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 runtime api library for high-level cuda programming, on top of the cuda driver api.