Where is cuOccupancyMaxActiveBlocksPerMultiprocessor defined?

Accelerated Computing CUDA CUDA Programming and Performance
#1

Hello, I’m trying to play with the occupancy. Unforunately I cannot find in which library/DLL the method is defined.

Thanks
Martin

#2

Start by looking in the cuda documentation:

http://docs.nvidia.com

Functions that begin with cu are part of the driver API. Functions that begin with cuda are part of the runtime API. So for this particular function, look in the driver API documentation in the occupancy section.

If you want to learn how to build a driver API app, follow the example of one of the cuda sample codes, which include make files and project files, such as vectrAdddrv

#3

CUDA Driver API
cuOccupancyMaxActiveBlocksPerMultiprocessor
http://docs.nvidia.com/cuda/cuda-driver-api/group__CUDA__OCCUPANCY.html#group__CUDA__OCCUPANCY

CUDA Runtime API
cudaOccupancyMaxActiveBlocksPerMultiprocessor
http://docs.nvidia.com/cuda/cuda-runtime-api/group__CUDART__OCCUPANCY.html#group__CUDART__OCCUPANCY

#4

OK. Thanks. I think there has been some issue with my CUDA libraries. After the update I could load the library and find the method.