File:  [local] / rpl / lapack / lapack / dlaisnan.f
Revision 1.7: download - view: text, annotated - select for diffs - revision graph
Fri Aug 13 21:03:49 2010 UTC (13 years, 9 months ago) by bertrand
Branches: MAIN
CVS tags: rpl-4_0_19, rpl-4_0_18, HEAD
Patches pour OS/2

    1:       LOGICAL FUNCTION DLAISNAN( DIN1, DIN2 )
    2: *
    3: *  -- LAPACK auxiliary routine (version 3.2.2) --
    4: *  -- LAPACK is a software package provided by Univ. of Tennessee,    --
    5: *  -- Univ. of California Berkeley, Univ. of Colorado Denver and NAG Ltd..--
    6: *     June 2010
    7: *
    8: *     .. Scalar Arguments ..
    9:       DOUBLE PRECISION   DIN1, DIN2
   10: *     ..
   11: *
   12: *  Purpose
   13: *  =======
   14: *
   15: *  This routine is not for general use.  It exists solely to avoid
   16: *  over-optimization in DISNAN.
   17: *
   18: *  DLAISNAN checks for NaNs by comparing its two arguments for
   19: *  inequality.  NaN is the only floating-point value where NaN != NaN
   20: *  returns .TRUE.  To check for NaNs, pass the same variable as both
   21: *  arguments.
   22: *
   23: *  A compiler must assume that the two arguments are
   24: *  not the same variable, and the test will not be optimized away.
   25: *  Interprocedural or whole-program optimization may delete this
   26: *  test.  The ISNAN functions will be replaced by the correct
   27: *  Fortran 03 intrinsic once the intrinsic is widely available.
   28: *
   29: *  Arguments
   30: *  =========
   31: *
   32: *  DIN1    (input) DOUBLE PRECISION
   33: *
   34: *  DIN2    (input) DOUBLE PRECISION
   35: *          Two numbers to compare for inequality.
   36: *
   37: *  =====================================================================
   38: *
   39: *  .. Executable Statements ..
   40:       DLAISNAN = (DIN1.NE.DIN2)
   41:       RETURN
   42:       END

CVSweb interface <joel.bertrand@systella.fr>