Closed Eecornwell closed 3 months ago
@Eecornwell , this is my env: https://github.com/lfranke/TRIPS/issues/37#issuecomment-1977824525
and i notice you have added these environment vars with cudnn about path, you don't need to do it, the cmake will look for it auto when running.
Thank you @luoww1992 I matched my environment with yours and can now compile!
Hello, I encountered the same error log as you did. Even after using the conda environment like @luoww1992 , the error persists. I would like to ask if there are any other solutions available. Thank you.
Building your solution for Windows, here is my environment:
I have added these environment vars for: CUDNN = C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.8\lib\x64;C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.8\include;C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.8\bin; Path = C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.8\lib\x64;C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.8\include;C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.8\bin;
I also copied the cudnn files over to the Toolkit directory. It will error during
cmake --build build --config RelWithDebInfo -j
Here is the configure: