块矩阵版Fox在MPI中的矩阵乘法（使用C）

Question

我很长时间以来一直坚持这个程序，所以基本上我必须在p x p笛卡尔网格上乘以2 n x n矩阵，p＆lt; = n（n可以被p整除）

在您的程序中，2D网格的过程0应该读取矩阵A和B. 从文件。示例：

＆LT;矩阵维数n>

＆LT;矩阵A＆gt;

＆LT;矩阵B＆gt;

4

-11.0 -12.0 6.0 9.0
-11.0 12.0 2.0 4.0
8.0 6.0 12.0 11.0
6.0 4.0 -10.0 1.0
2.0 11.0 1.0 7.0
3.0 1.0 -2.0 4.0
1.0 9.0 11.0 3.0
6.0 1.0 -1.0 5.0

我能够使用进程0将矩阵读入两个数组A和B，但我需要对程序的其余部分提供一些帮助。

    #include <mpi.h> 
    #include <stdio.h> 
    #include <stdlib.h> 

    int main( int argc, char *argv[] ) 
    { 
    int p, rank, k, n;
     float *A = NULL, *B=NULL;
     char filename[FILENAME_MAX];
     //MPI_Status status;
     FILE *file;
     MPI_Init( &argc, &argv );
     MPI_Comm_size( MPI_COMM_WORLD, &p );
     MPI_Comm_rank( MPI_COMM_WORLD, &rank );

     // asks for filename
     if ( rank == 0 ) { 
     printf("p=%d", p);
     printf("Enter filename: ");
     scanf("%s", filename);

     // file is opened
     file = fopen(filename, "r");
     fscanf( file, "%d", &n );
     printf( "\nMatrix size = %d\n", n ); 
     // As per the expected format of the file the size of the matrix comes first in the file
     } 
     // broadcasts the value of n to all the processes
     MPI_Bcast( &n, 1, MPI_INT, 0, MPI_COMM_WORLD );

     if ( n*n > p ) 
     {
     // since p perfectly divides the n x n matrixes assumed
     if ( rank == 0 ) 
     { 
     printf( "No. of processes must be at least %d\n", n*n );
     fclose( file );
     } 
     MPI_Finalize( );
     exit( 1 );
     }

     //  Read and print input data matrix A and B 
     if ( rank == 0 ) 
     {

     A = (float *) malloc( n*n*sizeof(float) );
     printf( "\nMatrix A:\n" );
     for ( k=0;k<n*n; k++ ) 
     { 
     fscanf( file, "%f", &A[k] );
     printf( "%7.1f", A[k] );
     if ( (k+1)%n == 0 ) printf( "\n" );
     }
     // reads matrix B
     B = (float *) malloc( n*n*sizeof(float) ); 
     printf( "\nMatrix B:\n" );
     for ( k=0;k<n*n; k++ ) 
     { 
     fscanf( file, "%f", &B[k] );
     printf( "%7.1f", B[k] );
     if ( (k+1)%n == 0 ) printf( "\n" );
     } fclose( file );

     }

     MPI_Finalize( ); return 0; 
     }