is a GPU-accelerated library for linear algebra that provides iterative solvers for sparse systems

Offering a capability set that is unmatched in a ready-to-integrate library

With many solvers and preconditioners to choose from, you have the power to choose the solution that is best for you. Best of all, it is easy to use, with no complicated callback interfaces and a straightforward configuration.

Solvers Preconditioners Data Formats
Biconjugate Gradient (BICG) Jacobi Double-precision real and complex
Biconjugate Gradient Stabilized (BICGSTAB) Block Jacobi Compressed Sparse Row (CSR)
Conjugate Gradient (CG) Incomplete LU (Ilu0) Compressed Sparse Column (CSC)
Generalized Minimum Residual (GMRES) Reordered (Ilu0) Coordinate (COO)
Minimum Residual (MINRES)    

With performance that is 10x faster than competing solutions

We are experts at getting the very best performance available. We combine powerful GPU-acceleration with capable numerical algorithms to yield solutions that reach levels of performance not previously possible.

And available in whatever form you need it.

CULA Sparse provides C, C++, and Fortran interfaces so that you can work in the language that is easiest for you. And with support for all standard CUDA platforms; Linux 32/64, Win 32/64, OSX, you can be assured that your solution runs wherever you need it to.

More Information

» Review the programmer's guide
» Downloads
» Visit the forums