feeks

Discussion of anything and everything relating to chess playing software and machines.

Moderators: bob, hgm, Harvey Williamson

Forum rules
This textbox is used to restore diagrams posted with the [d] tag before the upgrade.
Guenther
Posts: 3283
Joined: Wed Oct 01, 2008 4:33 am
Location: Regensburg, Germany
Full name: Guenther Simon
Contact:

Re: feeks

Post by Guenther » Tue Jan 02, 2018 4:16 pm

With pypy3 it looks better now - no depths below 4 and a win vs. same opponent.

Except for the first move it reports around twice as much nps, but I am not sure how reliable the nps count currently is.

[pgn][Event "RWBC"]
[Site "CAPPUCCINO"]
[Date "2018.01.02"]
[Round "-"]
[White "Feeks"]
[Black "Youk V1.05"]
[Result "1-0"]
[TimeControl "180+2"]
[Annotator "1. +0.62"]

1. e4 {+0.62/5} e5 2. Qh5 {+0.68/5 9} Qf6 3. Bb5 {+0.47/5 6} c6 4. Bc4
{+0.45/5 6} d5 5. exd5 {+0.31/5 10} Bc5 6. d4 {+0.70/5 9} Bxd4 7. Nf3
{+0.45/5 8} g6 8. Qh4 {+0.30/4 9} Qxh4 9. Nxh4 {+0.01/5 5} b5 10. Bd3
{+0.50/5 4} e4 11. Bxe4 {+0.92/5 7} Bf6 12. dxc6 {+1.87/5 5} Bxh4 13. c7
{+2.76/6 6} Nd7 14. Bxa8 {+3.83/5 4} Ne5 15. Be3 {+4.46/5 7} Kd7 16. Nc3
{+5.02/5 4} b4 17. Nb5 {+5.17/5 6} Ng4 18. O-O-O+ {+6.28/5 4} Ke6 19. Bd5+
{+6.06/4 5} Kf6 20. Bd4+ {+6.65/5 4} Kg5 21. Bxh8 {+9.30/4 3} Nxf2 22. g3
{+9.80/4 2.8} Bxg3 23. hxg3 {+10.80/4 4} a6 24. Nd6 {+11.42/4 2.2} Nxh1 25.
Rxh1 {+11.67/3 4} Ne7 26. Bf3 {+11.95/3 2.0} b3 27. Nxf7+ {+12.90/4 2.3}
Kf5 28. axb3 {+13.41/4 2.0} Ke6 29. Rxh7 {+13.73/4 3} Nf5 30. Ne5
{+13.72/4 1.4} Nxg3 31. Nf7 {+13.57/4 2.8} Nf5 32. Bc3 {+12.86/4 2.6} Ng3
33. Rg7 {+13.00/4 2.2} Nh5 34. Rxg6+ {+14.17/5 1.4} Kxf7 35. Bxh5
{+14.12/5 2.3} a5 36. Rg5+ {+14.85/5 2.2} Ke7 37. Rg8 {+16.02/5 2.0} Kd6
38. Rxc8 {+17.90/5 1.6} Kd7 39. Bxa5 {+23.11/6 1.8} Ke6 40. Rf8
{+28.83/5 1.5} Kd6 41. c8=Q {+29.63/5 2.3} Kd5 42. Qd7+ {+100.00/5 3} Kc5
43. Rc8# {+100.00/6 2.4}
{Xboard adjudication: Checkmate} 1-0[/pgn]

Decisive difference in that position before move 6:

with pypy3-5.11.0a0 (3.53 32-bit for Win - there is no recent for 64-bit Win)

Code: Select all

Tue Jan  2 16:55:59 2018 rnb1k1nr/pp3ppp/2p2q2/2bPp2Q/2B5/8/PPPP1PPP/RNB1K1NR w KQkq - 1 6
Tue Jan  2 16:55:59 2018 depth 1 score cp 117 time 141 nodes 111 pv b1c3
Tue Jan  2 16:55:59 2018 depth 2 score cp 67 time 611 nodes 461 pv a2a4
Tue Jan  2 16:56:02 2018 depth 3 score cp 111 time 3500 nodes 2416 pv d2d3
Tue Jan  2 16:56:03 2018 depth 4 score cp 65 time 4027 nodes 2879 pv g1f3 c6d5
Tue Jan  2 16:56:07 2018 depth 5 score cp 70 time 8626 nodes 6898 pv d2d4
Tue Jan  2 16&#58;56&#58;07 2018 <LegalMoveGenerator at 0xa2110f0 &#40;Qxh7, Qxf7+, Qh6, Qg6, Qg5, Qf5, Qxe5+, Qh4, Qg4, Qh3, Qf3, Qe2, Qd1, Ba6, Bb5, Bd3, Bb3, Be2, Bf1, Nh3, Nf3, Ne2, Ke2, Kf1, Kd1, Nc3, Na3, dxc6, d6, h3, g3, f3, d3, c3, b3, a3, h4, g4, f4, d4, b4, a4&#41;>
Tue Jan  2 16&#58;56&#58;07 2018 &#91;70.0, Move.from_uci&#40;'d2d4'), 5, 8.625999927520752&#93;
Tue Jan  2 16&#58;56&#58;07 2018 nps&#58; 798.749429, nodes&#58; 6898, tt_hits&#58; 2.083333%
with my standard python installation (winpython 3.63-64)

Code: Select all

Mon Jan  1 22&#58;45&#58;22 2018 rnb1k1nr/pp3ppp/2p2q2/2bPp2Q/2B5/8/PPPP1PPP/RNB1K1NR w KQkq - 1 6
Mon Jan  1 22&#58;45&#58;23 2018 depth 1 score cp 117 time 287 nodes 111 pv b1c3
Mon Jan  1 22&#58;45&#58;24 2018 depth 2 score cp 67 time 1160 nodes 461 pv a2a4
Mon Jan  1 22&#58;45&#58;28 2018 depth 3 score cp 111 time 6023 nodes 2416 pv d2d3
Mon Jan  1 22&#58;45&#58;28 2018 <LegalMoveGenerator at 0x2b15ac8 &#40;Qxh7, Qxf7+, Qh6, Qg6, Qg5, Qf5, Qxe5+, Qh4, Qg4, Qh3, Qf3, Qe2, Qd1, Ba6, Bb5, Bd3, Bb3, Be2, Bf1, Nh3, Nf3, Ne2, Ke2, Kf1, Kd1, Nc3, Na3, dxc6, d6, h3, g3, f3, d3, c3, b3, a3, h4, g4, f4, d4, b4, a4&#41;>
Mon Jan  1 22&#58;45&#58;28 2018 &#91;111.25, Move.from_uci&#40;'d2d3'), 3, 6.022000074386597&#93;
Mon Jan  1 22&#58;45&#58;28 2018 nps&#58; 400.862787, nodes&#58; 2416, tt_hits&#58; 3.370787%
https://rwbc-chess.de
Greg Strong@ovyron wrote: What I think is that you have a very, very, VERY strong desire to post, and very little self-control.
All the time. In every conversation. You must post. Something. It's like you just can't help yourself.

flok

Re: feeks

Post by flok » Tue Jan 02, 2018 7:06 pm

Norbert,

It does not know, sofar, the "stop" command.

flok

Re: feeks

Post by flok » Tue Jan 02, 2018 7:23 pm

flok wrote:Norbert,

It does not know, sofar, the "stop" command.
It now does :D

Guenther
Posts: 3283
Joined: Wed Oct 01, 2008 4:33 am
Location: Regensburg, Germany
Full name: Guenther Simon
Contact:

Re: feeks

Post by Guenther » Wed Jan 03, 2018 1:56 pm

flok wrote:
flok wrote:Norbert,

It does not know, sofar, the "stop" command.
It now does :D
Tried to run lastest master under pypy2 now for speed comparison with pypy3, but it crashes here.
Found too much hits in google for really looking at it, may be a problem with select() in Win.

error code:

Code: Select all

Wed Jan 03 14&#58;34&#58;03 2018 &#40;10093, 'Die Anwendung hat die Funktion WSAStartup nicht aufgerufen, oder bei dieser Funktion ist ein Fehler aufgetreten')
Wed Jan 03 14&#58;34&#58;03 2018 Traceback &#40;most recent call last&#41;&#58;
  File "C&#58;/Engines/UCIPG/Feeks-P2/main.py", line 191, in main
    rlist, _, _ = select&#40;&#91;sys.stdin&#93;, &#91;&#93;, &#91;&#93;, 0.01&#41;
error&#58; &#40;10093, 'Die Anwendung hat die Funktion WSAStartup nicht aufgerufen, oder bei dieser Funktion ist ein Fehler aufgetreten')
https://rwbc-chess.de
Greg Strong@ovyron wrote: What I think is that you have a very, very, VERY strong desire to post, and very little self-control.
All the time. In every conversation. You must post. Something. It's like you just can't help yourself.

Henk
Posts: 6424
Joined: Mon May 27, 2013 8:31 am

Re: feeks

Post by Henk » Wed Jan 03, 2018 2:54 pm

flok wrote:Not at all!
Fully a/b with psq, material, lmr, null move, tt
No nnetwork/deep learning experiments ?
Maybe after some years they all laugh about alpha beta chess engine implementations.

Also alpha beta playing not human like chess.

flok

Re: feeks

Post by flok » Wed Jan 03, 2018 6:23 pm

Also googled for it and it may have to do with import socket being required and/or select not working for other filedescriptors than sockets (and I use it on stdin).

flok

Re: feeks

Post by flok » Wed Jan 03, 2018 6:23 pm

Maybe some other time.

Guenther
Posts: 3283
Joined: Wed Oct 01, 2008 4:33 am
Location: Regensburg, Germany
Full name: Guenther Simon
Contact:

Re: feeks

Post by Guenther » Wed Jan 03, 2018 7:14 pm

flok wrote:Also googled for it and it may have to do with import socket being required and/or select not working for other filedescriptors than sockets (and I use it on stdin).
Right, after regoogling I come to the same conclusion that you cannot use select on stdin in Win.

https://stackoverflow.com/questions/108 ... er-windows
https://rwbc-chess.de
Greg Strong@ovyron wrote: What I think is that you have a very, very, VERY strong desire to post, and very little self-control.
All the time. In every conversation. You must post. Something. It's like you just can't help yourself.

flok

Re: feeks

Post by flok » Thu Jan 04, 2018 6:33 pm

[pgn][Event "Computer Chess Game"]
[Site "lenovo"]
[Date "2018.01.04"]
[Round "-"]
[White "Feeks"]
[Black "GNU Chess"]
[Result "0-1"]
[TimeControl "40/60"]
[Annotator "1. +0.16 1... -0.12"]

1. e3 {+0.16/6} Nc6 {-0.12/13 2.7} 2. Bd3 {+0.90/6 2.1} Nf6 {+0.20/12 3} 3.
f4 {+0.75/6 2.8} e5 {+0.49/11 2.3} 4. Qf3 {+0.60/5 3} d5 {+1.50/10 1.7} 5.
Nc3 {+0.30/4 3} e4 {+2.79/11 1.3} 6. Nxe4 {-1.95/6 2.6} dxe4 {+2.47/11 0.1}
7. Bxe4 {-2.15/6 1.8} Nb4 {+2.57/12 0.9} 8. f5 {-2.33/5 2.4} Nxe4
{+3.21/11 2.8} 9. Qxe4+ {-2.78/6 2.3} Qe7 {+3.13/12 0.1} 10. Qc4
{-2.18/5 2.2} Bxf5 {+4.05/10 1.7} 11. Qb5+ {-3.15/5 1.6} Bd7 {+5.91/11 1.1}
12. Qxd7+ {-3.33/5 1.3} Qxd7 {+13.61/11 1.2} 13. Kd1 {-11.72/5 1.4} Nxc2
{+14.41/10 1.7} 14. Nf3 {-10.75/5 1.5} Nxa1 {+15.74/10 1.1} 15. Ke2
{-16.92/4 1.3} Qa4 {+16.35/10 0.7} 16. Nd4 {-16.97/4 1.2} Qxa2
{+16.75/11 1.5} 17. Nb5 {-17.32/4 1.1} Qc4+ {+20.57/9 0.6} 18. d3
{-20.99/5 1.0} Qxb5 {+21.70/10 0.1} 19. Bd2 {-21.69/5 0.8} Rd8
{+23.69/10 1.1} 20. Bc3 {-22.02/5 0.9} Qxd3+ {+26.93/9 1.2} 21. Kf2
{-23.39/5 0.6} Nc2 {+28.04/9 1.3} 22. Rf1 {-22.64/4 0.8} Qxe3#
{+99.99/38 0.4}
{Xboard adjudication: Checkmate} 0-1[/pgn]

User avatar
CMCanavessi
Posts: 886
Joined: Thu Dec 28, 2017 3:06 pm
Location: Argentina

Re: feeks

Post by CMCanavessi » Thu Jan 04, 2018 6:37 pm

Are there windows binaries (.exe files) or the only way to use the engine is through a python interpreter?

Post Reply