Getting error for cudNN in MATLAB 2019

AI & Data Science Deep Learning (Training & Inference) cuDNN
1

I installed CUDA 9.2 for NVIDIA RTX 2060 along with cudNN. When tried to check the GPUinstaller in MATLAB 2019, got the below error:

Compatible GPU : PASSED
CUDA Environment : PASSED
Runtime : PASSED
cuFFT : PASSED
cuSOLVER : PASSED
cuBLAS : PASSED
cuDNN Environment : FAILED (Unable to find the ‘NVIDIA_CUDNN’ environment variable. Set ‘NVIDIA_CUDNN’ to point to the root directory of a NVIDIA cuDNN installation.)
TensorRT Environment : FAILED (Unable to find the ‘NVIDIA_TENSORRT’ environment variable. Set ‘NVIDIA_TENSORRT’ to point to the root directory of a TensorRT installation.)
Profiling Environment : PASSED
Basic Code Generation : FAILED (Test GPU code generation failed with the error ‘emlc:compilationError’. View report for further information: View report)

Need help to resolve the issue.

2

Hi,

Could you please check if all the environment variables are set correctly?
Please refer below link for more details:

https://in.mathworks.com/help/gpucoder/gs/setting-up-the-toolchain.html

Thanks

3

Hi,

I have installed CUDA v9.2 and in the root directory of CUDA, there is no cuDNN folder. I downloaded cudNN and after unzipping I found 3 folders named bin, include and lib and each of them containing one files. As per the instruction in NVIDIA, I copied and pasted these 3 files in the requite locations.

CUDA toolkit path in my case is given below:

C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v9.2\

The downloaded cudNN downloaded file named cudnn-9.2-windows10-x64-v7.6.5.32. Under this, CUDA named folder is there and within that bin, include and lib are located. If I modify this CUDA folder to cuDNN and paste in the CUDA toolkit root directory, then also it is not working.

1 Like