I’ve just installed CuDNN on my Ubuntu 18.04.3.
A bit earlier I’ve installed successfully CUDA 10.1.
I followed step by step the Installation guide: https://docs.nvidia.com/deeplearning/sdk/cudnn-install/index.html.
2.4 Verifying says:
“To verify that cuDNN is installed and is running properly, compile the mnistCUDNN sample located in the /usr/src/cudnn_samples_v7 directory in the debian file.”
Question how can I access that directory “in the debian file”? Which debian file?
Apparently (please update the guide) I had to download also:
- cuDNN Runtime Library for Ubuntu18.04 (Deb)
- cuDNN Code Samples and User Guide for Ubuntu18.04 (Deb)
However, now when I try to run:
sudo dpkg -i libcudnn7_220.127.116.11-1+cuda10.1_amd64.deb
I get the following error:
(Reading database ... 240864 files and directories currently installed.) Preparing to unpack libcudnn7_18.104.22.168-1+cuda10.1_amd64.deb ... Unpacking libcudnn7 (22.214.171.124-1+cuda10.1) over (126.96.36.199-1+cuda10.1) ... Setting up libcudnn7 (188.8.131.52-1+cuda10.1) ... Processing triggers for libc-bin (2.27-3ubuntu1) ... /sbin/ldconfig.real: /usr/local/cuda-10.1/targets/x86_64-linux/lib/libcudnn.so.7 is not a symbolic link
2 more suggestions:
- Make the documentation open source, so that it could be actively and more accurately maintained
- Switch to a better forum, this one lacks a lot of features and is very much of an outdated style