------- Comment #2 from burnus at gcc dot gnu dot org 2007-11-22 16:11 -------
I believe there is indeed a bug in gfortran. The standard says that the "the
nearest different machine-representable number" is returned.
However, gfortran returns a number which is not representable (denormal or
smaller?) and rounded to zero (= same number).
Result (gfortran 4.3.0):
a) The program is wrongly rejected
b) With -fno-range-check, the wrong number is used. For the example the dump
shows:
static real(kind=8) C.859 = 0.0;
The expected result is some denormal number like:
5.E-324, 4.9406564584124654E-324, 2.225073858507201E-308,
4.94065645841246544E-324, 4.940656458412465E-324.
(The ifort result of 2e-308 looks wrong in this regard as there is a nearer
number.)
gfortran 4.1.3 returns: 4.940656458412465E-324
gfortran 4.2.2 rejects it and has an ICE with -fno-range-check
Besides fixing this bug, we should also consider mentioning -fno-range-check in
the error message. (Though, in this case -fno-range-check would have concealed
the real bug.)
The Fortran 2003 standard states:
"NEAREST (X, S)
Description. Returns the nearest different machine-representable number in a
given direction.
X shall be of type real.
S shall be of type real and not equal to zero.
Result Characteristics. Same as X.
Result Value. The result has a value equal to the machine-representable number
distinct from X and nearest to it in the direction of the infinity with the
same sign as S.
NOTE 13.16 Unlike other floating-point manipulation functions, NEAREST operates
on machine-representable numbers rather than model numbers. On many systems
there are machine-representable numbers that lie between adjacent model
numbers."
--
burnus at gcc dot gnu dot org changed:
What |Removed |Added
----------------------------------------------------------------------------
CC| |burnus at gcc dot gnu dot
| |org
OtherBugsDependingO| |32834
nThis| |
GCC host triplet|Linux i686 |
Keywords| |rejects-valid, wrong-code
Known to fail| |4.2.2 4.3.0
Known to work| |4.1.3
Summary|problem with NEAREST |[4.2, 4.3 Regression]
|intrinsic function |NEAREST can return wrong
| |numbers
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=34192
