◄Up► ◄Contents► ◄Index► ◄Back►
─────$FLOATCALLS and $NOFLOATCALLS──────────────────────────────────────────
 
     Action
 
     $FLOATCALLS causes floating-point operations to be processed by
     calls to library subroutines; $NOFLOATCALLS causes floating-point
     operations to be processed by compiler-generated, inline
     coprocessor instructions.
 
     Syntax  $[NO]FLOATCALLS
 
     Remarks
 
     $NOFLOATCALLS is the default.
 
     Example
 
     $FLOATCALLS
           REAL  x, sine
 
           WRITE  (*, 100)
       100 FORMAT (1X, 'ENTER x: '\)
           READ   (*, '(F10.5)') x
 
           WRITE  (*, 200) x, SINE (x, .00001)
           200  FORMAT (1X, 'THE SINE OF ', F10.5, ' = ', F9.6)
           END
 
     C     The function calculates the sine of X using a power series.
     C     Successive terms are calculated until less than eps.
 
     C     Library calls are generated instead of in-line instructions,
     C     letting this routine run on machines without a coprocessor.
 
           REAL FUNCTION SINE (x, eps)
 
           REAL x, y, z, next, i, eps
 
           z    = AMOD (x, 6.2831853)
           y    = z
           i    = 4.0
           next = -z * z * z / 6.0
       100 IF (ABS (next) .GE. eps) THEN
               y    =  y + next
               next = -next * z * z / (i * (i + 1.0))
               i    =  i + 2.0
               GOTO 100
           END IF
           SINE = y
           END
                                    -♦-