Re: NEC-LIST: Portland Group's F77 and HPF compilers for Linux

From: David Gaudine <david_at_email.domain.hidden>
Date: Fri, 30 Jan 1998 10:12:40 -0500 (EST)

On Mon, 26 Jan 1998, David Gaudine wrote:

> Has anyone tried compiling any version of NEC with the Portland
> Group's F77 or HPF (F90) compilers for Linux? With or without
> success? I get long execution times and meaningless results, so I'd
> like to know if it's just me before I discuss it with the company.

On Wed, 28 Jan 1998, Pat Foster wrote:

> I have compiled NEC on LINUX using G77. This gave the same runtimes
> and results (correct) as NEC compiled using DECFortran for NT. We have
> used G77 on LINUX for a number of programs and have had no problems
> whatever.

I get rather slow results using g77, compared to Digital Visual
FORTRAN on Windows 95. Also, I get correct results with most data
files, but not all. The Digital compilers, both for the PC and my
DECStation, handle all my files perfectly. I suppose I'll have to dig
deeper though, saying "it must be a bug in the compiler" sounds rather
hollow when I try to say it about two different compilers.

On Thu, 29 Jan 1998, Steve Inge wrote:

> I have used the FORTRAN 77 that is available for Hewlett Packard G50
> machines using HP-UX. I can't remember the program name, but I have
> successfully compiled NEC on it.
>
> I did have a spot of difficulty, similar to above, but specifying a
> compiler switch which made all subroutine data static fixed it. Goes
> well now.

I tried this compiler's -norecursive option, which is similar to
-static or -fnoautomatic on some systems, no luck.
Received on Sat Jan 31 1998 - 11:47:16 EST

This archive was generated by hypermail 2.2.0 : Sat Oct 02 2010 - 00:10:38 EDT