int main(int argc, char* argv[])
//int main(void)
{
double t_start, t_end;
DATA_TYPE* A;
DATA_TYPE* B;
DATA_TYPE* C;
DATA_TYPE* D;
DATA_TYPE* E;
DATA_TYPE* F;
DATA_TYPE* G;
DATA_TYPE* G_outputFromGpu;
if(argc==2){
printf("arg 1 = %s\narg 2 = %s\n", argv[0], argv[1]);
cpu_offset = atoi(argv[1]);
}
A = (DATA_TYPE*)malloc(NI*NK*sizeof(DATA_TYPE));
B = (DATA_TYPE*)malloc(NK*NJ*sizeof(DATA_TYPE));
C = (DATA_TYPE*)malloc(NJ*NM*sizeof(DATA_TYPE));
D = (DATA_TYPE*)malloc(NM*NL*sizeof(DATA_TYPE));
E = (DATA_TYPE*)malloc(NI*NJ*sizeof(DATA_TYPE));
F = (DATA_TYPE*)malloc(NJ*NL*sizeof(DATA_TYPE));
G = (DATA_TYPE*)malloc(NI*NL*sizeof(DATA_TYPE));
G_outputFromGpu = (DATA_TYPE*)malloc(NI*NL*sizeof(DATA_TYPE));
int i;
init_array(A, B, C, D);
read_cl_file();
cl_initialization_fusion();
//cl_initialization();
cl_mem_init(A, B, C, D, E, F, G);
cl_load_prog();
cl_launch_kernel();
errcode = clEnqueueReadBuffer(clCommandQue[0], g_mem_obj, CL_TRUE, 0, sizeof(DATA_TYPE) * NI * NL, G_outputFromGpu, 0, NULL, NULL);
if(errcode != CL_SUCCESS) printf("Error in reading GPU mem\n");
t_start = rtclock();
mm3_cpu(A, B, C, D, E, F, G);
t_end = rtclock();
fprintf(stdout, "CPU Runtime: %0.6lfs\n", t_end - t_start);
compareResults(G, G_outputFromGpu);
cl_clean_up();
free(A);
free(B);
free(C);
free(D);
free(E);
free(F);
free(G);
free(G_outputFromGpu);
return 0;
}
int main(int argc, char** argv)
{
double t_start, t_end;
DATA_TYPE* A;
DATA_TYPE* B;
DATA_TYPE* C;
DATA_TYPE* D;
DATA_TYPE* E;
DATA_TYPE* F;
DATA_TYPE* G;
DATA_TYPE* G_outputFromGpu;
A = (DATA_TYPE*)malloc(NI*NK*sizeof(DATA_TYPE));
B = (DATA_TYPE*)malloc(NK*NJ*sizeof(DATA_TYPE));
C = (DATA_TYPE*)malloc(NJ*NM*sizeof(DATA_TYPE));
D = (DATA_TYPE*)malloc(NM*NL*sizeof(DATA_TYPE));
E = (DATA_TYPE*)malloc(NI*NJ*sizeof(DATA_TYPE));
F = (DATA_TYPE*)malloc(NJ*NL*sizeof(DATA_TYPE));
G = (DATA_TYPE*)malloc(NI*NL*sizeof(DATA_TYPE));
G_outputFromGpu = (DATA_TYPE*)malloc(NI*NL*sizeof(DATA_TYPE));
fprintf(stdout, "<< Linear Algebra: 3 Matrix Multiplications (E=A.B; F=C.D; G=E.F) >>\n");
init_array(A, B, C, D);
t_start = rtclock();
mm3_OMP(A, B, C, D, E, F, G_outputFromGpu);
t_end = rtclock();
fprintf(stdout, "GPU Runtime: %0.6lfs\n", t_end - t_start);
t_start = rtclock();
mm3_cpu(A, B, C, D, E, F, G);
t_end = rtclock();
fprintf(stdout, "CPU Runtime: %0.6lfs\n", t_end - t_start);
compareResults(G, G_outputFromGpu);
free(A);
free(B);
free(C);
free(D);
free(E);
free(F);
free(G);
free(G_outputFromGpu);
return 0;
}
