The XMULT function computes the matrix product like the matrix multiplication operator (*) except XMULT uses extended precision to accumulate sums of products. You should use the XMULT function only when you need great accuracy.

The following program uses the XMULT function:

  a=1e13;
  b=1e13;
  c=100*a;
  a=a+1;
  x=c || a || b || c;
  y=c || a || (-b) || (-c);

  z=xmult(x,y`);  /* correct answer */
  print z [format=16.0];

  wrong = x * y`;  /* loss of precision */
  print wrong [format=16.0];

                       Z

                 20000000000001

                     WRONG

                 19997367730176