GPU Glossary
/
Host Software·1 min read

CUDA Runtime API

A higher-level, closed-source API that wraps the CUDA Driver API and is the usual way applications use CUDA.

The CUDA Runtime API wraps the CUDA Driver API and exposes the same functionality at a higher level.

The CUDA Toolkit. The CUDA Runtime API wraps the CUDA Driver API to make it more amenable to application programming. Adapted from the Professional CUDA C Programming Guide.
The CUDA Toolkit. The CUDA Runtime API wraps the CUDA Driver API to make it more amenable to application programming. Adapted from the Professional CUDA C Programming Guide.

It is usually preferred over the Driver API because it is easier to use. The trade-off is slightly less control over kernel launches and context management. NVIDIA explains the differences in this section of the Runtime API docs.

Linking#

The Runtime API may be linked statically, per Attachment A of the NVIDIA CUDA Toolkit EULA, but it doesn't have to be. On Linux, the shared object for dynamic linking is usually named libcudart.so.

The Runtime API is closed source. Its documentation is here.