Better check evasion move sorting. #1685
Conversation
|
Congratulations! |
| @@ -73,8 +73,8 @@ MovePicker::MovePicker(const Position& p, Move ttm, Depth d, const ButterflyHist | |||
|
|
|||
| /// MovePicker constructor for quiescence search | |||
| MovePicker::MovePicker(const Position& p, Move ttm, Depth d, const ButterflyHistory* mh, | |||
| const CapturePieceToHistory* cph, Square rs) | |||
| : pos(p), mainHistory(mh), captureHistory(cph), recaptureSquare(rs), depth(d) { | |||
| const CapturePieceToHistory* cph, const PieceToHistory** ch, Square rs) | |||
There was a problem hiding this comment.
I think you can remove the double indirection here (const PieceToHistory** -> const PieceToHistory*) and just pass (ss-1)->contHistory to the constructor below (only the first member of the array is used by MovePicker). Or am I missing something?
There was a problem hiding this comment.
Yes that would be possible. But in the Movepicker object only a pointer to the array of the counter move histories is currently used (const PieceToHistory** contHistory;). There itself is no array of the four cmh pointers defined.
We would have to define it there like PieceToHistory* contHistory[4]; and modify both constructors. That seems not so elegant for me (but is debatable).
But an important reason is that we could easy test using the other histories (which i have already started)
There was a problem hiding this comment.
Oh, now I see that I missed that other constructors use the same pattern already and that in other places more elements of the array are used. You can disregard my previous comment then :)
There was a problem hiding this comment.
I have looked at this more and we can simply fake the array by using in the second Movepicker constructor contHistory(&ch) in the initiliazer list (so only the first index exists but the others are not used). But i would rewrite this only after a period where we have tested out the possibilities of the other histories.
Thanks for the inspiration!
|
We should improve the bench extraction for travis and appveyor checking because if a number other than from bench is in the commit message this goes wrong. So i have added new commits with a correct parseable bench. |
locutus2
changed the title
|
@candirufish |
|
Merged via 12e79be, congrats! :-) I have renamed "contHistory" to "continuationHistory" to ease the newcomers understanding of this part of the codebase. And I also agree that time is up for a regression test :-) |
Use in addition the counter move history table for sorting quiet check evasion moves in main and quiecence search. Also rename "contHistory" to "continuationHistory" while there. STC: LLR: 2.96 (-2.94,2.94) [0.00,5.00] Total: 73284 W: 16433 L: 15938 D: 40913 http://tests.stockfishchess.org/tests/view/5b4f526e0ebc5902bdb7a401 LTC: LLR: 2.96 (-2.94,2.94) [0.00,5.00] Total: 12135 W: 2171 L: 1997 D: 7967 http://tests.stockfishchess.org/tests/view/5b4fc0ef0ebc5902bdb7ae0e Closes official-stockfish#1685 Bench 4817583
Use in addition the counter move history table for sorting quiet check evasion moves in main and quience search.
STC
LLR: 2.96 (-2.94,2.94) [0.00,5.00]
Total: 73284 W: 16433 L: 15938 D: 40913
http://tests.stockfishchess.org/tests/view/5b4f526e0ebc5902bdb7a401
LTC
LLR: 2.96 (-2.94,2.94) [0.00,5.00]
Total: 12135 W: 2171 L: 1997 D: 7967
http://tests.stockfishchess.org/tests/view/5b4fc0ef0ebc5902bdb7ae0e
Bench 4817583