Discussion of anything and everything relating to chess playing software and machines.
Moderators: hgm , Rebel , chrisw
MikeB
Posts: 4889Joined: Thu Mar 09, 2006 6:34 amLocation: Pen Argyl, Pennsylvania
Post by MikeB Mon Apr 24, 2017 4:19 am
jdart wrote: Stockfish 8, 6 cores on a i7-6900k, 8G hash, 30 min search time:
latest yet to be released McBrain locks in Nc6 after 12 seconds and get a comparable score of ~400 after 3 1/2 minutes in a 12 core Mac Pro
Code: Select all
info depth 40 seldepth 74 multipv 1 score cp 281 nodes 2209145765 nps 23581074 hashfull 886 tbhits 11015 time 93683 pv d4c6 d8d1 c2d1 f7f8 c6e7 g8h8 f4c1 a6c5 c1a3 c8e6 e1e6 c5e6 e7g6 h7g6 a3f8 e6f8 f3b7 a8e8 g1f1 h8g7 d1d5 c7c5 d5d6 e8e6 d6b8 e6e7 b7d5 b6a5 b8d6 e7c7 d6d8 a5b6 g2g3 c7d7 d8e8 b6a5 e8e3 d7c7 f1e2 c7d7 e2e1 a5b6 e1f1 d7c7 f1e2 c7d7 e2e1 b6a5 e3c5 a5c3 e1f1 c3e5 c5e3 e5a1 f1g2 a1e5 a2a4 e5a1 a4a5 g5g4 e3d2
info depth 41 seldepth 74 multipv 1 score cp 273 nodes 2360858408 nps 23557464 hashfull 890 tbhits 12659 time 100217 pv d4c6 d8d1
info depth 41 seldepth 74 multipv 1 score cp 266 nodes 2450535949 nps 23561033 hashfull 891 tbhits 13909 time 104008 pv d4c6 d8d1
info depth 41 seldepth 74 multipv 1 score cp 255 nodes 2602453822 nps 23564626 hashfull 897 tbhits 16478 time 110439 pv d4c6 d8d1
info depth 41 seldepth 74 multipv 1 score cp 264 nodes 2672103325 nps 23568925 hashfull 900 tbhits 17365 time 113374 pv d4c6
info depth 41 seldepth 74 multipv 1 score cp 285 nodes 2949113082 nps 23620493 hashfull 907 tbhits 18785 time 124854 pv d4c6
info depth 41 seldepth 74 multipv 1 score cp 313 nodes 3221240187 nps 23696566 hashfull 914 tbhits 22711 time 135937 pv d4c6
info depth 41 seldepth 75 multipv 1 score cp 350 nodes 3604033705 nps 23794498 hashfull 925 tbhits 24806 time 151465 pv d4c6
info depth 41 seldepth 77 multipv 1 score cp 394 nodes 5030864561 nps 24088525 hashfull 941 tbhits 62548 time 208849 pv d4c6 d8d1 c2d1 f7f8 d1d5 g8h8 f4g5 f6g5 e1e7 b7c6 d5e5 h8g8 e5g5 g8h8 g5e5 h8g8 e7e8 c8f5 e8f8 a8f8 f3c6 f5g6 c6d5 g6f7 e5g5 g8h8 g5h6 b6c5 h6f6 h8g8 d5f7 f8f7 f6a6 g8g7 a6a5 c5f2 g1h1 f2b6 a5e5 g7g8 h2h4 c7c5 h4h5 g8f8 e5d6 f8e8 g2g3 f7e7 h1h2 e8f7 d6h6 f7e8 h2g2 b6a5 h6c6 e8f7 c6c5 a5c3 c5f5 f7g8 c4c5 a7a6 f5c8 g8f7 c8a6
info depth 42 seldepth 77 multipv 1 score cp 402 nodes 5694971191 nps 24205182 hashfull 946 tbhits 107430 time 235279 pv d4c6
info depth 42 seldepth 77 multipv 1 score cp 409 nodes 6455709768 nps 24302018 hashfull 953 tbhits 118881 time 265645 pv d4c6
info depth 42 seldepth 77 multipv 1 score cp 420 nodes 6742250719 nps 24333052 hashfull 955 tbhits 122083 time 277082 pv d4c6
info depth 42 seldepth 79 multipv 1 score cp 435 nodes 7273276682 nps 24398208 hashfull 956 tbhits 138200 time 298107 pv d4c6
info depth 42 seldepth 79 multipv 1 score cp 456 nodes 8158219743 nps 24489230 hashfull 961 tbhits 160924 time 333135 pv d4c6
info depth 42 seldepth 80 multipv 1 score cp 440 nodes 8249895042 nps 24506143 hashfull 961 tbhits 169488 time 336646 pv d4c6 d8d1 c2d1 f7f8 d1d5 g8h8 f4g5 f6g5 e1e7 b7c6 d5e5 h8g8 e5g5 g8h8 e7h7 h8h7 f3e4 c8f5 e4f5 f8f5 g5f5 h7g7 f5d7 g7h8 d7c6 a8f8 c6h6 h8g8 h6g6 g8h8 h2h4 b6f2 g1h2 a6c5 h4h5 c5d7 h2h3 f8f6 g6e8 d7f8 e8e5 f8d7 e5c7 f6f7 c7d6 f2c5 d6d5 f7h7 g2g4 c5f8 d5a8 h7f7 g4g5 f7f5 h3g4 f5c5 h5h6 c5c4 g4f5 c4c5 f5e6 c5g5 e6d7 g5g8 a8a7 f8h6 a7d4 g8g7 d7e6 h8h7 d4e4 h7g8 a2a4 h6c1 a4a5 c1b2
up to +481 in 7-1/2 minutes
info depth 43 seldepth 85 multipv 1 score cp 481 nodes 11178440452 nps 24719687 hashfull 967 tbhits 417677 time 452208 pv d4c6
and + 498 in less than 10 minutes
info depth 45 seldepth 92 multipv 1 score cp 498 nodes 14471994657 nps 24868192 hashfull 974 tbhits 601252 time 581948 pv d4c6 d8d1 c2d1 f7f8 d1d5 g8h8 f4g5 f6g5 e1e7 b7c6 d5e5 h8g8 e5g5 g8h8 e7h7 h8h7 f3e4 c8f5 e4f5 f8f5 g5f5 h7g7 f5d7 g7h8 d7c6 a8f8 c6h6 h8g8 h6g6 g8h8 h2h4 c7c6 g6c6 b6f2 g1h2 f2b6 h4h5 a6c7 c6h6 h8g8 h6g6 g8h8 h2h3 f8e8 g6h6 h8g8 c4c5 b6a5 h6g6 g8h8 g6f6 h8g8 c5c6 e8e3 h3h4 a5b6 f6g5 g8h8 h5h6 c7e8 g2g4 e3e4 g5g6 e4e5 g6f7 b6d8 h4g3 e5e3 g3f4 e3e1 f7a7 h8g8 g4g5 e8c7 f4g4 c7e6
jdart
Posts: 4366Joined: Fri Mar 10, 2006 5:23 amLocation: http://www.arasanchess.org
Post by jdart Mon Apr 24, 2017 5:06 am
The second one is hard for Arasan - I am not sure why.
Ferdy
Posts: 4833Joined: Sun Aug 10, 2008 3:15 pmLocation: Philippines
Post by Ferdy Mon Apr 24, 2017 5:23 am
jdart wrote: The second one is hard for Arasan - I am not sure why.
Deuterium (weaker than Arasan) started to like Ra6 at 8s.
[d]1rr5/5R2/6k1/3B2P1/1p2P1n1/1PpK4/8/R7 w - - bm Ra6+; id "arasan20.197";
Code: Select all
20/39 00:08 +1.55 1.Ra6+ Kxg5 2.Rg7+ Kf4 3.Ra1 Nf2+ 4.Kd4 c2 5.Rf7+ Kg3 6.Rg1+ Kh2 7.Rc1 Nd1 8.e5 Rc3
21/41 00:10 +1.39 1.Ra6+ Kxg5 2.Rg7+ Kf4 3.Ra1 Nf2+ 4.Kd4 c2 5.Rf7+ Kg3 6.Rg1+ Kh2 7.Rc1 Nd1 8.e5 Kg3
jdart
Posts: 4366Joined: Fri Mar 10, 2006 5:23 amLocation: http://www.arasanchess.org
Post by jdart Mon Apr 24, 2017 11:08 am
Arasan, latest dev version, i7-6900k (8 cores):
Code: Select all
bm Ra6+
1 0.00 Rf5 +1.02 374
1 0.00 Rf5 +1.02 484
2 0.00 Rf5 +1.07 1454
3 0.00 Rf5 +1.36 3007
4 0.00 -- +1.36 3558
4 0.00 Rf5 +0.82 4242
5 0.00 Rf5! +1.01 4395
5 0.00 Rf5 +1.10 5625
6 0.01 Rf5 +1.06 7941
7 0.01 Rf5 +0.95 11361
8 0.01 Rf5 +0.81 16393
9 0.01 Rf5 +0.97 23302
10 0.02 Rf5! +1.16 35083
10 0.02 Rf5! +1.54 37752
10 0.05 -- +1.54 140097
10 0.06 Rf5 +0.25 206822
11 0.07 Rf5! +0.43 222190
11 0.07 Rf5 +0.68 250260
12 0.08 Rf5 +0.74 271632
13 0.09 -- +0.74 311044
13 0.16 Rf4! +0.93 770894
13 0.17 Rf4 +0.84 867227
14 0.19 Rf4! +1.02 903794
14 0.19 Rf4 +1.06 947940
15 0.20 -- +1.06 1014705
15 0.27 Rf4 +0.94 1299092
16 0.28 -- +0.94 1413579
16 0.30 Rf4 +0.62 1520234
17 0.31 Rf4! +0.81 1601035
17 0.38 Rf4 +0.57 1973656
18 0.46 -- +0.57 2548546
18 0.48 Rf4 +0.43 2635526
19 0.50 Rf4! +0.62 2788900
19 0.72 Rf4 +0.36 4510611
20 0.80 Rf4! +0.54 5034806
20 0.88 Rf4 +0.65 5607688
21 0.99 Rf4 +0.50 6554294
22 1.48 Rf4 +0.68 9657785
23 1.97 -- +0.68 13632424
23 2.41 Rf4 +0.37 16886010
24 2.50 Rf4! +0.56 17582313
24 2.88 Rf4 +0.69 20430487
25 3.31 Rf4 +0.69 24223531
26 4.39 Rf4 +0.81 33661993
27 5.21 Rf4 +0.81 40904461
28 7.30 -- +0.81 60012118
28 7.73 Rf4 +0.80 63399243
29 10.33 -- +0.80 85222005
29 12.17 Rf4 +0.81 99759820
30 16.65 Rf4 +0.80 132203393
31 21.18 Rf4 +0.80 172209160
32 26.11 -- +0.80 219376628
32 48.62 Ra6+! +0.99 437256431
32 60.65 Ra6+! +1.74 515563480
32 81.10 Ra6+ +2.08 699391760
33 83.50 -- +2.08 722460991
33 86.44 Ra6+ +1.92 748553994
34 89.92 Ra6+ +1.86 784151114
35 96.14 Ra6+ +1.99 845298979
36 103.32 -- +1.99 907510746
36 108.62 Ra6+ +1.79 954654258
37 117.62 Ra6+! +1.97 1025154431
37 136.11 Ra6+ +2.26 1174000090
38 141.14 -- +2.26 1219913123
This is still better than Arasan 20.0. Could be over-pruning, but I have also seen cases where a relatively small eval difference prevents finding a move, or finding it fast. That is all you need to steer the search down a different line.
--Jon
Dann Corbit
Posts: 12540Joined: Wed Mar 08, 2006 8:57 pmLocation: Redmond, WA USA
Post by Dann Corbit Wed Apr 26, 2017 4:34 am
For the two new problems:
Code: Select all
MPV: 1
b1r1r1k1/p2n1p2/1p5p/2qp1Rn1/2P3pN/6P1/P2N1P1P/Q3RBK1 b - - acd 47/93-; acs 8385 ; acn 91500708k; ce 1536 ; bm Ra6+; pv Ra6+ Kh5;
b1r1r1k1/p2n1p2/1p5p/2qp1Rn1/2P3pN/6P1/P2N1P1P/Q3RBK1 b - - acd 47/93+; acs 7482 ; acn 81501665k; ce 1501 ; bm Ra6+; pv Ra6+;
b1r1r1k1/p2n1p2/1p5p/2qp1Rn1/2P3pN/6P1/P2N1P1P/Q3RBK1 b - - acd 47/93+; acs 8251 ; acn 90155947k; ce 1707 ; bm Ra6+; pv Ra6+;
b1r1r1k1/p2n1p2/1p5p/2qp1Rn1/2P3pN/6P1/P2N1P1P/Q3RBK1 b - - acd 47/93+; acs 8587 ; acn 93810938k; ce 1751 ; bm Ra6+; pv Ra6+;
b1r1r1k1/p2n1p2/1p5p/2qp1Rn1/2P3pN/6P1/P2N1P1P/Q3RBK1 b - - acd 47/99-; acs 13638 ; acn 146407540k; ce 2233 ; bm Ra6+; pv Ra6+ Kh5;
b1r1r1k1/p2n1p2/1p5p/2qp1Rn1/2P3pN/6P1/P2N1P1P/Q3RBK1 b - - acd 47/99-; acs 14709 ; acn 156108345k; ce 2104 ; bm Ra6+; pv Ra6+ Kh5;
b1r1r1k1/p2n1p2/1p5p/2qp1Rn1/2P3pN/6P1/P2N1P1P/Q3RBK1 b - - acd 47/99+; acs 13000 ; acn 141189707k; ce 2674 ; bm Ra6+; pv Ra6+;
b1r1r1k1/p2n1p2/1p5p/2qp1Rn1/2P3pN/6P1/P2N1P1P/Q3RBK1 b - - acd 47/99+; acs 13948 ; acn 149291231k; ce 2782 ; bm Ra6+; pv Ra6+;
b1r1r1k1/p2n1p2/1p5p/2qp1Rn1/2P3pN/6P1/P2N1P1P/Q3RBK1 b - - acd 47/99+; acs 14711 ; acn 156120450k; ce 2948 ; bm Ra6+; pv Ra6+;
b1r1r1k1/p2n1p2/1p5p/2qp1Rn1/2P3pN/6P1/P2N1P1P/Q3RBK1 b - - acd 47/101+; acs 19252 ; acn 205966321k; ce 6511 ; bm Ra6+; pv Ra6+;
MPV:3
b1r1r1k1/p2n1p2/1p5p/2qp1Rn1/2P3pN/6P1/P2N1P1P/Q3RBK1 b - - acd 48/80; acs 2264 ; acn 21999671k; ce 0 ; bm Re7; pv Re7 c2 Kd2 Rc3 Rg1 Rd8 Rb7 Rdc8 Kc1 Ne5 Rb6+ Kg7 Rb7+ Kg6;
b1r1r1k1/p2n1p2/1p5p/2qp1Rn1/2P3pN/6P1/P2N1P1P/Q3RBK1 b - - acd 48/80; acs 2264 ; acn 21999671k; ce 0 ; bm Rf4; pv Rf4 c2 Bc4 Ne5+ Kxc2 Nxc4 bxc4 Rxc4+ Kb2 Rc5 Rf6+ Kxg5 Rf2 Kg6 Rg1+ Rg5 Rh1 Re5 Rg1+ Rg5;
b1r1r1k1/p2n1p2/1p5p/2qp1Rn1/2P3pN/6P1/P2N1P1P/Q3RBK1 b - - acd 48/80; acs 2264 ; acn 21999671k; ce 897 ; bm Ra6+; pv Ra6+ Kxg5 Rg7+ Kf4 Ra1 Ne5+ Ke2 Ng4 Rf1+ Ke5 Kd3 Rf8 Rg5+ Kd6 Rxf8 Rxf8 Rxg4 Kc5 Rg5 Rf3+ Ke2 c2 Rg1 Rc3 Kd2 Kd4 Rc1 Rh3 Rxc2 Rh2+ Kc1 Rh1+ Kb2 Ke5 Rc4 Rh2+ Kb1 Rh1+ Rc1 Rh5 Rc6 Rh2 Kc1 Kf4 Rb6 Ke5 Rxb4 Rf2 Rc4 Rg2 Rc3 Rf2 Rc2 Rf8 Kb2 Rf3 Rc4 Rh3 Kc2 Rh2+ Kd3 Rh3+ Ke2 Rh1 Kd2 Rh2+ Kc3 Rh3+ Kb4 Rh7 Rc5;
MPV:1
rn2r1k1/p4pn1/1p2p3/qPppP1Q1/3P4/2P2N2/2P2PPP/1R3RK1 w - - acd 49/90-; acs 17430 ; acn 157161171k; ce 610 ; bm Nh4; pv Nh4 Nd7;
rn2r1k1/p4pn1/1p2p3/qPppP1Q1/3P4/2P2N2/2P2PPP/1R3RK1 w - - acd 49/90; acs 17480 ; acn 157634678k; ce 604 ; bm Nh4; pv Nh4 Nd7 Rbd1 c4 f4 Qxc3 f5 f6 exf6 Nxf6 Qxf6 Qe3+ Kh1 Rf8 Qg6 Qe4 f6 Qxg6 Nxg6 Nf5 Nxf8 Rxf8 Ra1 Nxd4 Rxa7 Rf7 Ra6 Nxb5 Rxb6 Nd4 Rf4 e5 Rg4+ Kh7 c3 Nf5 Rg5 Ne3 Rxe5 Ng4 Rxd5 Ra7 Rb1 Nxf6 Rc5 Ne4 Rxc4 Nd6 Rc6 Nf7 Kg1 Ne5 Re6 Nd3 Re4 Rg7 Rd4 Nc5 Rb5 Rg5 Rc4 Ne6 Rxg5 Nxg5 Kf2 Kg6 h4;
rn2r1k1/p4pn1/1p2p3/qPppP1Q1/3P4/2P2N2/2P2PPP/1R3RK1 w - - acd 49/90+; acs 16342 ; acn 146945845k; ce 610 ; bm Nh4; pv Nh4;
rn2r1k1/p4pn1/1p2p3/qPppP1Q1/3P4/2P2N2/2P2PPP/1R3RK1 w - - acd 49/90+; acs 17413 ; acn 157005023k; ce 617 ; bm Nh4; pv Nh4;
rn2r1k1/p4pn1/1p2p3/qPppP1Q1/3P4/2P2N2/2P2PPP/1R3RK1 w - - acd 50/90-; acs 19253 ; acn 173855577k; ce 612 ; bm Nh4; pv Nh4 Nd7;
rn2r1k1/p4pn1/1p2p3/qPppP1Q1/3P4/2P2N2/2P2PPP/1R3RK1 w - - acd 50/90-; acs 19467 ; acn 175803294k; ce 509 ; bm Nh4; pv Nh4 Nd7;
rn2r1k1/p4pn1/1p2p3/qPppP1Q1/3P4/2P2N2/2P2PPP/1R3RK1 w - - acd 50/90; acs 19659 ; acn 177530321k; ce 505 ; bm Nh4; pv Nh4 Nd7 Rbd1 c4 f4 Qxc3 f5 f6 exf6 Nxf6 Qxf6 Qe3+ Kh1 Rf8 Qg6 Qe4 f6 Qxg6 Nxg6 Nf5 Nxf8 Rxf8 Ra1 Nxd4 Rxa7 Rf7 Ra6 Nxb5 Rxb6 Nd4 Rf4 e5 Rg4+ Kh7 c3 Nf5 Rg5 Ne3 Rxe5 Ng4 Rxd5 Ra7 Rb1 Nxf6 Rd4 Rc7 Kg1 Kg6 h4 Kh5 Kf2 Rc5 Kf3 Rf5+ Kg3 Nd5 Rxc4;
rn2r1k1/p4pn1/1p2p3/qPppP1Q1/3P4/2P2N2/2P2PPP/1R3RK1 w - - acd 50/90+; acs 17775 ; acn 160368132k; ce 612 ; bm Nh4; pv Nh4;
rn2r1k1/p4pn1/1p2p3/qPppP1Q1/3P4/2P2N2/2P2PPP/1R3RK1 w - - acd 50/90+; acs 19243 ; acn 173765307k; ce 619 ; bm Nh4; pv Nh4;
rn2r1k1/p4pn1/1p2p3/qPppP1Q1/3P4/2P2N2/2P2PPP/1R3RK1 w - - acd 51/90+; acs 19953 ; acn 180261547k; ce 513 ; bm Nh4; pv Nh4;
MPV:3
rn2r1k1/p4pn1/1p2p3/qPppP1Q1/3P4/2P2N2/2P2PPP/1R3RK1 w - - acd 46/88; acs 25093 ; acn 172341030k; ce 139 ; bm Qf4; pv Qf4 Nd7 Ng5 Re7 Nh7 Kxh7 Qh4+ Kg8 Qxe7 Nf8 Qb7 Rd8 dxc5 bxc5 Ra1 Qxc3 Rxa7 Rd7 Qa8 Rxa7 Qxa7 Nf5 b6 Qxc2 b7 Nh4 Kh1 Nd7 b8R+ Nxb8 Qxb8+ Kh7 Qb1 Qxb1 Rxb1 c4 Rb7 Kg7 Rc7 Ng6 g3 Nxe5 f4 Nd3 Kg2 Kg6 Kf3 Ne1+ Kf2 Nd3+ Ke2 Kf6 h4 Kg6 Kf3 f5 Ke3 Ne1 Kd4 Nf3+ Kc5 Nd2 Rc6 Nb3+ Kb4 Nd4 Ra6 Kf7 h5 Ne2;
rn2r1k1/p4pn1/1p2p3/qPppP1Q1/3P4/2P2N2/2P2PPP/1R3RK1 w - - acd 47/88; acs 25093 ; acn 172341030k; ce 329 ; bm Rbd1; pv Rbd1 c4 Nh4 Qxc3 f4 Qe3+ Kh1 Qe2 g4 Nd7 f5 f6 exf6 Nxf6 Qxf6 exf5 gxf5 Re4 Qd6 Rxh4 Rde1 Qxc2 f6 Rf8 Qg3 Rxh2+ Qxh2 Qxh2+ Kxh2 Nh5 Re6 c3 Rc6 c2 Rxc2 Nxf6 Rg1+ Kf7 Rc7+ Ke6 Rc6+ Kf7 Ra1 Kg6 Kg2 Rf7 Rf1 Kg7 Rf5 Ne4 Rxd5 Nd2 Kh3 Ne4 Rd8 Nf6 Kg3 Ne4+ Kg4 Nf2+ Kh4 Nd3 d5 Ne5 d6 Kf6 d7+ Kf5 Rc7 Kf4;
rn2r1k1/p4pn1/1p2p3/qPppP1Q1/3P4/2P2N2/2P2PPP/1R3RK1 w - - acd 47/88; acs 25093 ; acn 172341030k; ce 578 ; bm Nh4; pv Nh4 Nd7 Rbd1 c4 f4 Qxc3 f5 f6 exf6 Nxf6 Qxf6 Qe3+ Kh1 Rf8 Qg6 Qe4 Qg3 Rf6 Qg5 Raf8 Ng6 exf5 Nxf8 Rxf8 Ra1 Qxd4 Rxa7 Qf6 Qg3 Kh7 Rd7 f4 Qf3 Qf5 Qxd5 Qxd5 Rxd5 Rf7 Rd4 Rf5 Rxc4 Rxb5 Rcxf4 Rg5 g4 Rc5 R1f2 Ne6 Rf7+ Kg8 Rb7 Rc7 Rxb6 Nc5 Rb4 Na6 Re4 Nc5 Re8+ Kh7 Kg2 Na6 Re4 Nc5 Rc4 Kg7 Kg3 Ne6 Rxc7+ Nxc7;
Taking ideas is not a vice, it is a virtue. We have another word for this. It is called learning.
MikeB
Posts: 4889Joined: Thu Mar 09, 2006 6:34 amLocation: Pen Argyl, Pennsylvania
Post by MikeB Wed Apr 26, 2017 5:04 am
let this run in "normal" search mode with latest McBrain (yet to be released), took just under an hour, but then it quickly saw the damage Rxg7 created , needed 85 billion nodes to see the move , just over 100 billion to pv the damage
[d]1r1rb1k1/5ppp/4p3/1p1p3P/1q2P2Q/pN3P2/PPP4P/1K1R2R1 w - -
Code: Select all
info depth 50 seldepth 84 multipv 1 score cp 112 nodes 47905892773 nps 22550996 hashfull 997 tbhits 7975452 time 2124336 pv d1d4 b4d6 h4f6 d6f8 e4d5 g8h8 f6h4 e6e5 h5h6 g7g5 h4g5 f7f6 g5g7 f8g7 h6g7 h8g8 d4d2 b8c8 b2a3 e8f7 d5d6 c8c6 d6d7 f7e6 f3f4 d8d7 f4e5 d7d2 b3d2 f6e5 g1g5 e6a2 b1b2 a2f7 g5e5 f7g6 c2c3 g6d3 e5d5 d3e2 d5g5 c6f6 d2b3 e2c4 b3d4 f6f2 b2c1 h7h6 g5g3 f2h2 d4f5 g8h7 f5d6 c4g8 d6e4 h2h1 c1d2 h1f1 d2e2 f1f7 e2e3 f7f5 g3g1 f5f7 g1g2 f7e7 e3d4 e7d7 d4c5 d7g7 e4f6 h7h8 g2g7 h8g7
info depth 51 seldepth 84 multipv 1 score cp 119 nodes 59919372327 nps 22528468 hashfull 997 tbhits 9963164 time 2659718 pv d1d4
info depth 51 seldepth 84 multipv 1 score cp 126 nodes 62671800359 nps 22540521 hashfull 997 tbhits 10564380 time 2780406 pv d1d4
info depth 51 seldepth 84 multipv 1 score cp 137 nodes 76928487721 nps 22948944 hashfull 998 tbhits 12353015 time 3352158 pv d1d4
info depth 51 seldepth 84 multipv 1 score cp 152 nodes 85052430577 nps 23331266 hashfull 998 tbhits 13507113 time 3645427 pv g1g7
info depth 51 seldepth 84 multipv 1 score cp 173 nodes 85057480694 nps 23331500 hashfull 998 tbhits 13508769 time 3645607 pv g1g7
info depth 51 seldepth 84 multipv 1 score cp 202 nodes 85059454877 nps 23331587 hashfull 998 tbhits 13509050 time 3645678 pv g1g7
info depth 51 seldepth 84 multipv 1 score cp 239 nodes 85062285276 nps 23331723 hashfull 998 tbhits 13509534 time 3645778 pv g1g7
info depth 51 seldepth 84 multipv 1 score cp 287 nodes 85075236072 nps 23332319 hashfull 998 tbhits 13511249 time 3646240 pv g1g7
info depth 51 seldepth 84 multipv 1 score cp 350 nodes 85140980137 nps 23335003 hashfull 998 tbhits 13520136 time 3648638 pv g1g7
info depth 51 seldepth 84 multipv 1 score cp 429 nodes 85146085344 nps 23335270 hashfull 998 tbhits 13520712 time 3648815 pv g1g7
info depth 51 seldepth 84 multipv 1 score cp 531 nodes 85150361867 nps 23335457 hashfull 998 tbhits 13521231 time 3648969 pv g1g7
info depth 51 seldepth 84 multipv 1 score cp 660 nodes 85207404925 nps 23338119 hashfull 998 tbhits 13530840 time 3650997 pv g1g7
info depth 51 seldepth 84 multipv 1 score cp 823 nodes 85302208138 nps 23339423 hashfull 998 tbhits 13548280 time 3654855 pv g1g7
info depth 51 seldepth 84 multipv 1 score cp 1029 nodes 87063274007 nps 23419334 hashfull 998 tbhits 13907481 time 3717581 pv g1g7
info depth 51 seldepth 90 multipv 1 score cp 1253 nodes 103397372427 nps 24348449 hashfull 998 tbhits 16184119 time 4246569 pv g1g7 g8g7 h5h6 g7f8 h4f6 e8c6 c2c3 b4b3 a2b3 f8e8 d1g1 e8d7 f6f7 d7c8 f7e6 c6d7 e6d5 b8b6 g1g8 d8e8 g8e8 d7e8 d5f5 c8d8 f5h7 e8g6 h7g8 d8e7 h6h7 g6h7 g8h7 e7f8 b2a3 b6a6 h7f5 f8g7 f5b5 a6g6 b5d7 g7h6 d7h3 h6g7 b3b4 g6g1 b1b2 g1g6 a3a4 g6a6 h3h5 a6a7 a4a5 a7a6 h5f5 g7h6 f5e5 a6a7 e5d5 h6g6 d5d6 g6h7 d6c5 a7a6 c5e5 h7h6 e5f4 h6h5
Vinvin
Posts: 5228Joined: Thu Mar 09, 2006 9:40 amFull name: Vincent Lejeune
Post by Vinvin Wed Apr 26, 2017 11:46 am
MikeB wrote: let this run in "normal" search mode with latest McBrain (yet to be released), took just under an hour, but then it quickly saw the damage Rxg7 created , needed 85 billion nodes to see the move , just over 100 billion to pv the damage
8 years ago, spark-0.3 found it after 383 Million nodes !
= 8m19s on an AMD 4200+, 2.2 GHz 1 thread.
http://www.talkchess.com/forum/viewtopi ... 794#300794
MikeB
Posts: 4889Joined: Thu Mar 09, 2006 6:34 amLocation: Pen Argyl, Pennsylvania
Post by MikeB Wed Apr 26, 2017 12:14 pm
Vinvin wrote: MikeB wrote: let this run in "normal" search mode with latest McBrain (yet to be released), took just under an hour, but then it quickly saw the damage Rxg7 created , needed 85 billion nodes to see the move , just over 100 billion to pv the damage
8 years ago, spark-0.3 found it after 383 Million nodes !
= 8m19s on an AMD 4200+, 2.2 GHz 1 thread.
http://www.talkchess.com/forum/viewtopi ... 794#300794
that's impressive - left it run overnight , it eventually hit a tb mate score
Code: Select all
info depth 56 seldepth 110 multipv 1 score cp 12824 nodes 1030538350922 nps 33698611 hashfull 999 tbhits 266539828 time 30581033 pv g1g7 g8g7 h5h6 g7f8 h4f6 e8c6 c2c3 b4b3 a2b3 f8e8 d1g1 a3a2 b1a2 b8a8 a2b1 e8d7 f6f7 d7c8 f7e6 c8b7 g1g7 b7b6 e4d5 d8d5 e6e3 b6a6 g7h7 d5d1 b1c2 d1d7 h7d7 c6d7 e3e5 d7e8 e5f6 a6b7 h6h7 e8c6 h7h8q a8h8 f6h8 b7c7 h8f6 c7d7 f6c6 d7c6 b3b4 c6d5 c2d3 d5d6 c3c4 d6c7 c4b5 c7b6 d3c2 b6c7 b5b6 c7b7 c2d3 b7a8 b6b7 a8b8 d3c2 b8a7 b7b8q a7b8
bestmove g1g7 ponder g8g7
Vinvin
Posts: 5228Joined: Thu Mar 09, 2006 9:40 amFull name: Vincent Lejeune
Post by Vinvin Wed Apr 26, 2017 11:50 pm
MikeB wrote: Vinvin wrote: MikeB wrote: let this run in "normal" search mode with latest McBrain (yet to be released), took just under an hour, but then it quickly saw the damage Rxg7 created , needed 85 billion nodes to see the move , just over 100 billion to pv the damage
8 years ago, spark-0.3 found it after 383 Million nodes !
= 8m19s on an AMD 4200+, 2.2 GHz 1 thread.
http://www.talkchess.com/forum/viewtopi ... 794#300794
that's impressive - left it run overnight , it eventually hit a tb mate score
Code: Select all
info depth 56 seldepth 110 multipv 1 score cp 12824 nodes 1030538350922 nps 33698611 hashfull 999 tbhits 266539828 time 30581033 pv g1g7 g8g7 h5h6 g7f8 h4f6 e8c6 c2c3 b4b3 a2b3 f8e8 d1g1 a3a2 b1a2 b8a8 a2b1 e8d7 f6f7 d7c8 f7e6 c8b7 g1g7 b7b6 e4d5 d8d5 e6e3 b6a6 g7h7 d5d1 b1c2 d1d7 h7d7 c6d7 e3e5 d7e8 e5f6 a6b7 h6h7 e8c6 h7h8q a8h8 f6h8 b7c7 h8f6 c7d7 f6c6 d7c6 b3b4 c6d5 c2d3 d5d6 c3c4 d6c7 c4b5 c7b6 d3c2 b6c7 b5b6 c7b7 c2d3 b7a8 b6b7 a8b8 d3c2 b8a7 b7b8q a7b8
bestmove g1g7 ponder g8g7
+result from the old HIARCS 13.1 on an i7-6*4GHz on 3runs : 1m34s, 2m16s and 2m26s
Code: Select all
Hiarcs 13.1:
...
21 00:50 606 975k 11 989k +0,43 1.Qf6 Qf8 2.exd5 exd5 3.bxa3 Rbc8 4.Rde1 d4 5.h4 Kh8 6.Qe7 Qxe7 7.Rxe7 Kg8 8.h6 g6 9.Rd1 Kf8 10.Re4 Rc3 11.Nxd4 Ra8 12.a4 bxa4 13.Kb2 Rac8 14.a3
22+ 01:34 1 111 790k 11 806k +0,51 1.Rxg7+
22+ 01:34 1 111 790k 11 806k +0,51 1.Rxg7+
22+ 03:22 2 364 019k 11 694k +4,41 1.Rxg7+
Hiarcs 13.1:
...
23 01:38 1 171 621k 11 938k +0,42 1.Qg5 g6 2.exd5 exd5 3.Rxd5 Rxd5 4.Qxd5 Qc4 5.Qd4 Qxd4 6.Nxd4 b4 7.b3 Rd8 8.Rd1 Rc8 9.hxg6 hxg6 10.Kc1 Kg7 11.Kd2 Bc6 12.Ke3 Bb7 13.Kf4 Rh8 14.Kg3 Re8 15.Nb5
23+ 02:16 1 628 305k 11 942k +0,52 1.Rxg7+
23+ 07:47 5 644 397k 12 066k +6,36 1.Rxg7+
Hiarcs 13.1:
...
23 01:25 1 014 271k 11 932k +0,35 1.Qf6 Qf8 2.exd5 Rxd5 3.Rxd5 exd5 4.Qe5 Ra8 5.Qxd5 Rc8 6.f4 h6 7.f5 Qe7 8.Qd2 b4 9.Qf4 Qf6 10.Qd4 Qxd4 11.Nxd4 Rd8 12.Rg4 Bd7 13.Kc1 Rc8 14.Rg3 axb2+ 15.Kxb2 Kf8 16.Kb3 Ra8 17.c4 bxc3 18.Rxc3 Rb8+ 19.Kc2
23+ 02:26 1 719 480k 11 743k +0,43 1.Rxg7+
23+ 05:14 3 728 207k 11 845k +4,33 1.Rxg7+
ZapMexII is even faster, around 1m20s :
Code: Select all
Zappa Mexico II x64:
17/53 00:41 236 222k 5 733k +0,64 1.Rd4 Qf8 2.exd5 Rxd5 3.Rxd5 exd5 4.h6 g6 5.Re1 b4 6.Nd4 axb2 7.Kxb2 Qc5 8.Kb1 Qc3
17/53 01:12 417 504k 5 785k +4,44 1.Rxg7+ Kxg7 2.h6+ Kf8 3.Qf6 Bc6 4.c3 Qc4 5.Rg1 Ke8 6.Rg7 Qd3+ 7.Kc1 axb2+ 8.Kxb2 Qe2+ 9.Kc1 Qf1+ 10.Kc2 Qf2+ 11.Nd2 Qa7 12.Rg8+ Kd7 13.Qxf7+ Kc8 14.Rxd8+ Kxd8 15.Qxa7
17/53 01:12 420 548k 5 765k +4,44 1.Rxg7+ Kxg7 2.h6+ Kf8 3.Qf6 Bc6 4.c3 Qc4 5.Rg1 Ke8 6.Rg7 Qd3+ 7.Kc1 axb2+ 8.Kxb2 Qe2+ 9.Kc1 Qf1+ 10.Kc2 Qf2+ 11.Nd2 Qa7 12.Rg8+ Kd7 13.Qxf7+ Kc8 14.Rxd8+ Kxd8 15.Qxa7
Zappa Mexico II x64:
18/55 00:55 329 856k 5 899k +0,73 1.Qf6 Qf8 2.exd5 Rxd5 3.Rxd5 exd5 4.Nd4 Bd7 5.h6 g6 6.Re1 b4 7.h4 axb2 8.Kxb2 Re8 9.Rxe8 Bxe8 10.Qd8 Ba4 11.Qxd5
18/55 01:49 626 788k 5 730k +6,34 1.Rxg7+ Kxg7 2.h6+ Kf8 3.Qf6 Bd7 4.exd5 Bc6 5.Rg1 Ke8 6.dxc6 Qf8 7.c7 Rbc8 8.cxd8Q+ Rxd8 9.bxa3 Rd7 10.Rg7 Rd1+ 11.Kb2 Kd7 12.Rxh7
18/55 01:51 638 124k 5 699k +6,34 1.Rxg7+ Kxg7 2.h6+ Kf8 3.Qf6 Bd7 4.exd5 Bc6 5.Rg1 Ke8 6.dxc6 Qf8 7.c7 Rbc8 8.cxd8Q+ Rxd8 9.bxa3 Rd7 10.Rg7 Rd1+ 11.Kb2 Kd7 12.Rxh7
Zappa Mexico II x64:
17/53 00:43 252 014k 5 789k +0,71 1.Qf6 Qf8 2.exd5 Rxd5 3.Rxd5 exd5 4.Nd4 Bd7 5.h6 g6 6.Re1 b4 7.h4 axb2 8.Re7 Rb6 9.Qg5 Rd6 10.Kxb2
17/53 01:14 421 778k 5 657k +5,86 1.Rxg7+ Kxg7 2.h6+ Kf8 3.Qf6 Bd7 4.exd5 Bc6 5.Rg1 Ke8 6.dxc6 Qf8 7.c7 Ra8 8.cxd8Q+ Rxd8 9.Nc5 axb2 10.Nxe6 Rd6 11.Nc7+ Kd7 12.Qg7 Qd8 13.Nxb5 Kc6 14.Nxd6 Qxd6 15.Qxh7
17/53 01:15 425 893k 5 635k +5,86 1.Rxg7+ Kxg7 2.h6+ Kf8 3.Qf6 Bd7 4.exd5 Bc6 5.Rg1 Ke8 6.dxc6 Qf8 7.c7 Ra8 8.cxd8Q+ Rxd8 9.Nc5 axb2 10.Nxe6 Rd6 11.Nc7+ Kd7 12.Qg7 Qd8 13.Nxb5 Kc6 14.Nxd6 Qxd6 15.Qxh7
Dann Corbit
Posts: 12540Joined: Wed Mar 08, 2006 8:57 pmLocation: Redmond, WA USA
Post by Dann Corbit Thu Apr 27, 2017 8:59 pm
For your move ordering spy, I analyzed the set a bit.
The "move ordering controversies" can be found here:
http://rybkaforum.net/cgi-bin/rybkaforu ... pid=572997
I am not sure which answers are right.
Some of them are only searched to 30 plies.
Taking ideas is not a vice, it is a virtue. We have another word for this. It is called learning.
