CULA Sparse Beta 2 is undergoing final packaging and testing to be sent out to our Beta testers very soon. This is a Feature update, with the following changes:
- Added the BiCGSTAB solver
- Added the BiCGSTAB(L) solver
- Complex (Z) data types available for all solvers
- Fortran module added
- Configuration parameter to return best experienced solution
- Maximum runtime configuration parameter
- New example for Fortran interface
- New example for MatrixMarket data
- Several important bug fixes, as noted by the Beta testers
This release also contains the first steps towards interoperability with CULA for dense linear algebra, which some hybrid methods require. A user will now need to link cula_core and cula_sparse rather than just the sparse lib. Full interoperability will require CULA R13, which is also coming soon.
We are still accepting Sparse Beta applications, so register soon!