getkeops / keops

KErnel OPerationS, on CPUs and GPUs, with autodiff and without memory overflows
https://www.kernel-operations.io
MIT License
1.03k stars 65 forks source link

Update installation.rst -- troubleshoot missing cuda.h #315

Closed rubenalv closed 11 months ago

rubenalv commented 1 year ago

Added a note that when running pykeops from a conda enviroment, pykeops.test_torch_bindings() will fail due to cuda.h missing unless a system-wide installation of CUDA toolkit is available. I included the link to the NVIDIA FAQ that explains that the conda-forge install of CUDA toolkit does NOT contain the full toolchain with the compilers. https://conda-forge.org/docs/user/faq.html?highlight=cuda, question "How can I compile CUDA (host or device) codes in my environment?"