I have followed the instructions here https://docs.nvidia.com/cuda/cuda-installation-guide-mac-os-x/index.html to install CUDA on a MacBookPro.
I am also able to build the examples. When running the examples, CUDA is not able to find a GPU, but reports “cudaGetDeviceCount returned 100” instead.
➜ release ./deviceQuery ./deviceQuery Starting... CUDA Device Query (Runtime API) version (CUDART static linking) cudaGetDeviceCount returned 100 -> no CUDA-capable device is detected Result = FAIL ➜ release
Where should I look to trouble shoot this issue?
My setup is:
MacBookPro Mid 2012 Retina
OSX High Sierra, 10.13.6
NVIDIA GeForce GT 650M 1024 MB
CUDA Driver Version: 418.163
CUDA 10.1.105 Toolkit
- Installation as per this page: https://docs.nvidia.com/cuda/cuda-installation-guide-mac-os-x/index.html
- An OSX related (apparently) tweak as described here, post #3 onwards: https://devtalk.nvidia.com/default/topic/1048339/cuda-setup-and-installation/cuda-driver-410-130-error-quot-update-required-quot-mac-os-10-11-6-/
- System Preferences -> Energy Saver -> Automatic graphics switching: unchecked
- kernel extension reports started:
➜ release kextstat | grep -i cuda 169 0 0xffffff7f83b66000 0x2000 0x2000 com.nvidia.CUDA (1.1.0) 50F4AE08-3D20-3B13-B36D-439DEEB1D49C <4 1> ➜ release
I appreciate any ideas, insight or help in finding steps forward,