Re: arbitrary precision gfortran
arbitrary precision has been implemented -- in library form, not as a
compiler intrinsic. the likely reasoning is that there is no
agreed-upon arbitrary precision standand (e.g. real*8 is an
implementation of the ieee floating point standard). however, there
are libraries and modules (in a sense, an extension to the language)
that implement the arbitrary precision type.
http://crd.lbl.gov/~dhbailey/mpdist/
regards,
mike
On 12/16/05, Yigal Weinstein <yigal.weinstein@gmail.com> wrote:
> No one has formatted fortran into arbitrary precision here? Well i'll be
> darned. Well thanks a bunch. Is it because real*8 has been acceptable or
> some other reason?
>
>
>
> On 16 Dec 2005 13:12:25 -0000, Euclid Uranium <euclid@miskatonic.edu> wrote:
> > > > Is there a simple way of implementing arbitary precision in gnu
> fortran
> > > > compilation? I got rather confused about what could be easily done
> and what
> > > > cannot be easily done researching it myself. So any feedback would be
> > > > great.
> > >
> > > You mean like having pi to any desired number of digits? As far as I
> > > know, there is no intrinsic support for this in any FORTRAN (or C, for
> > > that matter) compiler.
> >
> > Don't do it!!!
> >
> > http://www.everything2.com/index.pl?node_id=1302963
> >
> > pi is believed to be a normal number, ie containing all finite digit
> > sequences. Computer files are as you know just finite sequences of
> > numbers.
> >
> > The longer you calculate pi, the more likely you are to generate
> > child-porn, copyright-infringing, DMCA-violating, or
> > scientology-provoking files!
> >
> >
> > --
> > To UNSUBSCRIBE, email to
> debian-science-request@lists.debian.org
> > with a subject of "unsubscribe". Trouble? Contact
> listmaster@lists.debian.org
> >
> >
>
>
Reply to: