Closed flaport closed 1 year ago
The line
CMake Error at /tmp/pip-build-env-92yt10qq/overlay/lib/python3.10/site-packages/cmake/data/share/cmake-3.26/Modules/CMakeDetermineCompilerId.cmake:751 (message):
Compiling the CUDA compiler identification source file
"CMakeCUDACompilerId.cu" failed.
makes me think this is CUDA related.
I have actually been using CUDA 11.8 to build things, would you be able to try that? Please let me know whether it works or not.
I use the following to install it (on Ubuntu 22.04):
wget https://developer.download.nvidia.com/compute/cuda/11.8.0/local_installers/cuda_11.8.0_520.61.05_linux.run
sudo sh cuda_11.8.0_520.61.05_linux.run
Unfortunately that did not seem to solve it. I'll try a few more things and see if I can make it work. I'll report my findings afterwards
Thanks. Any additional information on your system, OS, etc... may be helpful (and of course the demo works in the colab notebook)
On Fri, May 5, 2023 at 3:53 PM Floris Laporte @.***> wrote:
Unfortunately that did not seem to solve it. I'll try a few more things and see if I can make it work. I'll report my findings afterwards
— Reply to this email directly, view it on GitHub https://github.com/spinsphotonics/fdtdz/issues/9#issuecomment-1536868894, or unsubscribe https://github.com/notifications/unsubscribe-auth/A3VGPQMUYJ3SS6PWTV4QQ7TXEWAFTANCNFSM6AAAAAAXWUJDJE . You are receiving this because you commented.Message ID: @.***>
Not sure if its the same error, but in my case (on Ubuntu 20.04) I had to manually install JAX with CUDA first.
From: https://github.com/google/jax#installation
pip install --upgrade pip
# CUDA 12 installation
# Note: wheels only available on linux.
pip install --upgrade "jax[cuda12_pip]" -f https://storage.googleapis.com/jax-releases/jax_cuda_releases.html
# CUDA 11 installation
# Note: wheels only available on linux.
pip install --upgrade "jax[cuda11_pip]" -f https://storage.googleapis.com/jax-releases/jax_cuda_releases.html
@danielpeace Thanks for mentioning this. Yes, I think this could be the issue as well. In my experience with installing JAX, the sticking point is often correctly installing the CUDNN package. Since fdtd-z is explicitly designed to run on the GPU, this is a must.
@flaport Please let me know if installing JAX with CUDNN solved this for you!
Installing jax with cuda support indeed solved the issue! Thanks @danielpeace :)
Hi, Thanks for this exciting new library! I'd love to give it a try!
However, I was trying to build the library but i get the following errors:
Any idea on how to solve this? Thanks