[FEA]: DeviceReduce argument version with user-defined reduction

### 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

CUB

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

I encountered this problem:
Given an unsigned integer array, find if there's any "1" bits, and if so, find the first location.
I can use DeviceReduce::Max to check if there's any "1"s, but I'm not sure how to do the other.
One way is to create a new array and copy the values as 0 if it's 0 and 1 if it's non-zero, then do an ArgMax to find the index, then copy that element in the original array to the host and find the bit, but it's kind of awkward. Or just copy the entire array to host and do it on CPU.

### Describe the solution you'd like

If there is an argument reduction with user-defined reduction, this would be much easier.

### Describe alternatives you've considered

_No response_

### Additional context

_No response_