This is the mail archive of the
cygwin
mailing list for the Cygwin project.
Re: g77 on cygwin64
- From: Marco Atzeri <marco dot atzeri at gmail dot com>
- To: cygwin at cygwin dot com
- Date: Wed, 12 Feb 2014 20:10:56 +0100
- Subject: Re: g77 on cygwin64
- Authentication-results: sourceware.org; auth=none
- References: <52FAF082 dot 6020805 at appstate dot edu> <52FB0686 dot 8080403 at gmail dot com> <52FB13E2 dot 1070107 at appstate dot edu> <724410821 dot 20140212111451 at mtu-net dot ru> <CAMCbSMoqb_TsY3BQSU-AwnCOuCYecwCw4YexBQuUhbAbpSNDQg at mail dot gmail dot com> <52FB8E3C dot 3030704 at appstate dot edu> <52FBAC0B dot 8000106 at gmail dot com> <52FBB0F8 dot 80606 at molconn dot com>
On 12/02/2014 18:35, LMH wrote:
> Coming back to your issue, it is really curious that gfortran compiles
> the code but behaves wrongly, I was expecting it rejects the code
> if not in line with latest standard.
This happens on some of my code as well and is in many ways a far worse
issue to deal with. Some of my code compiles and runs, but the floating
point answers to many of the calculations are wrong. This makes the
issue(s) very difficult to run down (and also makes me stick with gcc3
like the posters colleagues).
LMH
Could you share a test case ?
I found no difference in math behaviour on all the Fortran
codes in LAPACK and OCTAVE when we moved from G77 to GFORTRAN
I am really curious
Regards
Marco
--
Problem reports: http://cygwin.com/problems.html
FAQ: http://cygwin.com/faq/
Documentation: http://cygwin.com/docs.html
Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple