000000; iii++)if(stacking_target_array[iii]!=0.0)printf("%f\n",stacking_target_array[iii]); //printf(">>>>>>>>>>>>>>>>>>>>>>>>\n"); //int ppp; //scanf("%d",&ppp); start = CPU_TIME_wt; //We have to call different GPUs per MPI task!!! [GL] Loading Loading @@ -218,6 +223,8 @@ void gridding_data() start = CPU_TIME_wt; if (isector == rank) for (long isum = 0; isum < size_of_grid; isum++) gridss[isum] += grid[isum]; //Force to use MPI_Reduce when -fopenmp is not active #ifdef _OPENMP Loading Loading @@ -246,7 +253,6 @@ void gridding_data() #endif timing_wt.reduce += CPU_TIME_wt - start; printf("=========================\n"); // Go to next sector memset ( gridss, 0, 2*param.num_w_planes*xaxis*yaxis * sizeof(double) ); Loading
init.c +5 −0 Original line number Diff line number Diff line Loading @@ -92,6 +92,10 @@ void op_filename() { strcat(buf, outparam.fftfile_writedata2); strcpy(out.fftfile_writedata2, buf); #endif // CLAAAAAAAAAA strcpy(out.fftfile2,outparam.fftfile2); strcpy(out.fftfile3,outparam.fftfile3); /* strcpy(buf, num_buf); strcat(buf, outparam.fftfile2); strcpy(out.fftfile2, buf); Loading @@ -99,6 +103,7 @@ void op_filename() { strcpy(buf, num_buf); strcat(buf, outparam.fftfile3); strcpy(out.fftfile3, buf); */ strcpy(buf, num_buf); strcat(buf, outparam.logfile); Loading