Is this a duplicate?

• I confirmed there appear to be no duplicate issues for this request and that I agree to the Code of Conduct

Area

cuda.core

Is your feature request related to a problem? Please describe.

Both cupy and pytorch define __hash__() methods which use the pointer address (and sometimes device index) of the underlying CUDA API structure such two Python level objects which represent the same underlying CUDA API structure will hash the same.

https://github.com/pytorch/pytorch/blob/v2.7.0/torch/cuda/streams.py#L14
https://github.com/cupy/cupy/blob/v13.4.1/cupy/cuda/stream.pyx

Describe the solution you'd like

cuda.core should also implement __hash__() such that caching and comparison are easier to implement.

Describe alternatives you've considered

No response

Additional context

No response