[FEA]: Improve cub::DeviceHistogram performance

### 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.

General tracking issue for various optimizations that impact or target the high level `cub::DeviceHistogram` on multiple ends. Optimizations could be in any level (block, warp etc.)

- [x] [Use atomicAdd_block in device histogram](#4973)
- [ ] Use warp atomic aggregates
- [ ] Experiment with saturation small-sized counters
- [x] Evaluate `__match_sync_any` to replace warp-level atomic operations (slower)

### Describe the solution you'd like

Overall performance improvement of `cub::*Histogram`

### Describe alternatives you've considered

_No response_

### Additional context

_No response_