MPI matrix-vector-multiplication returns sometimes correct sometimes weird values

问题

I have the following code:

    //Start MPI...
MPI_Init(&argc, &argv);

int size = atoi(argv[1]);
int delta = 10;
int rnk;
int p;
int root = 0;

MPI_Status mystatus;
MPI_Comm_rank(MPI_COMM_WORLD, &rnk);
MPI_Comm_size(MPI_COMM_WORLD, &p);

//Checking compatibility of size and number of processors
assert(size % p == 0);

//Initialize vector...
double *vector = NULL;
vector = malloc(size*sizeof(double));
double *matrix = NULL;

//Rank 0 -----------------------------------
if (rnk == 0) {

    //Initialize vector...
    srand(1);
    for (int i = 0; i < size; i++) {
        vector[i] = rand() % delta + 1;
    }
    printf("Initial vector:");
    print_vector(vector, size);

    //Initialize matrix...
    matrix = malloc(size*size*sizeof(double));
    srand(2);
    for (int i = 0; i < (size*size); i++) {
        matrix[i] = rand() % delta + 1;
    }

    //Print matrix...
    printf("Initial matrix:");
    print_flat_matrix(matrix, size);

}

//Calculating chunk_size...
int chunk_size = size/p;

//Initialize submatrix..
double *submatrix = malloc(size*chunk_size*sizeof(double));

//Initialize result vector...
double *result = malloc(chunk_size*sizeof(double));

//Broadcasting vector...
MPI_Bcast(vector, size, MPI_DOUBLE, root, MPI_COMM_WORLD);
MPI_Barrier(MPI_COMM_WORLD);

//Scattering matrix...
MPI_Scatter(matrix, (size*chunk_size), MPI_DOUBLE, submatrix, (size*chunk_size), MPI_DOUBLE, root, MPI_COMM_WORLD);
MPI_Barrier(MPI_COMM_WORLD);

printf("I am rank %d and first element of my vector is: %f and of my matrix1: %f/matrix2: %f/matrix3: %f/matrix4: %f\n", rnk, vector[0], submatrix[0], submatrix[1], submatrix[2], submatrix[3]);

//Calculating...
for (int i = 0; i < chunk_size; i++) {
    for (int j = 0; j < size; j++) {
        result[i] += (submatrix[(i*size)+j] * vector[j]);
        printf("Rank %d; current result: %f, ", rnk, result[i]);
    }
    printf("\n");
    printf("Rank %d; result: %f...\n", rnk, result[i]);
}

printf("Rank: %d; first result: %f\n", rnk, result[0]);


double *final_result = NULL;
//Rank 0 -----------------------------------
if (rnk == 0) {
    final_result = malloc(size*sizeof(double));
}

//Gather...
MPI_Gather(result, chunk_size, MPI_DOUBLE, final_result, chunk_size, MPI_DOUBLE, root, MPI_COMM_WORLD);


//Rank 0 -----------------------------------
if (rnk == 0) {
    printf("Final result:\n");
    print_vector(final_result, size);

    free(matrix);
    free(final_result);
}

free(submatrix);
free(result);
free(vector);

MPI_Finalize();

When I run the program it runs to completion without errors, but the values I print at the end aren't always the correct ones. Sometimes I receive the vector with correct output, sometimes it's partially correct and sometimes completely wrong. The wrong values are either wrong by exactly the value of 2 or they are some very long useless sequence of number (Which seems to me there has to be a wrong memory access, but I can't find anything and also weird, because it sometimes works).

I also always choose my size so it'll fit the number of created processes by mpi. mpi creates 4 processes on my machine (tested and checked value), so for testing my algorithm I've always choosen 4 as the value for size. Same problem occurs also with bigger sizes.

Looking forward to your help and inputs guys, thanks in advance!

PS: I am in C

回答1:

Are you familiar with valgrind? It will draw your attention to the problematic line straight away.

Your trouble appears to be this line:

result[i] += (submatrix[(i*size)+j] * vector[j]);

What was result[] initially? It was pulled off the heap. Sometimes, if you are lucky, it will be zero. Don't count on luck with C.

There are many ways to initialize the array. Here are a few approaches, listed in order by most likely to be optimized:

Allocate result[] with calloc:

double *result = calloc(chunk_size , sizeof(double));

Or, initialize the array with memset:

double *result = malloc(chunk_size *sizeof(double));
memset(result, 0, chunk_size *sizeof(double));

or, one could loop over the array

for (i=0; i < chunk_size; i++)
    result[i] = 0.0

来源：https://stackoverflow.com/questions/28811791/mpi-matrix-vector-multiplication-returns-sometimes-correct-sometimes-weird-value

标签

mpi

matrix-multiplication