void main(){
float *x,*y, *temp_x, *temp_y, *alpha, *temp_alpha;
// int size_x=50;
// int size_y=10;
int size_x=1600;
int size_y=800;
if( NULL == (x = memalign(16, sizeof(float)*size_x))){
printf("\nfailed for x\n");
exit(1);
}
if(NULL == ( y = memalign(16, sizeof(float)*size_y))){
printf("\nfailed for y\n");
free(x);
exit(1);
}
if(NULL == (alpha = memalign(16, sizeof(float)*4))){
printf("\nfailed for alpha\n");
free(x);
free(y);
exit(1);
}
int i;
//fill up arrays
for(i=0; i < size_x; i++){
x[i]= (float)(i*1000);
}
for(i=0; i < size_y; i++){
y[i]= (float)(i*1000);
}
for(i=0; i <4; i++){
alpha[i]=(float)1;
}
print_arrays(x, size_x, y, size_y, alpha,1);
//warmup_scalar(x, y, size_y, alpha);
warmup_vector(x, y, size_y, alpha);
print_arrays(x, size_x, y, size_y, alpha,1);
free(x);
free(y);
free(alpha);
}
int main()
{
float cel[N], far[N];
init_array(cel, N);
cel2far(cel, far, N);
print_arrays(cel, far, N);
system("pause");
return 0;
}
void *pthreads_each(void *rank_ptr) {
uint32_t rank;
#if !defined(DEBUG)
uint32_t bytes_transferred;
#endif
#if defined(PAPI_ENABLED) && !defined(DEBUG)
int num_sets;
PAPI_event_set_wrapper_t* event_sets;
papi_filter_events(desired_events, num_desired, &event_sets, &num_sets);
#endif
rank = *((uint32_t *) rank_ptr);
#if defined(AFFINITY_ENABLED)
Affinity_Bind_Thread(rank);
Affinity_Bind_Memory(rank);
#endif
#if defined(DEBUG)
init_arrays(rank);
barrier_wait(&my_barrier, rank);
stream_per_thread[read_arrays_case][write_arrays_case][write_type_case][load_type_case][loop_unroll_case][prefetch_distance_case](rank);
barrier_wait(&my_barrier, rank);
if (rank == 0) {
print_arrays();
}
#else
TIMER_MAKE_MEASUREMENTS(stream_per_thread[read_arrays_case][write_arrays_case][write_type_case][load_type_case][loop_unroll_case][prefetch_distance_case](rank), results, rank, NUM_TRIALS);
if (rank == 0) {
printf("TIME (IN SECONDS)\n");
print_max_timer_measurements(results, numThreads, NUM_TRIALS, median_counts_per_sec);
printf("\nBANDWIDTH (IN GB/S, WHERE 1 GB/S = 10^9 B/S)\n");
bytes_transferred = 0;
if (write_type_case == 0) {
bytes_transferred = (numReadArraysPerThread + 2*numWriteArraysPerThread) * (numThreads * threadArrayLength * NUM_BYTES_PER_DOUBLE);
}
else if (write_type_case == 1) {
bytes_transferred = (numReadArraysPerThread + numWriteArraysPerThread) * (numThreads * threadArrayLength * NUM_BYTES_PER_DOUBLE);
}
print_bandwidth_measurements(results, bytes_transferred, numThreads, NUM_TRIALS, median_counts_per_sec);
printf("\n\n");
}
#endif
pthread_exit((void*) 0);
}
int main()
{
int idx, bound, temp;
bound = 16;
populate_arrays();
print_arrays();
bound = 16;
idx = 0;
while (idx < bound - 1)
{
if (array_1[idx] > array_1[idx + 1])
{
temp = array_1[idx];
array_1[idx] = array_1[idx + 1];
array_1[idx + 1] = temp;
idx = 0;
continue;
}
idx = idx + 1;
}
idx = 0;
while (idx < bound - 1)
{
if (array_2[idx] > array_2[idx + 1])
{
temp = array_2[idx];
array_2[idx] = array_2[idx + 1];
array_2[idx + 1] = temp;
idx = 0;
continue;
}
idx = idx + 1;
}
idx = 0;
while (idx < bound - 1)
{
if (array_3[idx] > array_3[idx + 1])
{
temp = array_1[idx];
array_3[idx] = array_3[idx + 1];
array_3[idx + 1] = temp;
idx = 0;
continue;
}
idx = idx + 1;
}
idx = 0;
while (idx < bound - 1)
{
if (array_4[idx] > array_4[idx + 1])
{
temp = array_4[idx];
array_4[idx] = array_4[idx + 1];
array_4[idx + 1] = temp;
idx = 0;
continue;
}
idx = idx + 1;
}
print_arrays();
}
