矩阵相乘

 

/*
** Multiply two matrices together.
*/

void matrix_multiply( int *m1, int *m2, register int *r,
    int x, int y, int z )
{
 register int *m1p;
 register int *m2p;
 register int k;
 int row;
 int column;

 /*
 ** The outer two loops go through the solution matrix element
 ** by element.  Because this is done in storage order, we
 ** can access these elements using indirection on r.
 */
 for( row = 0; row < x; row += 1 ){
  for( column = 0; column < z; column += 1 ){
   /*
   ** Compute one value in the result.  This is
   ** done by getting pointers to the proper
   ** elements in m1 and m2, and then advancing
   ** them as we go through the loop.
   */
   m1p = m1 + row * y;
   m2p = m2 + column;
   *r = 0;

   for( k = 0; k < y; k += 1 ){
    *r += *m1p * *m2p;
    m1p += 1;
    m2p += z;
   }

   /*
   ** Advance r to point to the next element.
   */
   r++;
  }
 }
}