Hello,
I completely rewrote my old chess engine squared-chess from scratch, using pieces of it, so it got a new name.
- move generation is now staged
- transposition table uses two separate tables: always replace and a selective replace
- PVS search is now implemented
- aspiration window
- a million bugs got removed
- code looks nicer
- option: adjustable linear skill levels 0-99
- option: randomized search
- nicer logo
- nicer name
- can now also easily be compiled with Mircosoft Visual Studio (and with a makefile on Linux)
Elo improvement compared to squared-chess 1.3.0 is probably between 80 and 160 (~2150 CCRL 40/4).
Googleplex Starthinker 1.3
"Googleplex Starthinker" chess engine
Moderator: Ras
-
tsoj
- Posts: 35
- Joined: Thu Oct 19, 2017 4:59 pm
- Location: Germany, Berlin
- Full name: Jost Triller
-
Guenther
- Posts: 4718
- Joined: Wed Oct 01, 2008 6:33 am
- Location: Regensburg, Germany
- Full name: Guenther Simon
Re: "Googleplex Starthinker" chess engine
Nice improvement Jost, still I dislike the name of the new program or let's say the 'commercial part' of it.tsoj wrote: ↑Sun Dec 02, 2018 4:41 am Hello,
I completely rewrote my old chess engine squared-chess from scratch, using pieces of it, so it got a new name.
- move generation is now staged
- transposition table uses two separate tables: always replace and a selective replace
- PVS search is now implemented
- aspiration window
- a million bugs got removed
- code looks nicer
- option: adjustable linear skill levels 0-99
- option: randomized search
- nicer logo
- nicer name
- can now also easily be compiled with Mircosoft Visual Studio (and with a makefile on Linux)
Elo improvement compared to squared-chess 1.3.0 is probably between 80 and 160 (~2150 CCRL 40/4).
Googleplex Starthinker 1.3
Think about an 'Amazon Stockfish' or 'Red Bull Baron' or 'McDonalds Komodo' brrrrrr....
Starthinker is good though.
-
tsoj
- Posts: 35
- Joined: Thu Oct 19, 2017 4:59 pm
- Location: Germany, Berlin
- Full name: Jost Triller
-
Guenther
- Posts: 4718
- Joined: Wed Oct 01, 2008 6:33 am
- Location: Regensburg, Germany
- Full name: Guenther Simon
Re: "Googleplex Starthinker" chess engine
Hi Jost,
I am sorry, I completely missed that 'Googleplex Starthinker' was meant as a reference to the supercomputer in a 'hitchhiker's guide to the galaxy'
and not to the googleplex building
-
tsoj
- Posts: 35
- Joined: Thu Oct 19, 2017 4:59 pm
- Location: Germany, Berlin
- Full name: Jost Triller
Re: "Googleplex Starthinker" chess engine
There is a new release (1.4), mainly fixed segfaults that happen because I thought that 64 quiet moves are enough
https://gitlab.com/tsoj/Googleplex_Starthinker/tags
[pgn]
[Site "https://lichess.org/bm3cO1vP"]
[Date "2018.12.02"]
[Round "14"]
[White "Googleplex Starthinker"]
[Black "Sungorus 1.4"]
[Result "1-0"]
[WhiteElo "?"]
[BlackElo "2311 CCRL"]
[ECO "E68"]
[GameDuration "00:05:33"]
[GameEndTime "2018-12-02T22:08:09.518 CET"]
[GameStartTime "2018-12-02T22:02:35.623 CET"]
[Opening "King's Indian, Fianchetto, Classical Variation, 8. e4"]
[PlyCount "113"]
[TimeControl "40/120"]
1. d4 {book} Nf6 {book} 2. c4 {book} g6 {book} 3. Nc3 {book} Bg7 {book}
4. g3 {book} O-O {book} 5. Bg2 {book} d6 {book} 6. Nf3 {book} Nbd7 {book}
7. O-O {book} e5 {book} 8. e4 {book}{ E68 King's Indian, Fianchetto, Classical Variation, 8. e4 } exd4 {-0.07/12 3.6s}
9. Nxd4 {-0.04/13 3.5s} Kh8 {-0.18/11 3.6s} 10. Qe2 {+0.45/12 2.8s}
Ne8 {+0.07/11 3.6s} 11. Be3 {+0.82/14 5.1s} c5 {+0.33/13 3.6s}
12. Nc2 {+0.60/15 12s} Bxc3 {+0.44/12 3.6s} 13. bxc3 {+0.65/13 2.6s}
Qa5 {+0.69/12 3.6s} 14. Bh6 {+0.52/15 3.9s} Rg8 {+0.71/13 3.6s}
15. Ne3 {+0.77/15 5.1s} Qxc3 {+0.79/12 3.6s} 16. Nd5 {+0.91/15 5.2s}
Qe5 {+0.83/12 3.6s} 17. f4 {+0.83/15 7.8s} Qe6 {+0.86/11 3.6s}
18. Qb2+ {+0.82/13 4.5s} f6 {+0.93/11 3.6s} 19. Qc3 {+0.86/12 2.4s}
Nb6 {+0.95/12 3.6s} 20. a4 {+0.78/13 8.3s} Qf7 {+1.01/12 3.6s}
21. e5 {+0.84/12 2.1s} Nxd5 {+0.83/13 3.6s} 22. cxd5 {+0.96/15 3.7s}
b6 {+0.79/12 3.6s} 23. e6 {+1.13/14 4.7s} Qe7 {+0.64/13 3.6s}
24. g4 {+1.08/14 3.0s} Ba6 {+0.64/11 3.6s} 25. Rfe1 {+1.11/15 4.9s}
Bb7 {+0.65/12 3.6s} 26. g5 {+1.56/16 3.7s} Ba6 {+0.58/11 3.6s}
27. Rab1 {+1.58/13 3.8s} Rc8 {+0.57/12 3.6s} 28. Be4 {+1.56/14 4.0s}
Rb8 {+0.57/12 3.6s} 29. Kh1 {+1.57/11 1.6s} Rb7 {+0.17/12 3.6s}
30. f5 {+1.70/13 3.3s} gxf5 {+0.19/13 3.6s} 31. Bxf5 {+1.71/13 2.2s}
b5 {+0.14/12 3.6s} 32. axb5 {+1.78/12 2.2s} Rxb5 {-0.05/12 3.6s}
33. Bxh7 {+2.58/12 3.7s} Qxh7 {-1.55/12 3.6s} 34. gxf6 {+3.34/13 2.6s}
Qxh6 {-1.71/13 3.6s} 35. f7+ {+3.97/13 5.8s} Ng7 {-1.99/13 3.6s}
36. e7 {+4.33/10 0.61s} Rxb1 {-3.27/13 3.6s} 37. Rxb1 {+5.14/12 1.1s}
Qh5 {-3.55/13 3.6s} 38. f8=Q {+5.29/12 1.4s} Qxd5+ {-4.74/12 3.7s}
39. Qff3 {+5.45/11 1.2s} Qxf3+ {-5.44/12 3.7s} 40. Qxf3 {+5.78/10 0.53s}
Bc8 {-8.40/13 3.3s} 41. Rg1 {+8.62/16 17s} Bd7 {-8.88/14 3.0s}
42. Rg3 {+8.57/13 2.2s} Be8 {-10.58/16 3.0s} 43. Rh3+ {+10.62/15 4.3s}
Nh5 {-9.86/16 3.0s} 44. Rxh5+ {+10.57/14 3.8s} Bxh5 {-10.93/16 3.0s}
45. Qxh5+ {+10.83/14 3.3s} Kg7 {-16.18/15 3.0s} 46. Qg5+ {+10.80/13 2.9s}
Kf7 {-16.54/15 3.0s} 47. Qxg8+ {+10.57/11 2.2s} Kxe7 {-17.16/14 3.0s}
48. h4 {+15.45/12 3.8s} Kd7 {-17.96/13 3.0s} 49. h5 {+M17/13 4.5s}
c4 {-19.58/12 3.0s} 50. Qxc4 {+M15/12 1.9s} d5 {-M14/13 3.0s}
51. Qxd5+ {+M13/10 0.59s} Ke7 {-M12/14 3.0s} 52. Qg5+ {+M11/9 0.24s}
Ke6 {-M10/14 3.0s} 53. h6 {+M9/8 0.17s} a6 {-M8/12 3.0s} 54. h7 {+M7/6 0.097s}
Kf7 {-M6/11 3.0s} 55. h8=Q {+M5/4 0.068s} Ke6 {-M4/10 3.0s}
56. Qge5+ {+M3/2 0.040s} Kf7 {-M2/11 3.0s} 57. Qef6# {+M1/1 0.030s, White mates}
1-0
[/pgn]