[FEA]: Validate cuda.parallel type matching in build and execution

[gevtushenko]

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

General CCCL

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

Today, cuda.parallel API consists of build and execution stages. During the build stage, we instantiate C++ templates and form a cubin. During execution stage, we use this cubin. There's no type validation during execution step. This might create issues where user build cuda.parallel algorithm for, say, int, but execute it on double leading to a silent corruption.

Describe the solution you'd like

We should validate that types used during execution step match ones provided in build here:

cccl/python/cuda_parallel/cuda/parallel/experimental/__init__.py

Line 215 in 4088134

# TODO Assert that types match the ones used in the constructor

This PR can be closed by a test illustrating that an exception is thrown when passing not matching types during execution step.

Describe alternatives you've considered

No response

Additional context

No response