Museum

VEC_$MAX_I                        Domain/OS                         VEC_$MAX_I


NAME
     vec_$max_i - find the maximum absolute value in a vector from a single-
     precision matrix

SYNOPSIS (C)
     #include <apollo/base.h>
     #include <apollo/vec.h>

     void vec_$max_i(
          float *vector,
          long int &inc,
          long int &length,
          float *result,
          long int *location)

SYNOPSIS (Pascal)
     %include '/sys/ins/base.ins.pas';
     %include '/sys/ins/vec.ins.pas';

     procedure vec_$max_i(
          in vector: univ vec_$real_vector;
          in inc: integer32;
          in length: integer32;
          out result: real;
          out location: integer32);

SYNOPSIS (FORTRAN)
     %include '/sys/ins/base.ins.ftn'
     %include '/sys/ins/vec.ins.ftn'

           parameter (nvec = 10)

           real vector(nvec), result
           integer*4 length, inc, location

           call vec_$max_i(vector, inc, length, result, location)

DESCRIPTION
     Vec_$max_i searches through the length elements of vector selected by
     inc, and supplies the value and location of the element with the greatest
     absolute value.

     Through appropriate choice of inc, a program can use vec_$max_i to search
     a vector within a matrix.  To search the Mth vector in a matrix, choose
     inc equal to the number of vectors in the matrix, and place the Mth ele-
     ment of the matrix array at the beginning of vector.

     In C, the resulting operation is

          result = (float)fabs(vector[0]);
          location = 1;
          j = inc;
          for (i = 1; i < length, ++i) {
               if ((float)fabs(vector[j]) > result) {
                    location = i + 1;
                    result = (float)fabs(vector[j]);
               }
               j += inc;
          }

     In Pascal, the resulting operation is

          result := abs(vector[1]);
          location := 1;
          j := 1 + inc;
          for 10 i := 2 to length do
               begin
               if (abs(vector[j]) > result) then
                    begin
                    location := i;
                    result := abs(vector[j]);
                    end
               j := j + inc;
               end

     In FORTRAN, the resulting operation is

           result = abs(vector(1))
           location = 1
           j = 1 + inc
           do 10 i = 2, length
               if (abs(vector(j)) .gt. result) then
                    location = i
                    result = abs(vector(j))
               endif
           j = j + inc
       10  continue

     vector
          The array to search.

     inc  An increment for the index of vector that selects the elements to
          search.

     length
          The number of elements to search.

     result
          The maximum absolute value of all the elements searched.

     location
          The location of the element with the greatest absolute value.  The
          location supplied in location is just the index of the element with
          the greatest absolute value in FORTRAN or Pascal (if vector is
          declared to begin with index 1).  In C, location - 1 is the index of
          the element.

NOTES
     In C and Pascal, vec_$max_i searches a column vector; whereas in FORTRAN,
     it searches a row vector.

SEE ALSO
     vec_$dmax_i, vec_$imax16_i, vec_$imax_i, vec_$max.