[Haskell-cafe] Is anyone working on a sparse matrix library in Haskell?
carter.schonwald at gmail.com
Fri Nov 30 23:12:12 CET 2012
I look forward to see what comes of this!
On Fri, Nov 30, 2012 at 4:58 PM, Mark Flamer <mark at flamerassoc.com> wrote:
> Thanks for all the replies,
> It sounds like there is enough interest and even some potential
> collaborators out there. I have created a few data structures to represent
> sparse vectors and matrices. The vector was a simple binary tree and the
> matrix a quad tree. As I suspected a standard IntMap was around 3X as fast
> as my binary tree, so I have switched to the IntMap for now. I was hoping
> hold on to my quad tree for the matrix rep. because I like the elegance of
> the recursive algorithms like Strassen’s for multiplication. In the end it
> will most likely be best to have a few different matrix representations
> anyway. For instance, I know that compressed row form is most efficient for
> certain algorithms. I have a Gauss iterative solver working currently and
> thinking of moving on to a parallel Conjugate gradient or direct solver
> using LU decomposition. I’m no expert in Haskell or numeric methods but I
> my best. I’ll clean up what I have and open up the project on Github or
> Bitbucket. Any other thoughts or ideas are welcome.
> I’m currently using the Matrix Market files for testing. It would be nice
> benchmark this against an industry standard solver in C or Fortran, without
> having to do a lot of work to set it up. Does anyone know of an easy way to
> do this? I’m just trying to get results to compare in orders of magnitude
> for now. It would be motivating to see these comparisons.
> View this message in context:
> Sent from the Haskell - Haskell-Cafe mailing list archive at Nabble.com.
> Haskell-Cafe mailing list
> Haskell-Cafe at haskell.org
-------------- next part --------------
An HTML attachment was scrubbed...
More information about the Haskell-Cafe