From bug-octave-request at bevo dot che dot wisc dot edu  Wed May 27 03:34:30 1998
Subject: Re[2]: FFLAGS on Alpha  
From: "PETER HOPFGARTNER"<phopfgartner at memc dot inet dot it>
To: <bug-octave at bevo dot che dot wisc dot edu>
Date: Wed, 27 May 98 10:22:59 +0100


     OK, learned something more! Thanks.
     
     Next time I will compile Octave I will try out -fpe3.
     
     Reagarding the flags passed to gcc/g77 you seem to be right, too.
     
                                        Peter


______________________________ Reply Separator _________________________________
Subject: Re: FFLAGS on Alpha  
Author:  <bug-octave at bevo dot che dot wisc dot edu > at NOVARALINK
Date:    5/26/98 10:54 AM


On 25-May-1998, PETER HOPFGARTNER <phopfgartner at memc dot inet dot it> wrote:
     
|      ...and maybe -fpe4 in Digital's f77 is used
     
Is better to have `gradual underflow' instead of setting the result of 
an underflow to zero?  Can someone explain why that might be?
     
If so, I think -fpe3 is better than -fpe4 because with -fpe4, you'll 
get annoying warning messages printed for each overflow and division 
by zero.
     
jwe