NEC-LIST: Fortran execution speed

I would like to echo the comments of wadavis on his post about the
execution methods of NEC under Windows. I had a program where I needed
to model many thousands of segments. I used GNEC (a very good
commercial version of NEC-4) under Widows XP. Due to the memory
allocation limits that wadavis mentioned, my models failed at about
12,000 segments.

I abandoned that platform and compiled the NEC-4 code from source using
g77 on a Macintosh running the MacOS X (Unix-based) system. Using the
resources posted on this list by John B. Wood 03/04/04:

> Hello. I ran a thread on this subject a while back. NEC 4.1 can be
> somewhat optimized to run on a Linux box by compiling with g77 as
> follows:
>
> g77 nec41.f -O3 -fno-automatic
>
> where nec41.f is the NEC 4.1 f77 source code file. The default
> executable filename will be nec41.o. The above gave me results
> consistent with those I had obained on an old HPUX workstation. You
> might also try
>
> g77 -O3 -fno-emulate-complex -funroll-loops nec41.f
>
> This version of the compile was published in the Nov. '98 ACES
> newletter but I never could get the executable to run reliably on the
> laptop I was using (Panasonic Toughbook CF-27). Sincerely,
>
At that time I used both of the above proceedures to compile NEC-4 on
Mac OS X and got good results for double-precision models of about
16,000 wire segments. Later, I recompiled for 30,000 segments and got
what seemed to be good results for my model.

In my opinion, unless you absolutely must use Windows (especially for
large models) a good version of Linux or MacOS X probably would be a
better choice of platform. I do not know if the "Cygwin" environment
mentioned by john.kot has avoided the Windows memory-limit factor. If
anyone knows this, it would beneficial.

Hope this helps.

-- 
The NEC-List mailing list
NEC-List_at_robomod.net
http://www.robomod.net/mailman/listinfo/nec-list