[FEA]: Enable Windows GPU Testing

### Is this a duplicate?

- [x] I confirmed there appear to be no [duplicate issues](https://github.com/NVIDIA/cccl/issues) for this request and that I agree to the [Code of Conduct](CODE_OF_CONDUCT.md)

### Area

Infrastructure

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

Today, CCCL only tests _builds_ in Windows environments. I would like CCCL tests to actually be _executed_ on a GPU in a Windows environment.

### Describe the solution you'd like

For the time being, we can use [GitHub's Windows-enabled GPU runners](https://docs.github.com/en/actions/concepts/runners/about-larger-runners#specifications-for-gpu-larger-runners).

The CCCL repo now has access to a runner group called `cuda-python-windows-gpu-github`. We should use the runners from this pool to enable nightly GPU testing of the following in priority order

- [ ] Enable nightly libcudacxx Windows testing
- [ ] Enable nightly CUB Windows testing
- [ ] Enable nightly Thrust Windows testing
- [ ] Enable nightly cuda.cccl Python Windows testing

Here is an example from the `cuda-python` repo of using these runners: https://github.com/NVIDIA/cuda-python/blob/ed12c8301c6f9b23e6db9829e66e4ec745a76a7a/.github/workflows/test-wheel-windows.yml#L77-L85

### Describe alternatives you've considered

_No response_

### Additional context

_No response_