Floating Point Underflow

[Tobias Schlüter]

Ray Nachlinger wrote:
> On Fri, 26 Sep 2008, Steve Kargl wrote:
> 
>> On Fri, Sep 26, 2008 at 10:21:27AM -0500, Ray Nachlinger wrote:
>>> On Fri, 26 Sep 2008, Steve Kargl wrote:
>>>
>>>> On Fri, Sep 26, 2008 at 09:43:15AM -0500, Ray Nachlinger wrote:
>>>>> Is there a gfortran compiler option that will
>>>>> prevent a fatal "Floating Point Underflow" error
>>>>> message on 64bit Windows. This problem does not
>>>>> occur on 32bit systems.
>>>>>
>>>> Is this a compile time or runtime problem?
>>>   Only runtime. Have big program that works nicely on
>>>   Solaris, 32, 64 bit Linux, 32bit OS X, and 32 bit
>>>   Windows (mingw). Received this error report from
>>>   a user on 64bit Windows. Traceback from place
>>>   what an underflow should be expected.
>> Underflow should never be a fatal error.   Is the error
>> message a gfortran error message or an OS error message?
>> It appears that you need to contact the OS vendor.
> 
>   I agree it never should be but have had it before
>   (first sparcs required -abrupt_underflow not to
>   abort). This is definitely a WINDOWS message, 
>   am not looking forward to contacting Microsoft.

Chapter 11 in Metcalf should be able to solve your user's problem, see 
http://books.google.de/books?id=iLj_xAmxscgC&pg=PA219&lpg=PA219&dq=ieee+floating+point+exception+fortran&source=web&ots=8CKY_UazsA&sig=Qw2iCgcfnoeVZwnawHyFjbt2ozs&hl=en&sa=X&oi=book_result&resnum=9&ct=result#PPA219,M1

Cheers,
  -Tobi

-- 
Tobias Schlüter
Am Coulombwall 1, Zi. 326
85748 Garching b. München
Tel.: +49/89/289-14139