Re: Tensor in GSL.

[Jordi Burguet Castell <jordi dot burguet-castell at cern dot ch>]

> I'm wondering if tensors (i.e. arrays with more than 2 dimensions) will
> be implemented in GSL or if someone is already working on this. I need
> high dimensional arrays for my own code, and I was considering to
> encode such a data structure. Perhaps it could be included in GSL when
> (and if) I get it working.

Hi,

I did a simple library in C++ to do simple tensor algebra. I also started an 
attempt to gslize it. You can see my post at the list archives, 2003-05-25. I 
think it is the one Brian refers to.

The C++ one worked fine, but for lack of knowledge about GSL & lack of time & 
maybe laziness, the gsl version was far from finished.

In my old post I described it a bit, but let me recall that it can do 
arbitrary tensors (any rank, any dimension, as far as there is memory) and 
simple algebra (tensor product, index contraction,...). Also, a drawback I 
can think of: there are no strides (maybe easy to add, though).

I was counting with this xmas to get it finished or at least usable for GSL, 
but if anyone gets something before based or not on my code, I will be very 
happy.

Jordi


PS: Sorry Andrea, you'll get this twice :)