Original-Sender: [log in to unmask]
Date: Wed, 22 Nov 2000 09:39:46 -0700
From: Walt Brainerd <[log in to unmask]>
Wei-Yin Loh wrote:
> I haven't had good experience with F because of its restrictions. If
> you want compatibility with Lahey, I would recommend Lahey's elf90. It
> used to be free but now costs a little more than F. I've written and
> compiled a 35,000 line program in elf90 and it ported without any
> problems on all platforms available to me (LF90 and DVF on
> Intel/Windows, FortranPlus on Intel/OS2, f90 on DEC and SUN, and
> Absoft Fortran on Mac and Intel/Linux).
I find this all a little puzzling:
The F and ELF restrictions are very similar, although it is
true that F is more restrictive than ELF. Which differences
cause problems?
I don't recall exactly what caused the problems since it has been two
years at least. I needed to use several lapack routines and had to
convert them from f77 to f90 by hand. There were some f77 constructs
that I couldn't easily work around within the F restrictions but I
could within elf90.
Re: "costs a little more than F": F is free:
ftp://web01.meltingsand.com/pub/F
I know this. In my previous message, I was referring the original
poster's statement that the F CD costs $60.
--
Wei-Yin Loh [log in to unmask]
Department of Statistics phone: 608/262-2598
1210 West Dayton Street fax: 608/262-0032
University of Wisconsin http://www.stat.wisc.edu/~loh/
Madison, WI 53706
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
|