[Beowulf] NCSU and FORTRAN

[Geoff Jacobs]

Wallace Pitts wrote:
> Robert;
> 
> My wife is a biomathematics student at NCSU.  She is currently working
> on a Markov chain simulation using MATLAB.  The goal is to use some sort
> of search routine to find a set of transition matrix parameters that
> minimize the sum of squares.  The problem simulated say 10,000 molecules
> for say 400 states.  The search process is slow as mollassesb (we use
> routines in the optimization toolbox as well as the Genetic algorithm
> and direct search toolbox).
>
> I used FORTRAN in a former life, and still have a compiler or 2.  In
> addition I have several optimization routines lifted  from "Optimization
> with FORTRAN.
> 
> Any suggestions as to compilers (have g77).
gfortran is a vast improvement on g77. On the commercial front, all my
experience is with the stuff from Portland Group, which is good.

> I have looked at the
> propaganda on Intel's website, and they claim that their Visual FORTRAN
> will handle multi core processors.
Do they mean that they support OpenMP? This is hardly remarkable in a
commercial product.

> Regards,
> 
> Wallace Pitts.

-- 
Geoffrey D. Jacobs

Go to the Chinese Restaurant,
Order the Special