The compute unified device architecture (cuda) enables nvidia graphics processing units (gpus) to be used for massively parallel general purpose computation.

this package contains the cuda sparse matrix runtime library.