[FEA]: Make `pathfinder` usable as a full replacement for `cuda_paths.py` in `numba-cuda`

[rwgk]

Support for locating these needs to be added:

rwgk-win11.localdomain:/usr/local/cuda-13.0 $ find . -name libdevice.10.bc -o -name nvdisasm -o -name cuobjdump -o -name libcudadevrt.a
./nvvm/libdevice/libdevice.10.bc
./targets/x86_64-linux/lib/libcudadevrt.a
./bin/cuobjdump
./bin/nvdisasm

rwgk-win11.localdomain:/usr/local/cuda-12.9 $ find . -name libdevice.10.bc -o -name nvdisasm -o -name cuobjdump -o -name libcudadevrt.a
./nvvm/libdevice/libdevice.10.bc
./targets/x86_64-linux/lib/libcudadevrt.a
./bin/cuobjdump
./bin/nvdisasm

Sub-issues: #762, #716

Also needed most likely: #759 — Support LoadedDL.from_distribution

This is what we want to replace:

https://github.com/NVIDIA/numba-cuda/blob/main/numba_cuda/numba/cuda/cuda_paths.py (URL for later when it is gone)

xref: https://github.com/NVIDIA/numba-cuda/issues/302

xref: https://github.com/NVIDIA/numba-cuda/pull/308 — very closely related (to this issue) exploratory work by @brandon-b-miller

[leofang]

Isn't this a duplicate of https://github.com/NVIDIA/cuda-python/issues/762 + https://github.com/NVIDIA/cuda-python/issues/716?

[rwgk]

Isn't this a duplicate of #762 + #716?

Thanks for pointing those out! I marked them as sub-issues of this one now.

[cpcloud]

Do we need duplicate issues tracking this work on both the numba-cuda and cuda-python trackers? It's really difficult to follow the issue true with this duplication. I'm going to close out the numba issue, since this issue is more complete than that one.

[leofang]

Works need to happen in both repos, so I am not sure what's the duplication?