help-gsl
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

[Help-gsl] Re: C++ wrapper for linear algebra


From: Rodney Sparapani
Subject: [Help-gsl] Re: C++ wrapper for linear algebra
Date: Mon, 05 Nov 2007 10:09:43 -0600
User-agent: Thunderbird 2.0.0.6 (X11/20070802)

Hasan Hasanov wrote:
Hello,

I am looking for a C++ wrapper for the GSL. I will be using lots of linear
algebra, and I am looking for a wrapper that will have operator overlaoding
for matrix-matrix and matrix-vector multiplication/addition in the very
least. It is also very important that it implements an efficient matrix
inversion operator.

So far all I could find on the web are:
1) http://gslwrap.sourceforge.net
2) http://cholm.web.cern.ch/cholm/misc/#gslmm

Which one of these would you recommend? Are there any wrappers that are
actually part of the GSL project itself?

I have already implemented my project in Matlab, and I will be translating
it into C++.  Since Octave is the GNU analogue of Matlab I am wondering if
Octave has an C++ API for matrix operations that would be an alternative to
the above two wrappers?


Regards,
Hasan

Hi Hasan:

I worked on the original version of GSLwrap. However, it has changed quite a bit. I have also ported my pre-GSL Vector/Matrix class library to use GSL (C++ abstraction is beautiful). I haven't gotten around to submitting it as a GSL add-on. But, it only deals with linear algebra and pseudo RNG, so it's quite limited.

Why don't you give those 2 projects a try first? If you then find you still want more, then drop me a line and I can post some of my code. However, be warned, it is not documented. But, most of the overloadings are quite short: 5 lines or less.

Oh, I'm not much of a Matlab user (I have used Scilab occasionally), but what do the vector/matrix indices start with 0 or 1 (I forget)? My class library supports both which I really like. I personally use 1 since I have spent so much time with other software that also uses 1
(and of course the mathematics also uses 1, another reason to use C++
abstraction :o).

Thanks,

Rodney





reply via email to

[Prev in Thread] Current Thread [Next in Thread]