Finally official Lc0 v0.31.0-rc1
Moderators: hgm, Rebel, chrisw
-
- Posts: 3425
- Joined: Wed Mar 08, 2006 8:15 pm
Finally official Lc0 v0.31.0-rc1
here https://github.com/LeelaChessZero/lc0/r ... 0.31.0-rc1. For Apple users: Multihead arch metal added. But I think biggest problem for Lc0 that there is no official net! So we will NEVER know how strong it really is.
Jouni
-
- Posts: 153
- Joined: Sun May 16, 2021 5:33 pm
- Full name: Aron Rodgriges
Re: Finally official Lc0 v0.31.0-rc1
But what we DO KNOW, is that LC0 is faster and better on most modern cheap nvidia laptops, than on an úberpriced Apple computer
-
- Posts: 613
- Joined: Sun Aug 04, 2013 1:19 pm
Re: Finally official Lc0 v0.31.0-rc1
BT4-1024x15x32h-swa-6315000.pb.gzJouni wrote: ↑Tue Mar 26, 2024 6:02 pm here https://github.com/LeelaChessZero/lc0/r ... 0.31.0-rc1. For Apple users: Multihead arch metal added. But I think biggest problem for Lc0 that there is no official net! So we will NEVER know how strong it really is.
It runs fine and it's strong enough.
-
- Posts: 2490
- Joined: Wed Mar 08, 2006 9:40 pm
- Location: Eden Prairie, Minnesota
- Full name: Stephen Ham
Finally official Lc0 v0.31.0-rc1
I agree. Although the standard Leela site continues to declare that the latest "stable" iteration is 30.0, I suspect that the Leela team just hasn't updated all their websites.
I'm running Leela 31.0 now with the BT4 net, and everything seems fine and "stable" to me.
-Steve-
I'm running Leela 31.0 now with the BT4 net, and everything seems fine and "stable" to me.
-Steve-
-
- Posts: 613
- Joined: Sun Aug 04, 2013 1:19 pm
Re: Finally official Lc0 v0.31.0-rc1
Yes and because of úberpriced Apple computer, everybody is buying thosewickedpotus wrote: ↑Tue Mar 26, 2024 10:08 pm But what we DO KNOW, is that LC0 is faster and better on most modern cheap nvidia laptops, than on an úberpriced Apple computer
Don't forget to mention that only some LC0 developers are tuning LC0 from the beginning for Nvidia GPUs, because they have only these GPUs and no AMD or Apple GPUs. So it's completely basic that LC0 is faster on Nvidia GPUs after all these years.
They could easily use some AMD GPUs and improve the speed of LC0 on those GPUs or on Intel GPUs.
The developers mentions years ago that they could easily improve the speed of LC0 by a factor of 5x but they are to much Nvidia guys.
= It's a developers / money problem and not an Apple, AMD or Intel problem.
-
- Posts: 613
- Joined: Sun Aug 04, 2013 1:19 pm
Re: Finally official Lc0 v0.31.0-rc1
They have problems to compile v0.31.0-rc1 on Apple devices https://github.com/LeelaChessZero/lc0/r ... 0.31.0-rc1Stephen Ham wrote: ↑Tue Mar 26, 2024 10:57 pm I agree. Although the standard Leela site continues to declare that the latest "stable" iteration is 30.0, I suspect that the Leela team just hasn't updated all their websites.
I'm running Leela 31.0 now with the BT4 net, and everything seems fine and "stable" to me.
-Steve-
See last message: https://github.com/LeelaChessZero/lc0/issues/1795
-
- Posts: 3601
- Joined: Thu Jun 07, 2012 11:02 pm
Re: Finally official Lc0 v0.31.0-rc1
Official releases of Lc0 in the past always came with a bundled net that was by definition the official net. Do they not do that now ?Jouni wrote: ↑Tue Mar 26, 2024 6:02 pm here https://github.com/LeelaChessZero/lc0/r ... 0.31.0-rc1. For Apple users: Multihead arch metal added. But I think biggest problem for Lc0 that there is no official net! So we will NEVER know how strong it really is.
-
- Posts: 12662
- Joined: Wed Mar 08, 2006 8:57 pm
- Location: Redmond, WA USA
Re: Finally official Lc0 v0.31.0-rc1
This is my quick verification test of 120 positions. I have much larger tests of similar quality, but nobody will have the patience to run them.
Here are the results for the new Lc0 (which was notoriously poor at tactics):
Medium net is
C:\Users\dannc\chess\lc0-v0.31.0-rc1\768x15x24h-t82-swa-7464000.pb.gz
03/26/2024 04:48 PM 170,339,855 768x15x24h-t82-swa-7464000.pb.gz
Mega net is
C:\Users\dannc\chess\lc0-v0.31.0-rc1\BT4-1024x15x32h-swa-6147500.pb.gz
03/26/2024 04:48 PM 335,897,546 BT4-1024x15x32h-swa-6147500.pb.gz
I was unable to get this network to function:
03/26/2024 04:47 PM 190,937,780 BT3-768x15x24h-swa-2790000.pb.gz
which is (according to my understanding) the one used by TCEC.
There is no screen activity nor any CPU effort by Lc0 when I use that network.
No idea why.
A suggestion for Lc0 team:
Scan the hardware and use what is learned to set the UCI parameters automatically. A thousand users will bless you.
Me at the top of the list.
Code: Select all
1B2b2k/7q/4n2p/3Np1PQ/8/1p6/5N1P/1K1B4 w - - bm g6; id "TTT2.014";
1B4n1/p2Pp2q/4Pp2/bP3Pp1/6Pp/7k/4B3/7K w - - bm b6; id "ACT1-B.157";
1K5k/1p4pP/1p3bP1/1Pp5/2P1P3/8/2P1P1P1/8 w - - bm Ka7; id "ACT4.053";
1K6/RP3pk1/4p3/2rbP1P1/8/8/P7/8 b - - bm Rc2; id "TTT2.031";
1R6/p4p2/2k1pP2/2nr4/2p2P1P/1p4P1/1Pq5/1R2Q1K1 b - - bm Nd7; id "TTT2.086";
1b2b1BN/2r1pp1n/2qpR2p/r1Np1Rnk/p5p1/Q5Kp/8/4B3 w - - bm Qf3; id "ACT3.089";
1b2r1r1/5p1k/3pb1pB/2pnP1Q1/q2P1N2/p1P5/1p3PP1/1R2R1K1 w - - bm Re4; id "TTT1.050";
1b3K2/1p6/P1P2pp1/kN2n3/8/P7/P7/8 w - - bm a7; id "ACT1-B.056";
1b6/r2p1k2/8/1R1b3R/8/1n1K4/8/1nB5 w - - bm Bb2; id "ACT1-B.057";
1k1r1b1r/ppnqp1pp/b3Rp2/4n3/2Pp1B2/N5PB/PP3P1P/3QNRK1 w - - bm Nd3 Nf3; id "TTT1.014";
1k2b3/4p1rP/1P1pK3/1Pp2P2/Rp2P3/p2P4/2P5/b7 w - - bm Ra8+; id "ACT1-B.012";
1k5b/p2K3P/P6p/p4pp1/1p6/4B3/2P5/8 w - - bm Bc1; id "ACT3.084";
1k5r/pbq2p2/4pP2/p1P3P1/5P1r/2p1Q3/P3B1P1/3R1RK1 w - - bm Rb1; id "TTT1.026";
1k6/1P2bp2/PpP1p2P/N2p4/2p5/1p4KP/2b1BpPn/5R2 w - - bm h7; id "ACT3.067";
1kb5/R7/1Pp4p/2N1pp2/2P2np1/2K5/P4r1P/3B4 b - - bm e4; id "ACT4.011";
1q1r4/pp1b1k1p/4pppP/Q2n4/1B1P4/1P3PP1/4R1K1/2R5 w - - bm Qxd5; id "TTT2.061";
1qQ4r/p2Kpk2/5p2/1B2PPp1/6Pp/7P/P7/8 w - - bm Qxh8; id "ACT3.033";
1qrr1b2/6p1/ppkn1P2/3pP3/1PP5/2BP1pP1/BQ3P2/1RRN1n1K b - - bm Be7; id "TTT1.011";
1r3rk1/1pp2p2/1b1pbq2/p3n3/P2NP1pP/1BP1N3/1PQ2PP1/3R1RK1 b - - bm Bxd4; id "TTT2.071";
1r4k1/3N2p1/8/P1q1p3/4P3/2p2RPP/4n1K1/Q7 b - - bm Qd4; id "TTT1.047";
1r4kb/4pr2/Qppp2q1/6p1/2PP1p2/P1N5/1P1RbPB1/R3B1K1 b - - bm Bh5; id "TTT2.060";
1r5k/5P1r/2pR2pp/7P/3P4/8/1B3K2/8 w - - bm Rd8+; id "ACT1-B.021";
1r5k/6pp/p2P2Q1/p3p3/b6q/3B2P1/3P1R2/6K1 w - - bm Qf5; id "ACT5.054";
1r6/1b6/1P2k3/4Np2/2R2PpP/2K1p1P1/8/8 w - - bm Rc7; id "TTT2.008";
1r6/4rnpk/1p6/p1p1Pp1q/3p1P1P/1P1Q1RR1/P3N1P1/6K1 w - - bm Rg5; id "TTT2.024";
1rb2rk1/2q3pp/p2ppbP1/n1nB4/1p1NP2P/2N1B3/PPP1Q3/2KR2R1 w - - bm Nf5; id "TTT1.036";
1rk1br1q/1pbp1n2/5p2/4P3/PpQR1PPN/4R1p1/1PK4n/6B1 w - - bm Rc3; id "ACT2.045";
1rrb1n2/1q5k/p3pPbN/3pP1Q1/1pnP2P1/8/4RPB1/1R5K w - - bm f4; id "TTT1.012";
2B4n/p4p1P/n1p2Pp1/p1P3P1/P1q1NN2/4K3/P2P4/5k2 w - - bm Bxa6; id "ACT4.087";
2B4n/p4p1P/r1p2Pp1/p1P3P1/P1q1NN2/4K3/3P4/5k2 w - - bm Bxa6; id "ACT5.010";
2K3b1/8/pPP1p1P1/2p1k2N/2P1pN1P/6PP/n1pQ2p1/1q3r2 w - - bm Kc7; id "ACT3.048";
2K5/4NB2/3k1B2/ppp2p2/2p1p1pp/6rr/2N5/8 w - - bm Ng6; id "ACT2.093";
2K5/p1p4k/p1p5/2p3p1/6Pb/6p1/p5P1/B7 w - - bm Kd7 Kd8; id "ACT4.094";
2N3q1/pp1N4/5K2/7k/7P/3Q4/1P6/r7 w - - bm Ne7; id "ACT5.100";
2N3qk/1p6/1p4pr/6N1/8/pR2Q3/2pp2K1/4b3 w - - bm Ne7; id "ACT2.070";
2R3bk/K3p1rp/6pb/4B1p1/6P1/8/8/8 w - - bm Ra8; id "ACT2.069";
2b2B1r/3p4/pbpp4/5nR1/kp2p3/N2P1p1p/PPP3Nn/K7 w - - bm b3+; id "ACT3.054";
2bR4/N3n1Pq/p2n1rB1/1PR2p2/KP1k4/3Ppr2/2Ppp3/4B3 w - - bm Bh4; id "ACT1-B.180";
2k5/1p2K3/1B6/8/8/P7/1N2b3/8 w - - bm Ba7; id "ACT5.075";
2k5/K1P1p3/1p1PpP2/2P5/4P3/8/pP6/8 w - - bm d7+; id "ACT1-B.150";
2k5/Q1np1N2/2p2Bpp/1p1P4/pPP1p2P/P5q1/8/1K6 w - - bm Qxc7+; id "ACT4.035";
2k5/p2p4/1P6/P3p3/1q2N3/2p5/8/2K4Q w - - bm Nd6+; id "ACT3.015";
2n1n1r1/1qp2R1B/bp4N1/1p1rP3/1p1b2p1/6R1/P2p1p2/3K1k1N w - - bm Nh4; id "ACT3.079";
2n3R1/5p2/P7/5P1k/4BK1p/2q3Pp/7P/8 w - - bm Rh8+; id "ACT1-B.067";
2n5/6pB/6p1/2N2k1b/2NBprpK/P3ppP1/rP1P4/5n2 w - - bm Bxg7; id "ACT2.053";
2nB1k2/6pp/3p2p1/n2K4/3Qp1N1/1Np4R/rb1p4/1Rq3r1 w - - bm Ke6; id "ACT2.074";
2nk4/6p1/2P5/1P6/p5K1/7p/1P5B/8 w - - bm Kf3; id "ACT5.096";
2q1k1b1/7p/4p1r1/Pp1pP2Q/1P1P2PB/2P5/1KP5/8 w - - bm Qxg6+; id "ACT4.066";
2r2rk1/1q3p2/3p2p1/2nPpPPp/ppR1P2P/5QN1/PP4K1/5R2 w - - bm Rxc5; id "TTT1.091";
2r3k1/b5pp/3n4/p2Pp3/P3P2q/2NP1P1P/3Q1pK1/1R3R2 b - - bm h5; id "TTT1.018";
2r5/3Np2q/r5Rp/3p1p2/2pNkB1b/p3P3/K3B2p/8 w - - bm Nf3; id "ACT1-B.162";
2r5/pb2kp2/6r1/8/3p2pP/1P4P1/P1B2P2/2Q3K1 b - - bm Bf3; id "ACT4.065";
2rq1rk1/5p2/4pP1B/4P3/1pp3Pp/p2P4/4Q1P1/5R1K w - - bm Rf5; id "TTT1.037";
3B4/2p1PpP1/2P3pN/R2P3k/p3q3/8/P5pK/8 w - - bm e8=Q e8=R; id "ACT2.041";
3K1R2/8/8/4B2p/8/6pp/6kp/7b w - - bm Bb8; id "ACT2.023";
3R1nk1/6pr/3K1b1p/7P/4P1P1/p6B/P7/8 w - - bm g5; id "ACT4.004";
3bn3/p6b/2R5/R1pN4/2p4p/6kP/ppn3P1/6BK w - - bm Raa6; id "ACT1-B.176";
3k4/8/2p5/4P2P/2pp4/b7/P1B1K3/8 w - - bm Kf3 Bb1; id "ACT5.062";
3k4/8/p4ppp/Pp1p4/1Pp2PPP/P1P5/8/3K4 w - - bm a4; id "ACT4.007";
3q2k1/1pr3pp/5P2/6pP/2p1B2r/4QP1P/R5bP/6K1 w - - bm Bd5+; id "ACT2.049";
3r1k2/pp4pp/5p2/3P1R2/5P2/1PP4P/5P2/1K6 b - - bm Rd6 b6; id "ACT4.024";
3r1rk1/1bqn1pp1/1p1bp3/p1n1N1PQ/2PB4/1PN5/3R1P1P/3R2K1 w - - bm Rd3; id "TTT1.033";
4B3/1pk1p3/q3p3/1nP1Pp2/PPQ2P2/PK2P3/4b3/8 w - - bm Qxb5; id "ACT3.021";
4K3/1B1p4/pn1Npb2/k3p3/P1pn4/P7/7B/3N4 w - - bm Bc6; id "ACT3.029";
4N3/pp3np1/6Rn/4p1pr/1P1p4/1K3p1p/1b3R2/1k6 w - - bm Rf1+; id "ACT1-B.160";
4R3/4P1Pp/2p1p3/7p/2p1B1rk/6p1/7p/5K2 w - - bm Bh1; id "ACT4.010";
4k2q/p4p1B/1pp3pP/2N2pP1/2P2P1N/8/PP2P3/2K5 w - - bm Nxg6; id "ACT5.011";
4k3/2p4p/8/p2n1p2/5P1P/4BP2/1PN1K1nP/8 b - - bm Ngxe3 Ndxe3; id "ACT3.055";
4k3/5pBr/4b1pP/3pP1P1/2Pp4/P7/1PP5/1K6 w - - bm cxd5; id "TTT1.078";
4n1k1/8/1b1P3P/1qB3Np/6PK/7Q/p2p1p2/8 w - - bm h7+; id "ACT5.049";
4n3/8/8/8/4pNp1/8/1p2K1pk/B2Q3b w - - bm Nh3; id "ACT3.005";
4r3/1ppk2p1/3pr2p/5B2/p3RP2/2P5/P2P2PP/7K w - - bm c4; id "TTT2.090";
4rkr1/1b2np2/q4Bp1/6P1/2Q5/pP6/K1P3B1/2R5 w - - bm Bxb7; id "ACT4.001";
5B2/2Q1p3/6R1/1Pr5/2n5/8/4P1bp/3K2kq w - - bm Qf4; id "ACT5.038";
5B2/8/Q2pq2p/2p3p1/5p1k/2P1nP1P/P5PK/8 b - - bm h5; id "ACT2.066";
5K1k/rp6/8/3Pppp1/4p1p1/3p4/5p2/1Q1n4 w - - bm Qb6; id "ACT3.017";
5K2/1Pk5/1pP4n/8/2Np3b/4pB2/8/8 w - - bm Ne5; id "ACT1-B.044";
5K2/kp3P1p/2p5/2Pp4/3P4/r7/p7/6R1 w - - bm Ke7 Ke8; id "TTT2.010";
5KR1/QB4pk/P2p3B/1p5p/3b4/5nPp/q2pp3/7n w - - bm Rh8+; id "ACT1-B.084";
5Kbk/4B1pp/3N1R2/p3P2n/3P2p1/2P2p2/1P1pP1r1/1B4n1 w - - bm Rh6; id "ACT2.047";
5N1r/5n1n/ppp3R1/5K2/7k/6p1/6PN/8 w - - bm Rg4+; id "ACT3.049";
5N2/rp6/pB1p1N2/pb2k1K1/3p2p1/1R2pp2/4B1n1/6bn w - - bm Rc3; id "ACT3.011";
5b2/8/q3n3/4k3/7R/1p4R1/bP2P2B/K7 w - - bm Ra4; id "ACT2.021";
5bb1/8/P1p1P3/BN2p1p1/1p4P1/pK1P1p2/P2R4/4r2k w - - bm Nd4; id "ACT3.010";
5brk/2p1p1p1/4P1p1/3K2P1/2p5/8/2P5/1B6 w - - bm c3; id "ACT3.022";
5k1q/1rp2P2/1p3p2/1P1N4/5K2/1P3PQ1/8/8 w - - bm Qg8+; id "ACT4.062";
5k1r/7p/2p3P1/2p5/1p4PK/p6R/7P/4B3 w - - bm Kg5; id "ACT3.047";
5k2/3b4/8/5PQ1/1KPP1PpP/PP5p/2R5/3q4 b - - bm Qe1+; id "ACT2.024";
5k2/p7/3p4/1p1B1p1p/1P3P2/P1rP3P/6P1/n1BK4 w - - bm Bb2; id "TTT2.040";
5kbQ/pp4p1/n3p2p/3p2r1/7N/2PPP2P/PP3PP1/4n2K b - - bm Rxg2 Nxg2; id "ACT5.029";
5qk1/pp4p1/4ppp1/2Q3P1/7P/P1P5/1P4P1/6K1 w - - bm Qxf8+; id "TTT2.073";
5r1k/p2b1pR1/3p3B/2p1pP2/4P2q/8/PPP5/1K3Q1n w - - bm Qg1; id "TTT1.052";
5r2/NBk4P/8/8/1P2K3/pp6/1p6/1N6 w - - bm Bd5; id "ACT5.095";
5rqk/4R2p/5Pp1/1p4P1/4Q3/2p3P1/6K1/8 w - - bm Qe5 Qc6 Qb7 Kg1; id "TTT1.064";
6B1/2N5/1n6/8/6p1/3b2p1/3K1prp/5kbQ w - - bm Na8; id "ACT2.090";
6B1/6K1/8/8/7N/8/P1PkN1pp/8 w - - bm Ng1; id "ACT5.069";
6K1/1N1N4/6p1/5p1p/1k6/1p1p4/1PP5/2B1q2B w - - bm Bd2+; id "ACT3.063";
6N1/8/2ppr3/2pNp3/2k5/5P2/1PP1K3/8 w - - bm b3+; id "ACT2.009";
6bk/5pp1/3N1p1b/8/6pN/5pPp/4pP1P/4K3 w - - bm Ng6+; id "ACT3.051";
6bk/8/5R2/8/7p/5NpP/4p1P1/5rBK b - - bm Kh7 Kg7; id "ACT3.060";
6k1/1N1p3p/8/4N3/r2pp1K1/p3p1B1/p1p5/6R1 w - - bm Bf4; id "ACT1-B.091";
6k1/5pp1/7p/pBP5/P5K1/1P4P1/3Q1PbP/2r1r3 b - - bm Rcd1; id "TTT2.062";
6k1/6pp/2b5/P6p/3P1B1P/5PK1/8/8 w - - bm Bh6; id "ACT3.095";
6k1/8/ppp2pP1/4P1p1/7p/PPP1K3/8/8 w - - bm e6; id "ACT3.093";
6k1/P7/1n5P/p1p1K3/8/1p6/1N6/8 w - - bm Ke6 Kf6; id "ACT4.009";
6k1/ppp2pp1/8/1n1P4/3N2P1/2P5/P4P1P/6K1 b - - bm Nxd4; id "TTT2.098";
6nN/2p5/P1k2b2/2Bpp3/8/3bR3/8/K7 w - - bm Rxd3; id "ACT4.084";
6r1/ppk2p1p/2b5/2R1pPP1/4P2P/8/4BK2/8 w - - bm Bc4; id "TTT1.092";
6rk/2p1np2/1r1pb3/1p2p3/p1PPP1p1/1P2N1Pq/P1BQRP2/3R2K1 w - - bm Nf1; id "TTT2.085";
6rk/4qpr1/1b1p4/pp2p3/P3Pp2/2pR1Np1/P1B2PPN/3Q1R1K b - - bm Rh7; id "TTT1.071";
6rq/6p1/pQp1p1p1/P1p1P2p/p1Pk3B/N7/3K4/1b5n w - - bm Nc2+; id "ACT1-B.094";
7k/1p2B2p/1P2K1p1/3p3P/8/4b1P1/1pP2N2/8 w - - bm Kf7; id "ACT3.070";
7k/2p3pp/4pP2/4pb2/3p4/B2N2P1/3Pq1PQ/6K1 w - - bm Qh6; id "ACT1-B.053";
7k/6R1/8/8/2P2P2/P7/pP6/3K3B w - - bm Rh7+; id "ACT2.038";
7k/8/7K/p1n1p1p1/B7/4N1p1/5B2/8 w - - bm Bg1; id "ACT5.064";
7k/8/pp2p1p1/bN1p2pb/1p4p1/1P6/6P1/4K2R w - - bm g3; id "ACT4.051";
7k/Bp1R1b1p/1p2pK1n/8/4P1P1/6pB/4p3/8 w - - bm Bb8; id "ACT3.008";
7k/R7/7p/p7/2B5/K7/PPp5/3b4 w - - bm Bg8; id "ACT3.019";
7k/n4p1b/2R1PP2/8/8/2B2r1n/pK6/8 w - - bm Rc8+; id "ACT3.043";
7k/pp6/2r2n2/4p1PN/1K2p1P1/1P1PP3/5PB1/8 b - - bm Nxh5; id "ACT3.100";
Code: Select all
--------------------------------------------------------------------------------
Analysis of H:\hardest.epd
Analyzing engine: lc0-v0.31.0-rc1-medium
3/26/2024 5:28:09 PM Level: 60 Seconds
Source : List of EPD/PGN files
H:\hardest.epd
Colours : White moves=True, Black moves=True
Direction : forward
Games (PGN) : From 1 to 999999
Moves : All moves
Positions (EPD) : All positions
Use Engine(s) : List
lc0-v0.31.0-rc1-medium
lc0-v0.31.0-rc1-mega
Use seconds per move : 60 Seconds per move
Analysis Lines : Minimum search depth=2
Activate abort analysis : True. Minimal analysis depth 9. Abort, if the correct move is held 220 additional ply
1) g5-g6 Kb1-b2
2) b5-b6 Be2-f1
3) Kb8-a7 Kb8-a7 * 0 Seconds
4) .. Rc5-c2 Kg7-g6
5) .. Nc5-d7 Nc5-d7 * 1 Second
6) Qa3-f3 Nc5-d3
7) Re1-e4 Re1-e4 * 0 Seconds
8) a6-a7 a6-a7 * 1 Second
9) Bc1-b2 Bc1-b2 * 0 Seconds
10) Ne1-d3, Ne1-f3 Ne1-d3 * 0 Seconds
11) Ra4-a8 Ra4-a8 * 36 Seconds
12) Be3-c1 Be3-c1 * 54 Seconds
13) Rd1-b1 Rd1-b1 * 1 Second
14) h6-h7 c6-c7
15) .. e5-e4 e5-e4 * 1 Second
16) Qa5xd5 Qa5xd5 * 2 Seconds
17) Qc8xh8 Qc8xh8 * 9 Seconds
18) .. Bf8-e7 g7xf6
19) .. Bb6xd4 Qf6xh4
20) .. Qc5-d4 Qc5-d4 * 40 Seconds
21) .. Be2-h5 Be2-h5 * 7 Seconds
22) Rd6-d8 Rd6-d8 * 1 Second
23) Qg6-f5 Qg6-f5 * 4 Seconds
24) Rc4-c7 Rc4-c7 * 3 Seconds
25) Rg3-g5 Rg3-g5 * 1 Second
26) Nd4-f5 Nd4-f5 * 27 Seconds
27) Re3-c3 Nh4-f5
28) f2-f4 f2-f4 * 0 Seconds
29) Bc8xa6 d2-d4
30) Bc8xa6 d2-d4
31) Kc8-c7 Kc8-c7 * 56 Seconds
32) Ne7-g6 Ne7-g6 * 0 Seconds
33) Kc8-d7, Kc8-d8 Kc8-d8 * 0 Seconds
34) Nc8-e7 Nc8-d6
35) Nc8-e7 Nc8-e7 * 2 Seconds
36) Rc8-a8 Rc8-a8 * 0 Seconds
37) b2-b3 Na3-c4
38) Be1-h4 Be1-h4 * 9 Seconds
39) Bb6-a7 Bb6-a7 * 0 Seconds
40) d6-d7 d6-d7 * 0 Seconds
41) Qa7xc7 Qa7-c5
42) Ne4-d6 Ne4-d6 * 0 Seconds
43) Ng6-h4 Rg3xg4
44) Rg8-h8 g3-g4
45) Bd4xg7 Bd4xg7 * 0 Seconds
46) Kd5-e6 Qd4-f2
47) Kg4-f3 Kg4-f3 * 0 Seconds
48) Qh5xg6 Bh4-f6
49) Rc4xc5 Rc4xc5 * 0 Seconds
50) .. h7-h5 h7-h5 * 5 Seconds
51) Nd4-f3 Rg6xa6
52) .. Bb7-f3 Bb7-f3 * 1 Second
53) Rf1-f5 Rf1-f5 * 0 Seconds
54) e7-e8Q, e7-e8R e7-e8Q * 12 Seconds
55) Be5-b8 Be5-d4
56) g4-g5 g4-g5 * 0 Seconds
57) Ra5-a6 Nd5-c3
58) Ke2-f3, Bc2-b1 Ke2-f3 * 5 Seconds
59) a3-a4 h4-h5
60) Be4-d5 Be4-d5 * 0 Seconds
61) .. Rd8-d6, b7-b6 Rd8-d6 * 0 Seconds
62) Rd2-d3 Rd2-d3 * 8 Seconds
63) Qc4xb5 Qc4xb5 * 0 Seconds
64) Bb7-c6 Nd1-b2
65) Rf2-f1 Rf2-f1 * 0 Seconds
66) Be4-h1 Be4-f3
67) Nh4xg6 Nh4xg6 * 0 Seconds
68) .. Ng2xe3, Nd5xe3 Ng2xe3 * 0 Seconds
69) c4xd5 c4xd5 * 13 Seconds
70) h6-h7 Qh3-c3
71) Nf4-h3 Nf4-h3 * 21 Seconds
72) c3-c4 c3-c4 * 9 Seconds
73) Bg2xb7 Qc4-h4
74) Qc7-f4 Qc7xc5
75) .. h6-h5 h6-h5 * 49 Seconds
76) Qb1-b6 Qb1-b6 * 2 Seconds
77) Nc4-e5 Nc4-e5 * 4 Seconds
78) Kf8-e7, Kf8-e8 Kf8-e7 * 0 Seconds
79) Rg8-h8 Rg8-h8 * 3 Seconds
80) Rf6-h6 Rf6-f7
81) Rg6-g4 Nf8-e6
82) Rb3-c3 Be2xb5
83) Rh4-a4 Rh4-a4 * 0 Seconds
84) Nb5-d4 Nb5-d4 * 1 Second
85) c2-c3 c2-c3 * 2 Seconds
86) Qg3-g8 Qg3-g8 * 0 Seconds
87) Kh4-g5 Kh4-g5 * 6 Seconds
88) .. Qd1-e1 Qd1-e1 * 0 Seconds
89) Bc1-b2 Bc1-b2 * 0 Seconds
90) .. Rg5xg2, Ne1xg2 Ne1xg2 * 0 Seconds
91) Qc5xf8 Qc5xf8 * 0 Seconds
92) Qf1-g1 Qf1-g1 * 3 Seconds
93) Bb7-d5 Na7-b5
94) Qe4-e5, Qe4-c6, Qe4-b7, Kg2-g1 Kg2-g1 * 1 Second
95) Nc7-a8 Nc7-a8 * 0 Seconds
96) Ne2-g1 Ne2-g1 * 6 Seconds
97) Bc1-d2 Bc1-d2 * 2 Seconds
98) b2-b3 b2-b3 * 3 Seconds
99) Nh4-g6 Nh4-g6 * 0 Seconds
100) .. Kh8-h7, Kh8-g7 Kh8-h7 * 0 Seconds
101) Bg3-f4 Bg3-f4 * 0 Seconds
102) .. Rc1-d1 Rc1-d1 * 0 Seconds
103) Bf4-h6 Bf4-h6 * 4 Seconds
104) e5-e6 e5xf6
105) Ke5-e6, Ke5-f6 Ke5-e6 * 1 Second
106) .. Nb5xd4 Nb5xd4 * 21 Seconds
107) Re3xd3 Re3xd3 * 0 Seconds
108) Be2-c4 Be2-c4 * 0 Seconds
109) Ne3-f1 Ne3-f1 * 12 Seconds
110) .. Rg7-h7 Rg7-h7 * 0 Seconds
111) Na3-c2 Na3-c2 * 9 Seconds
112) Ke6-f7 Ke6-f7 * 58 Seconds
113) Qh2-h6 Qh2-h6 * 0 Seconds
114) Rg7-h7 Rg7-b7
115) Bf2-g1 Bf2-g1 * 2 Seconds
116) g2-g3 g2-g3 * 0 Seconds
117) Ba7-b8 Ba7-b8 * 0 Seconds
118) Bc4-g8 Bc4-g8 * 0 Seconds
119) Rc6-c8 Rc6-c8 * 14 Seconds
120) .. Nf6xh5 Nf6xh5 * 0 Seconds
88 of 120 matching moves
3/26/2024 7:30:05 PM, Total time: 2:01:56 AM Rated time: 40:52 = 2452 Seconds
--------------------------------------------------------------------------------
Analysis of H:\hardest.epd
Analyzing engine: lc0-v0.31.0-rc1-mega
3/26/2024 5:28:09 PM Level: 60 Seconds
Source : List of EPD/PGN files
H:\hardest.epd
Colours : White moves=True, Black moves=True
Direction : forward
Games (PGN) : From 1 to 999999
Moves : All moves
Positions (EPD) : All positions
Use Engine(s) : List
lc0-v0.31.0-rc1-medium
lc0-v0.31.0-rc1-mega
Use seconds per move : 60 Seconds per move
Analysis Lines : Minimum search depth=2
Activate abort analysis : True. Minimal analysis depth 9. Abort, if the correct move is held 220 additional ply
1) g5-g6 Kb1-b2
2) b5-b6 Be2-f1
3) Kb8-a7 Kb8xb7
4) .. Rc5-c2 Kg7-g6
5) .. Nc5-d7 Nc5-d7 * 5 Seconds
6) Qa3-f3 Nc5-d3
7) Re1-e4 Re1-e4 * 0 Seconds
8) a6-a7 a6-a7 * 2 Seconds
9) Bc1-b2 Bc1-b2 * 0 Seconds
10) Ne1-d3, Ne1-f3 Ne1-d3 * 0 Seconds
11) Ra4-a8 Ra4-a8 * 19 Seconds
12) Be3-c1 Be3-c1 * 12 Seconds
13) Rd1-b1 Rd1-b1 * 12 Seconds
14) h6-h7 c6-c7
15) .. e5-e4 e5-e4 * 0 Seconds
16) Qa5xd5 Qa5xd5 * 2 Seconds
17) Qc8xh8 Qc8xh8 * 1 Second
18) .. Bf8-e7 Bf8-e7 * 0 Seconds
19) .. Bb6xd4 Bb6xd4 * 1 Second
20) .. Qc5-d4 Qc5-d4 * 59 Seconds
21) .. Be2-h5 Be2-h5 * 4 Seconds
22) Rd6-d8 Rd6-d8 * 0 Seconds
23) Qg6-f5 Qg6-f5 * 1 Second
24) Rc4-c7 Rc4-c7 * 12 Seconds
25) Rg3-g5 Rg3-g5 * 2 Seconds
26) Nd4-f5 Nd4-f5 * 10 Seconds
27) Re3-c3 Re3-c3 * 10 Seconds
28) f2-f4 f2-f4 * 0 Seconds
29) Bc8xa6 Bc8xa6 * 0 Seconds
30) Bc8xa6 Bc8xa6 * 53 Seconds
31) Kc8-c7 Kc8-c7 * 1 Second
32) Ne7-g6 Ne7-g6 * 0 Seconds
33) Kc8-d7, Kc8-d8 Kc8-d7 * 0 Seconds
34) Nc8-e7 Nc8-e7 * 29 Seconds
35) Nc8-e7 Nc8-e7 * 16 Seconds
36) Rc8-a8 Rc8-a8 * 0 Seconds
37) b2-b3 b2-b3 * 7 Seconds
38) Be1-h4 Be1-h4 * 1 Second
39) Bb6-a7 Bb6-a7 * 0 Seconds
40) d6-d7 d6-d7 * 0 Seconds
41) Qa7xc7 Qa7-c5
42) Ne4-d6 Ne4-d6 * 0 Seconds
43) Ng6-h4 Rg3xg4
44) Rg8-h8 g3-g4
45) Bd4xg7 Bd4xg7 * 0 Seconds
46) Kd5-e6 Rb1xc1
47) Kg4-f3 Kg4-f3 * 0 Seconds
48) Qh5xg6 Bh4-f6
49) Rc4xc5 Rc4xc5 * 1 Second
50) .. h7-h5 h7-h5 * 9 Seconds
51) Nd4-f3 Rg6xa6
52) .. Bb7-f3 Bb7-f3 * 0 Seconds
53) Rf1-f5 Rf1-f5 * 0 Seconds
54) e7-e8Q, e7-e8R e7-e8Q * 7 Seconds
55) Be5-b8 Be5-d4
56) g4-g5 g4-g5 * 0 Seconds
57) Ra5-a6 Nd5-c3
58) Ke2-f3, Bc2-b1 Ke2-f3 * 2 Seconds
59) a3-a4 h4-h5
60) Be4-d5 Be4-d5 * 0 Seconds
61) .. Rd8-d6, b7-b6 Rd8-d6 * 0 Seconds
62) Rd2-d3 Rd2-d3 * 2 Seconds
63) Qc4xb5 Qc4xb5 * 0 Seconds
64) Bb7-c6 Nd1-b2
65) Rf2-f1 Rf2-f1 * 3 Seconds
66) Be4-h1 Be4-f3
67) Nh4xg6 Nh4xg6 * 0 Seconds
68) .. Ng2xe3, Nd5xe3 Ng2xe3 * 0 Seconds
69) c4xd5 c4xd5 * 20 Seconds
70) h6-h7 Qh3-c3
71) Nf4-h3 Nf4xg2
72) c3-c4 c3-c4 * 0 Seconds
73) Bg2xb7 Qc4-h4
74) Qc7-f4 Qc7xc5
75) .. h6-h5 h6-h5 * 16 Seconds
76) Qb1-b6 Qb1-b6 * 6 Seconds
77) Nc4-e5 Nc4-e5 * 0 Seconds
78) Kf8-e7, Kf8-e8 Kf8-e7 * 0 Seconds
79) Rg8-h8 Rg8-h8 * 4 Seconds
80) Rf6-h6 Rf6-f7
81) Rg6-g4 Nf8-e6
82) Rb3-c3 Be2xb5
83) Rh4-a4 Rh4-a4 * 0 Seconds
84) Nb5-d4 Nb5-d4 * 1 Second
85) c2-c3 c2-c3 * 2 Seconds
86) Qg3-g8 Qg3-g8 * 0 Seconds
87) Kh4-g5 Kh4-g5 * 0 Seconds
88) .. Qd1-e1 Qd1-e1 * 0 Seconds
89) Bc1-b2 Bc1-b2 * 0 Seconds
90) .. Rg5xg2, Ne1xg2 Ne1xg2 * 0 Seconds
91) Qc5xf8 Qc5xf8 * 0 Seconds
92) Qf1-g1 Qf1-g1 * 10 Seconds
93) Bb7-d5 Bb7-d5 * 31 Seconds
94) Qe4-e5, Qe4-c6, Qe4-b7, Kg2-g1 Kg2-g1 * 0 Seconds
95) Nc7-a8 Nc7-a8 * 0 Seconds
96) Ne2-g1 Ne2-g1 * 1 Second
97) Bc1-d2 Bc1-d2 * 6 Seconds
98) b2-b3 b2-b3 * 4 Seconds
99) Nh4-g6 Nh4-g6 * 0 Seconds
100) .. Kh8-h7, Kh8-g7 Kh8-h7 * 1 Second
101) Bg3-f4 Bg3-f4 * 3 Seconds
102) .. Rc1-d1 Rc1-d1 * 0 Seconds
103) Bf4-h6 Bf4-h6 * 5 Seconds
104) e5-e6 e5-e6 * 11 Seconds
105) Ke5-e6, Ke5-f6 Ke5-e6 * 2 Seconds
106) .. Nb5xd4 Nb5xd4 * 1 Second
107) Re3xd3 Re3xd3 * 1 Second
108) Be2-c4 Be2-c4 * 0 Seconds
109) Ne3-f1 Ne3-f1 * 4 Seconds
110) .. Rg7-h7 Rg7-h7 * 0 Seconds
111) Na3-c2 Na3-c2 * 10 Seconds
112) Ke6-f7 Ke6-f7 * 57 Seconds
113) Qh2-h6 Qh2-h6 * 0 Seconds
114) Rg7-h7 Rg7-h7 * 2 Seconds
115) Bf2-g1 Bf2-g1 * 0 Seconds
116) g2-g3 g2-g3 * 0 Seconds
117) Ba7-b8 Ba7-b8 * 0 Seconds
118) Bc4-g8 Bc4-g8 * 0 Seconds
119) Rc6-c8 Rc6-c8 * 3 Seconds
120) .. Nf6xh5 Nf6xh5 * 1 Second
96 of 120 matching moves
3/26/2024 9:32:00 PM, Total time: 4:03:51 AM Rated time: 32:05 = 1925 Seconds
C:\Users\dannc\chess\lc0-v0.31.0-rc1\768x15x24h-t82-swa-7464000.pb.gz
03/26/2024 04:48 PM 170,339,855 768x15x24h-t82-swa-7464000.pb.gz
Mega net is
C:\Users\dannc\chess\lc0-v0.31.0-rc1\BT4-1024x15x32h-swa-6147500.pb.gz
03/26/2024 04:48 PM 335,897,546 BT4-1024x15x32h-swa-6147500.pb.gz
I was unable to get this network to function:
03/26/2024 04:47 PM 190,937,780 BT3-768x15x24h-swa-2790000.pb.gz
which is (according to my understanding) the one used by TCEC.
There is no screen activity nor any CPU effort by Lc0 when I use that network.
No idea why.
A suggestion for Lc0 team:
Scan the hardware and use what is learned to set the UCI parameters automatically. A thousand users will bless you.
Me at the top of the list.
Taking ideas is not a vice, it is a virtue. We have another word for this. It is called learning.
But sharing ideas is an even greater virtue. We have another word for this. It is called teaching.
But sharing ideas is an even greater virtue. We have another word for this. It is called teaching.
-
- Posts: 582
- Joined: Wed May 10, 2006 7:28 pm
- Location: Birmingham, England
Re: Finally official Lc0 v0.31.0-rc1
Many thanks.Dann Corbit wrote: ↑Wed Mar 27, 2024 11:49 am
<snip>
Here are the results for the new Lc0 (which was notoriously poor at tactics):
<snip>
So, what is the comparison to the old Lc0? In your own words.
Marek Soszynski
-
- Posts: 268
- Joined: Thu Jul 21, 2022 12:30 am
- Full name: Chesskobra
Re: Finally official Lc0 v0.31.0-rc1
What network should I use on an i7, 13 gen, desktop with 32 GB RAM?