round(3F)

NAME
     round: anint, dnint, qnint, nint, inint, jnint, idnint, iqnint, iidnnt,
     jidnnt, iiqnnt, jiqnnt, kiqnnt - FORTRAN nearest integer functions

SYNOPSIS
     integer i
     integer*2 ii
     integer*4 ji
     integer*8 ki
     real r1, r2
     real*4 r3
     double precision dp1, dp2
     real*16 qp1, qp2
     real*8 dp3
     real*16 qp3

     r2 = anint(r1)

     dp2 = dnint(dp1)
     dp2 = anint(dp1)

     qp2 = qnint(qp1)
     qp2 = anint(qp1)

     i = nint(r2)

     ii = inint(r3)
     ii = nint(r3)

     ji = jnint(r3)
     ji = nint(r3)

     i = idnint(dp1)
     i = nint(dp1)

     ii = iidnnt(dp3)
     ji = jidnnt(dp3)

     i = iqnint(qp1)
     i = nint(qp1)

     ii = iiqnnt(qp3)
     ji = jiqnnt(qp3)
     ki = kiqnnt(qp3)

DESCRIPTION
     anint returns the nearest whole real number to its real argument (i.e.,
     int(a+0.5) if a > 0, int(a-0.5) otherwise).  dnint does the same for its
     double-precision argument, qnint does the same for its real*16 argument.
     anint is the generic form of anint, dnint, and qnint, performing the same
     operation and returning the data type of its argument.
     nint returns the nearest integer to its real argument.  inint returns the
     nearest integer*2 to its real*4 argument.  jnint returns the nearest
     integer*4 to its real*4 argument.  idnint returns the nearest integer to
     its double precision argument.  iqnint returns the nearest integer to its
     real*16 argument.  nint is the generic form of inint, jnint, idnint and
     iqnint. idnint is also the generic form for iidnnt, which returns the
     nearest integer*2 to its real*8 argument, and jidnnt, which returns the
     nearest integer*4 to its real*8 argument.  iqnint is also the generic
     form for iiqnnt, which returns the nearest integer*2 to its real*16
     argument, jiqnnt, which returns the nearest integer*4 to its real*16
     argument, and kiqnnt, which returns the nearest integer*8 to its real*16
     argument.  When nint, idnint, or iqnint is specified as an argument in a
     subroutine call or function reference, the compiler supplies either an
     integer*2, integer*4, or integer*8 function depending on the -i2 command
     line option.