Re: search of a code in C

[John Turner]

>>>>> "JWE" == John W Eaton <address@hidden> writes:

JWE> Too bad there is no standard for matrix math libraries in C++.

Indeed.  That's one place where F9x wins *big* time.  Arrays as
1st-class objects.  Ahhh...

Back to C++... others mentioned newmat and TNT.  We've also used TNT's
predecessor, MV++, as well as SparseLib++ and IML++ (why does
everything written in C++ have to end with "++"?), which are built
on MV++.  See:

  http://math.nist.gov/mv++/
  http://math.nist.gov/sparselib/
  http://math.nist.gov/iml++/

but they are somewhat spotty.  A decent start, but it looks like they
were dropped before maturity in favor of TNT.

There's also A++/P++:

  http://www.c3.lanl.gov/~dquinlan/A++P++.html

which is a monstrous (over a million lines of code, last I spoke with
Dan about it), but mature and solid.  Really strives to provide
F9x'ish semantics and capabilities, in serial and parallel, plus
more.  Nice, but heavy.

I only needed something fairly simple, so I just chose MV++ and
extended it as needed.  But I imagine I'll move to something else at
some point.

-- 
John A. Turner, Ph.D.                Senior Research Associate
Blue Sky | VIFX                      http://www.bluesky-vifx.com/
One South Road, Harrison, NY 10528   http://www.lanl.gov/home/turner/
Phone: 914-381-8400                  Fax: 914-381-9790/1