Strange SIGFPE (underflow?) with gfortran [ -mieee ?]

Rob spamrefuse@yahoo.com
Fri Oct 20 23:30:00 GMT 2006


Brooks Moses wrote:
> 
> Given the quoted gdb results, it appears that this
> omitted negative sign is just a typo in the email
> message, and that the actual line in the program
> is the correct "R1=-708.396", with the sign
> unchanged.

Yes, indeed that was a 'cut&paste' typo; sorry.
All the numbers were negative.

I've come to following conconlusion.
One of two flags solved my runtime SIGFPE:
  -mieee
or
  -mfp-trap-mode=su

I've tried each of them separately and successfully.
The first one compromises on the code efficiency,
so I concluded that the second one would be my best
choice.

Note: this is specifically an Alpha system issue.
Is there a GCC mailing list where I can find
people that are familiar with these problems and
Alpha related issues? I'd like to double check
whether I'm doing the right thing here.

Thanks,
Rob.


__________________________________________________
Do You Yahoo!?
Tired of spam?  Yahoo! Mail has the best spam protection around 
http://mail.yahoo.com 



More information about the Fortran mailing list