parallel_matmul.c

//
// Created by gunavaran on 12/8/20.
//

#include <stdlib.h>
#include <stdio.h>
#include <time.h>
#include <pthread.h>

#define n 1024
#define s 512
#define NUM_THREADS 2

double A[n][n];
double B[n][n];
double C[n][n];
double D[n][n];

void compare_matrix(double *matrixA, double *matrixB) {

    for (int j = 0; j < n; j++)
        for (int i = 0; i < n; i++)
            if (matrixA[i + j * n] != matrixB[i + j * n]) {
                printf("FAILED \n");
            }

}

void *parallel_matrixmul(void *threadarg) {
    int threadId = (int) threadarg;
    int lower_limit = (n / NUM_THREADS) * threadId;
    int upper_limt = lower_limit + n / NUM_THREADS;

    for (int ih = lower_limit; ih < upper_limt; ih += s) {
        for (int jh = 0; jh < n; jh += s) {
            for (int kh = 0; kh < n; kh += s) {
                for (int il = 0; il < s; il++) {
                    for (int kl = 0; kl < s; kl++) {
                        for (int jl = 0; jl < s; jl++) {
                            C[ih + il][jh + jl] += A[ih + il][kh + kl] * B[kh + kl][jh + jl];
                        }
                    }
                }
            }

        }
    }

    pthread_exit(NULL);
}

int main() {
    int rc;
    //populate the matrices with random values between 0.0 and 1.0
    for (int i = 0; i < n; i++) {
        for (int j = 0; j < n; j++) {

            A[i][j] = (double) rand() / (double) RAND_MAX;
            B[i][j] = (double) rand() / (double) RAND_MAX;
            C[i][j] = 0;
        }
    }

    struct timespec start, end;
    double time_spent;

//    for (int i = 0; i < n; i++) {
//        for (int k = 0; k < n; k++) {
//            for (int j = 0; j < n; j++) {
//                D[i][j] += A[i][k] * B[k][j];
//            }
//        }
//    }


    clock_gettime(CLOCK_REALTIME, &start);

    //initialize threads
    pthread_t threads[NUM_THREADS];

    for (int i = 0; i < NUM_THREADS; i++) {
        rc = pthread_create(&threads[i], NULL, parallel_matrixmul, (void *) i);
        if (rc) {
            printf("ERROR; return code from pthread_join() is %d\n", rc);
            exit(-1);
        }
    }

    for (int i = 0; i < NUM_THREADS; i++) {
        rc = pthread_join(threads[i], NULL);
        if (rc) {
            printf("ERROR; return code from pthread_join() is %d\n", rc);
            exit(-1);
        }
    }

    clock_gettime(CLOCK_REALTIME, &end);
    time_spent = (end.tv_sec - start.tv_sec) + (end.tv_nsec - start.tv_nsec) / 1000000000.0;
    printf("Elapsed time in seconds: %f \n", time_spent);

//    compare_matrix(C, D);
    return 0;

}