Question about the cuda Driver api


How to understand the description in the CUDA Driver API manual?
It shows that “using the runtime API to work with it”, but cudaDeviceRest() function also the cuda runtime API.
So, I found confusing when reading this.