🧬 Staged movegen [WIP]

[eduherminio]

Trying to build on top of #810

Initial implementation, where an IEnumerable that marks the end of the current movegen stage is used and an array is passed (beware, allocs!)

Perft and all tests passing, but big elo loss. Different bench.

Test  | movegen/staged
Elo   | -58.85 +- 14.83 (95%)
SPRT  | 8.0+0.08s Threads=1 Hash=32MB
LLR   | -2.30 (-2.25, 2.89) [0.00, 3.00]
Games | 1192: +278 -478 =436
Penta | [80, 184, 211, 98, 23]
https://openbench.lynx-chess.com/test/449/

Generate captires before castling moves

Test  | movegen/staged
Elo   | -65.25 +- 15.76 (95%)
SPRT  | 8.0+0.08s Threads=1 Hash=32MB
LLR   | -2.33 (-2.25, 2.89) [0.00, 3.00]
Games | 1142: +261 -473 =408
Penta | [86, 182, 186, 92, 25]
https://openbench.lynx-chess.com/test/451/

Generate all the moves at the same time but return them in stages

Test  | movegen/staged
Elo   | -52.54 +- 14.14 (95%)
SPRT  | 8.0+0.08s Threads=1 Hash=32MB
LLR   | -2.27 (-2.25, 2.89) [0.00, 3.00]
Games | 1306: +312 -508 =486
Penta | [85, 190, 243, 106, 29]
https://openbench.lynx-chess.com/test/452/