// MatrixTranspose by Amir // I implemented this code long ago when I was a student (almost 2013). // I don't know why it gives "Time Limit Exceeded". even with small numbers. maybe it won't link. #include #include #include #include #define NUM_ROWS 100 #define NUM_COLS 60 #define NUM_THREADS 2 typedef struct { int *a, *out; int rows, cols, tid; } argument; void *parallelTranspose(void *arg) { argument *args = (argument *) arg; int *a = args->a; int *out = args->out; int tid = args->tid; int rows = args->rows; int cols = args->cols; int chunkSize = rows / NUM_THREADS; int startRow = tid * chunkSize; int endRow = startRow + chunkSize; for (int i = startRow; i < endRow; ++i) { for (int j = 0; j < cols; ++j) { out[j * rows + i] = a[i * cols + j]; } } return NULL; } void sequntialTranspose(int *out, int *a, int rows, int cols) { for (int i = 0; i < rows; ++i) { for (int j = 0; j < cols; ++j) { out[j * rows + i] = a[i * cols + j]; } } } void printMatrix(int *a, int rows, int cols) { for (int i = 0; i < rows; ++i) { for (int j = 0; j < cols; ++j) { printf("%5d", a[i * cols + j]); } printf("\n"); } printf("\n=============\n"); } void parallel(int *out, int *a) { pthread_t threads[NUM_THREADS]; argument args[NUM_THREADS]; for (int i = 0; i < NUM_THREADS; ++i) { args[i].a = a; args[i].out = out; args[i].rows = NUM_ROWS; args[i].cols = NUM_COLS; args[i].tid = i; pthread_create(&threads[i], NULL, parallelTranspose, (void *) &args[i]); } int *status; for (int i = 0; i < NUM_THREADS; ++i) { pthread_join(threads[i], (void **) &status); } } int main() { int *a = (int *) malloc(sizeof(int) * NUM_COLS * NUM_ROWS); int *out = (int *) malloc(sizeof(int) * NUM_COLS * NUM_ROWS); srand(time(NULL)); for (int i = 0; i < NUM_ROWS; ++i) { for (int j = 0; j < NUM_COLS; ++j) { a[i * NUM_COLS + j] = rand() % 10; } } sequntialTranspose(out, a, NUM_ROWS, NUM_COLS); parallel(out, a); return 0; }