Hello, I have an error while importing Theano when using the CVMFS stack on the helios cluster. I want to use this stack as it is the only way to have MAGMA on the Helios cluster. The error is :
pygpu.gpuarray.GpuArrayException: b'Could not load "libnvrtc.so": > libnvrtc.so: cannot open shared object file: No such file or directory' > The stack loads the following modules (this is the output of 'module list'): Currently Loaded Modules: 1) nixpkgs/16.09 (S) 2) icc/.2016.4.258 (H) 3) gcccore/.5.4.0 (H) 4) ifort/.2016.4.258 (H) 5) intel/2016.4 (t) 6) imkl/11.3.4.258 (math) 7) StdEnv/2016.4 (S) 8) cuda/8.0.44 (t) 9) openmpi/2.1.1 (m) Where: S: Module is Sticky, requires --force to unload or purge m: MPI implementations / Implémentations MPI math: Mathematical libraries / Bibliothèques mathématiques t: Tools for development / Outils de développement H: Hidden Module And my .theanorc is the following : [global] device = cuda floatX = float32 intX = int32 optimizer_including = cudnn [dnn] base_path = /cvmfs/soft.computecanada.ca/easybuild/software/2017/avx/Compiler/intel2016.4/cuda/8.0.44 [gpuarray] preallocate=0.9 [blas] ldflags = -lmkl -lguide -lpthread The base_path is the value of the environment variable $CUDA_PATH which is automatically set when loading any cuda module. I tried to ask for help to Calcul Québec but they could not help me with Theano. Thank you ! -- --- You received this message because you are subscribed to the Google Groups "theano-users" group. To unsubscribe from this group and stop receiving emails from it, send an email to [email protected]. For more options, visit https://groups.google.com/d/optout.
