How to prevent GPU or cuDNN from firing up during inference

Hi I have been testing two computers with the same CPU speed (3GHz). One of them has a GPU card and the same exact model (on same version of python, pymc3 and mkl) runs a few times slower. I have tried to turn off GPU by

import os
os.environ[‘THEANO_FLAGS’] =‘mode=FAST_RUN,floatX=float32,device=cpu’

but it appears that cuDNN is still running because I get
lib/python3.7/site-packages/theano/gpuarray/ UserWarning: Your cuDNN version is more recent than Theano. If you encounter problems, try updating Theano or downgrading cuDNN to a version >= v5 and <= v7.
warnings.warn("Your cuDNN version is more recent than "

Any ideas? Thanks!