Is there an efficient solution implementation for a sparse matrix linear equation using CUDA?

Question

Is there any effective tool for solving a linear sparse matrix equation using CUDA?

Answer 1

Like other answers, you can also check out cusp-library

Answer 2

With the new 3.2 Toolkit (I think it was released on 2010-09-22, although RC is currently available). NVIDIA providing CUSPARSE, the Solver Matrix library. This will be used in the same way as CUFFT and other libraries from NVIDIA, and full documentation is available on their web page.

Answer 3

NVidia .

: CUSPARSE NVIDIA CUDA Toolkit 3.2 / .

Answer 4

you can try openNL , there is also a commercial cula package that has some basic materials for free. there is no magic created for sparse matrix solvers, it depends a lot on the sparseness matrix template and how you want to use it, do you have a solution? matrix dose remains unchanged? In my case, we ended up writing our own solver for the reasons mentioned above.