Improve performance of select-top-k WARP_SORT implementation (#606)

A few simplifications and tricks to improve the performance of the kernel:

- Promote some constants to static constexpr
- Allow `capacity < WarpSize`
- Reduce the frequency of `sort` operations for `filtered` version
- Remove `warp_sort::load` to simplify the api and implementation

Authors:
  - Artem M. Chirkin (https://github.com/achirkin)

Approvers:
  - Tamas Bela Feher (https://github.com/tfeher)

URL: #606