Investigate how to retain WGPU and also using upstream cuda oxide: https://github.com/NVlabs/cuda-oxide
Investigate how to retain WGPU and also using upstream cuda oxide: https://github.com/NVlabs/cuda-oxide