This code is used to test the sampling of pseudo random fields
on different architectures.  It currently supports different IBM,
SGI, LINUX DEC, and CRAY fft libraries.

The algorithm is the one derived in Evensen (1994, 2003, 2004) for
the generation of the pseudo random fields and the improved sampling.

Note that for IBM the ESSL library is used.

For LINUX we have decided to use the FFTW3 library which is 
available for many unix systems from http://www.fftw.org/

When testing, do not set nre too large since this will 
really slow down the production of fields. Try with 1, 2, 4, and
maybe 6 or 8 for smaller 2D grids.

Before compiling define your achitecture in the file MODEL.CPP

You are probably using the mkdepend.pl script found elsewhere on 
this site.  Note that on Linux the path to cpp needs to be changed.

This code is developed by Geir.Evensen@hydro.com and/or Geir.Evensen@nersc.no
The code can be freely used but improvements, upgrades and bugfixes
should be reported,  and as everything free, it comes with no guarantee.