Geek Answers Handbook

Installing cuda via brew and dmg

After trying to install the nvidia toolkit on the MAC, following the manual: http://docs.nvidia.com/cuda/cuda-installation-guide-mac-os-x/index.html#axzz4FPTBCf7X I received the error "Package manifest analysis failed "which led me to this: NVidia CUDA toolkit 7.5.27 could not be installed on OS X. I turned off dmg, and the result was that instead of getting the "Parquet Parse Analysis Error", the installation program did not start (it seemed to start for a short time, and then exit).

Installing with the brew install Caskroom/cask/cuda ( CUDA 7.5 Mac installation does not have nvrtc ), it seems to have successfully installed cuda. A.

nvcc --version command returns:

 nvcc: NVIDIA (R) Cuda compiler driver Copyright (c) 2005-2015 NVIDIA Corporation Built on Mon_Apr_11_13:23:40_CDT_2016 Cuda compilation tools, release 7.5, V7.5.26

I built an example in /Developer/NVIDIA/CUDA-7.5/samples/1_Utilities with:

 make -C bandwidthTest/

It is executed without errors.

brew install Caskroom/cask/cuda installing with brew install Caskroom/cask/cuda be a safe installation method? What is the difference between this installation method and installation via a DMG file from nvidia?

Caskroom is a brew extension for installing GUI applications: https://github.com/caskroom/homebrew-cask

Should the IDE also be installed as part of the cuda installation?

+6
source share
2 answers

Both methods are downloaded and installed from the same .dmg file from NVidia.

The homebrew-cask structure is the preferred method for installing software distributed as binary files in the homegrown paradigm.

That is my understanding.

+4
source

Currently, to install cuda using brew, follow these steps:

 brew tap caskroom/drivers brew cask install nvidia-cuda

See https://github.com/caskroom/homebrew-cask/issues/38325 . Then you also need to add the following to your ~/.bash_profile file:

 export PATH=/Developer/NVIDIA/CUDA-9.0/bin${PATH:+:${PATH}} export DYLD_LIBRARY_PATH=/Developer/NVIDIA/CUDA-9.0/lib${DYLD_LIBRARY_PATH:+:${DYLD_LIBRARY_PATH}}

See http://docs.nvidia.com/cuda/cuda-installation-guide-mac-os-x/index.html .

UPDATE Newer versions of Mac OS X with activated SIP (System Integrity Protection) will prevent DYLD_LIBRARY_PATH from changing (see https://groups.google.com/forum/#!topic/caffe-users/waugt62RQMU ). You can check that through

 source ~/.bash_profile env | grep DYLD_LIBRARY_PATH

If the output of this command is empty, SIP is active, and you can disable it, as described in https://www.macworld.com/article/2986118/security/how-to-modify-system-integrity-protection-in-el- capitan.html . After that you should see

 env | grep DYLD_LIBRARY_PATH DYLD_LIBRARY_PATH=/Developer/NVIDIA/CUDA-9.0/lib
+4
source

More articles:


All Articles