[Discuss] FORTRAN -> ???

[Jerry Feldman]

On 01/25/2012 10:48 AM, Jay Kramer wrote:
> The Fortran IV program that I am referring to was one of the first 3D
> finite element programs.  It ran on a Digital and/or Cray
> supercomputer.  It was back in the early 80's. It took hours and cost
> thousands of dollars to perform one analysis consisting of about 500
> 2D elements (I can not remember the amount of CPU time).  The original
> software that I modified was called SAP4.  It was created at UC
> Berkley and was open source.  I spent years modifying the code to suit
> my needs in geomechanical engineering.  Those were the good old days
> when programs were written so efficiently that they were works of art.
While you should be able to get a halfway decent C implementation using
f2c, by reworking it in C++, you can re-engineer the code to be much
more maintainable. If I recall, FORTRAN IV had very small variable
names, where most modern languages have larger names that you can make
descriptive, the result of recoding and parallel testing on Linux might
work better. In addition, FORTRAN was very von Neumannish where modern C
or C++ (or Java) can use threads to perform analysis in parallel.

-- 
Jerry Feldman <gaf at blu.org>
Boston Linux and Unix
PGP key id:3BC1EB90 
PGP Key fingerprint: 49E2 C52A FC5A A31F 8D66  C0AF 7CEA 30FC 3BC1 EB90