Sergio NN 40x512

[Eduard]

A 512x40 test is running here:

https://www.comp.nus.edu.sg/~sergio-v/test/

[Hai]

A 512x40 test is running here:

https://www.comp.nus.edu.sg/~sergio-v/test/

Thank you very much. It looks much better running LC0 Sergio 40x512 net with my 2x RTX 2080 Ti.

[Nordlandia]

A 512x40 test is running here:

https://www.comp.nus.edu.sg/~sergio-v/test/

Thank you very much. It looks much better running LC0 Sergio 40x512 net with my 2x RTX 2080 Ti.

This is amazing

Hello Hai. What speed do you get in this FEN position. I'm asking because i want to compare the speed difference. I'm running 2x RTX 2080 Super OC (ASUS ROG STRIX) Identical cards.

[d]1q6/1r2k1p1/4pp1p/1P1b1P2/3Q4/7P/4B1P1/2R3K1 w

Analyse for 1 minute with 512x40 net.

I'm using following settings tweaked from the defaults:

<------->
Roundrobin
(backend=cudnn-fp16,gpu=0),(backend=cudnn-fp16,gpu=1)
CPuct: 3.4
CPuctBase: 10000
MiniBatchSize: 512
<------->

Tablebases disabled.

Assuming you have two identical cards try same settings or multiplexing if they're not intentical and run that test for speed comparison

1 minute search:

1q6/1r2k1p1/4pp1p/1P1b1P2/3Q4/7P/4B1P1/2R3K1 w - - 0 1

Analysis by Lc0 v0.23.2+git.c8d9095:

1.Qg4 exf5 2.Qxg7+ Ke6 3.Qg6 Qe5 4.Qe8+ Re7 5.Qg8+ Rf7 6.Re1 Kd6 7.b6 Rd7 8.Qf8+ Re7 9.Qb8+ Ke6 10.Qc8+ Kd6 11.Kh1 Bxg2+ 12.Kxg2 Qe4+ 13.Kf2 Qe3+ 14.Kf1
White is better: +/- (1.14) Depth: 12/35 00:01:06 329kN
(12.01.2020)

2x 2080 Super output: speed 5544 N/s after about one minute.

[Hai]

A 512x40 test is running here:

https://www.comp.nus.edu.sg/~sergio-v/test/

Thank you very much. It looks much better running LC0 Sergio 40x512 net with my 2x RTX 2080 Ti.

This is amazing

Hello Hai. What speed do you get in this FEN position. I'm asking because i want to compare the speed difference. I'm running 2x RTX 2080 Super OC (ASUS ROG STRIX) Identical cards.

[d]1q6/1r2k1p1/4pp1p/1P1b1P2/3Q4/7P/4B1P1/2R3K1 w

Analyse for 1 minute with 512x40 net.

I'm using following settings tweaked from the defaults:

<------->
Roundrobin
(backend=cudnn-fp16,gpu=0),(backend=cudnn-fp16,gpu=1)
CPuct: 3.4
CPuctBase: 10000
MiniBatchSize: 512
<------->

Tablebases disabled.

Assuming you have two identical cards try same settings or multiplexing if they're not intentical and run that test for speed comparison

1 minute search:

1q6/1r2k1p1/4pp1p/1P1b1P2/3Q4/7P/4B1P1/2R3K1 w - - 0 1

Analysis by Lc0 v0.23.2+git.c8d9095:

1.Qg4 exf5 2.Qxg7+ Ke6 3.Qg6 Qe5 4.Qe8+ Re7 5.Qg8+ Rf7 6.Re1 Kd6 7.b6 Rd7 8.Qf8+ Re7 9.Qb8+ Ke6 10.Qc8+ Kd6 11.Kh1 Bxg2+ 12.Kxg2 Qe4+ 13.Kf2 Qe3+ 14.Kf1
White is better: +/- (1.14) Depth: 12/35 00:01:06 329kN
(12.01.2020)

2x 2080 Super output: speed 5544 N/s after about one minute.

Hi Nordlandia,

we have nearly the same N/s and kN.

I'm using default settings.
And also:
Roundrobin
(backend=cudnn-fp16,gpu=0),(backend=cudnn-fp16,gpu=1)
+ 64 GB RAM

You have 2x 2080 Super and OC.
I have 2x 2080 Ti without Super and without OC.
Also I use power limit of 95% per GPU.
Also I use temperature limit of 70 per GPU.
So I've got lower power consumption, lower heat and it's also really quiet .
But GPU load is jumping between 60 to 90% on both GPUs when running your position.

When I run chess start position I have GPU load of 93% on both GPUs.
And CPU heat is 50.

[mwyoung]

A 512x40 test is running here:

https://www.comp.nus.edu.sg/~sergio-v/test/

Thank you very much. It looks much better running LC0 Sergio 40x512 net with my 2x RTX 2080 Ti.

This is amazing

Hello Hai. What speed do you get in this FEN position. I'm asking because i want to compare the speed difference. I'm running 2x RTX 2080 Super OC (ASUS ROG STRIX) Identical cards.

[d]1q6/1r2k1p1/4pp1p/1P1b1P2/3Q4/7P/4B1P1/2R3K1 w

Analyse for 1 minute with 512x40 net.

I'm using following settings tweaked from the defaults:

<------->
Roundrobin
(backend=cudnn-fp16,gpu=0),(backend=cudnn-fp16,gpu=1)
CPuct: 3.4
CPuctBase: 10000
MiniBatchSize: 512
<------->

Tablebases disabled.

Assuming you have two identical cards try same settings or multiplexing if they're not intentical and run that test for speed comparison

1 minute search:

1q6/1r2k1p1/4pp1p/1P1b1P2/3Q4/7P/4B1P1/2R3K1 w - - 0 1

Analysis by Lc0 v0.23.2+git.c8d9095:

1.Qg4 exf5 2.Qxg7+ Ke6 3.Qg6 Qe5 4.Qe8+ Re7 5.Qg8+ Rf7 6.Re1 Kd6 7.b6 Rd7 8.Qf8+ Re7 9.Qb8+ Ke6 10.Qc8+ Kd6 11.Kh1 Bxg2+ 12.Kxg2 Qe4+ 13.Kf2 Qe3+ 14.Kf1
White is better: +/- (1.14) Depth: 12/35 00:01:06 329kN
(12.01.2020)

2x 2080 Super output: speed 5544 N/s after about one minute.

Hi Nordlandia,

we have nearly the same N/s and kN.

I'm using default settings.
And also:
Roundrobin
(backend=cudnn-fp16,gpu=0),(backend=cudnn-fp16,gpu=1)
+ 64 GB RAM

You have 2x 2080 Super and OC.
I have 2x 2080 Ti without Super and without OC.
Also I use power limit of 95% per GPU.
Also I use temperature limit of 70 per GPU.
So I've got lower power consumption, lower heat and it's also really quiet .
But GPU load is jumping between 60 to 90% on both GPUs when running your position.

When I run chess start position I have GPU load of 93% on both GPUs.
And CPU heat is 50.

I guess the question is does this network 512x40 need 1, 2,3,4....High end graphic cards. Assuming it is a good net to be the best. Have you run any game test with this Fat Network. I will download it for a full round of testing. Have we now reached the limits of 1 RTX graphics card?

I can test if this Fat Network can top the best on 1 RTX....

[OneTrickPony]

Guys, right now those new test networks are too early in the training to say anything about potential strength.

[mwyoung]

Guys, right now those new test networks are too early in the training to say anything about potential strength.

It it ready or still training? Regardless I will see what it can do now.

[Hai]

A 512x40 test is running here:

https://www.comp.nus.edu.sg/~sergio-v/test/

Thank you very much. It looks much better running LC0 Sergio 40x512 net with my 2x RTX 2080 Ti.

This is amazing

Hello Hai. What speed do you get in this FEN position. I'm asking because i want to compare the speed difference. I'm running 2x RTX 2080 Super OC (ASUS ROG STRIX) Identical cards.

[d]1q6/1r2k1p1/4pp1p/1P1b1P2/3Q4/7P/4B1P1/2R3K1 w

Analyse for 1 minute with 512x40 net.

I'm using following settings tweaked from the defaults:

<------->
Roundrobin
(backend=cudnn-fp16,gpu=0),(backend=cudnn-fp16,gpu=1)
CPuct: 3.4
CPuctBase: 10000
MiniBatchSize: 512
<------->

Tablebases disabled.

Assuming you have two identical cards try same settings or multiplexing if they're not intentical and run that test for speed comparison

1 minute search:

1q6/1r2k1p1/4pp1p/1P1b1P2/3Q4/7P/4B1P1/2R3K1 w - - 0 1

Analysis by Lc0 v0.23.2+git.c8d9095:

1.Qg4 exf5 2.Qxg7+ Ke6 3.Qg6 Qe5 4.Qe8+ Re7 5.Qg8+ Rf7 6.Re1 Kd6 7.b6 Rd7 8.Qf8+ Re7 9.Qb8+ Ke6 10.Qc8+ Kd6 11.Kh1 Bxg2+ 12.Kxg2 Qe4+ 13.Kf2 Qe3+ 14.Kf1
White is better: +/- (1.14) Depth: 12/35 00:01:06 329kN
(12.01.2020)

2x 2080 Super output: speed 5544 N/s after about one minute.

Hi Nordlandia,

we have nearly the same N/s and kN.

I'm using default settings.
And also:
Roundrobin
(backend=cudnn-fp16,gpu=0),(backend=cudnn-fp16,gpu=1)
+ 64 GB RAM

You have 2x 2080 Super and OC.
I have 2x 2080 Ti without Super and without OC.
Also I use power limit of 95% per GPU.
Also I use temperature limit of 70 per GPU.
So I've got lower power consumption, lower heat and it's also really quiet .
But GPU load is jumping between 60 to 90% on both GPUs when running your position.

When I run chess start position I have GPU load of 93% on both GPUs.
And CPU heat is 50.

I guess the question is does this network 512x40 need 1, 2,3,4....High end graphic cards. Assuming it is a good net to be the best. Have you run any game test with this Fat Network. I will download it for a full round of testing. Have we now reached the limits of 1 RTX graphics card?

I can test if this Fat Network can top the best on 1 RTX....

I'm interested in chess analysis and not in engine-super-lightning-bullet-tc-chess .

Like for every big net, at the beginning you should have 2x RTX GPUs, 2080 Ti if possible.
Like for every big net, later one RTX GPU, 2080 Ti if possible, will be enough.
-30x384-T60-2, 131 MB, : Ongoing, at 3rd LR (0.002) - That explains why 1 RTX GPU is enough.
-40x512, 318 MB, already started to train = less knowledge at the moment = to compensate it you sould use 2x RTX GPUs. But it also benefit much more from 3 or 4 RTX GPUs than the other nets.

-I can do now much longer analysis, when using the LC0 Sergio 40x512 net.
-I need much lesser RAM when running the 40x512 net.
-It is also possible to run it much longer than a 30x384 net before I run out of memory with my 64 GB RAM.

I expect the 40x512 net to get ~60 points in the ERET Test, because it's an untrained net at the moment.
Of course when using longer time control, the result will be better.
I have tested it also for analysis and games playing style.
It can already find some very difficult moves when using it on some fen positions to analyse. That's really amazing.
Strategically it is weaker at the moment and the positional style is also weaker at the moment.
But tactics are better .

[mwyoung]

A 512x40 test is running here:

https://www.comp.nus.edu.sg/~sergio-v/test/

Thank you very much. It looks much better running LC0 Sergio 40x512 net with my 2x RTX 2080 Ti.

This is amazing

Hello Hai. What speed do you get in this FEN position. I'm asking because i want to compare the speed difference. I'm running 2x RTX 2080 Super OC (ASUS ROG STRIX) Identical cards.

[d]1q6/1r2k1p1/4pp1p/1P1b1P2/3Q4/7P/4B1P1/2R3K1 w

Analyse for 1 minute with 512x40 net.

I'm using following settings tweaked from the defaults:

<------->
Roundrobin
(backend=cudnn-fp16,gpu=0),(backend=cudnn-fp16,gpu=1)
CPuct: 3.4
CPuctBase: 10000
MiniBatchSize: 512
<------->

Tablebases disabled.

Assuming you have two identical cards try same settings or multiplexing if they're not intentical and run that test for speed comparison

1 minute search:

1q6/1r2k1p1/4pp1p/1P1b1P2/3Q4/7P/4B1P1/2R3K1 w - - 0 1

Analysis by Lc0 v0.23.2+git.c8d9095:

1.Qg4 exf5 2.Qxg7+ Ke6 3.Qg6 Qe5 4.Qe8+ Re7 5.Qg8+ Rf7 6.Re1 Kd6 7.b6 Rd7 8.Qf8+ Re7 9.Qb8+ Ke6 10.Qc8+ Kd6 11.Kh1 Bxg2+ 12.Kxg2 Qe4+ 13.Kf2 Qe3+ 14.Kf1
White is better: +/- (1.14) Depth: 12/35 00:01:06 329kN
(12.01.2020)

2x 2080 Super output: speed 5544 N/s after about one minute.

Hi Nordlandia,

we have nearly the same N/s and kN.

I'm using default settings.
And also:
Roundrobin
(backend=cudnn-fp16,gpu=0),(backend=cudnn-fp16,gpu=1)
+ 64 GB RAM

You have 2x 2080 Super and OC.
I have 2x 2080 Ti without Super and without OC.
Also I use power limit of 95% per GPU.
Also I use temperature limit of 70 per GPU.
So I've got lower power consumption, lower heat and it's also really quiet .
But GPU load is jumping between 60 to 90% on both GPUs when running your position.

When I run chess start position I have GPU load of 93% on both GPUs.
And CPU heat is 50.

I guess the question is does this network 512x40 need 1, 2,3,4....High end graphic cards. Assuming it is a good net to be the best. Have you run any game test with this Fat Network. I will download it for a full round of testing. Have we now reached the limits of 1 RTX graphics card?

I can test if this Fat Network can top the best on 1 RTX....

I'm interested in chess analysis and not in engine-super-lightning-bullet-tc-chess .

Like for every big net, at the beginning you should have 2x RTX GPUs, 2080 Ti if possible.
Like for every big net, later one RTX GPU, 2080 Ti if possible, will be enough.
-30x384-T60-2, 131 MB, : Ongoing, at 3rd LR (0.002) - That explains why 1 RTX GPU is enough.
-40x512, 318 MB, already started to train = less knowledge at the moment = to compensate it you sould use 2x RTX GPUs. But it also benefit much more from 3 or 4 RTX GPUs than the other nets.

-I can do now much longer analysis, when using the LC0 Sergio 40x512 net.
-I need much lesser RAM when running the 40x512 net.
-It is also possible to run it much longer than a 30x384 net before I run out of memory with my 64 GB RAM.

I expect the 40x512 net to get ~60 points in the ERET Test, because it's an untrained net at the moment.
Of course when using longer time control, the result will be better.
I have tested it also for analysis and games playing style.
It can already find some very difficult moves when using it on some fen positions to analyse. That's really amazing.
Strategically it is weaker at the moment and the positional style is also weaker at the moment.
But tactics are better .

Yes I don't know why the testing sites only play super lightning blitz tournament on only default settings and not the best settings except for me to rate NN engines. I am glad I do not. I guess they stupidly assume that all NN are created equal in regards to the default settings.
I have found the bigger nets to be better at shorter time controls and longer time controls. NPS is a very poor gauge on what is the best NN at any time control. I have show this in my testing.

And I assume you know that every position in a chess game is a series of positions to analyze. And the engine that is most correct general wins the game. That is why we test. General the strongest chess playing engine will give you the best analysis.

Lets see where Sergio 40x512 stands in regards to the other engines and NN.

[Nordlandia]

In my case 2x 2080 Super is not strong enough to benefit it to the fullest. I think CCC can benefit from it.

TimeControl "3600+30"

[pgn][Event "i7-5930K 4.5GHz"]
[Site "i7-5930K 4.5GHz"]
[Date "2020.01.12"]
[Round "?"]
[White "lc0"]
[Black "Stockfish 10.01.20 Ultimaiq BM"]
[Result "0-1"]
[SetUp "1"]
[FEN "rnbqrbnk/pppp1ppp/4p3/8/8/3P4/PPP1PPPP/KNBRQBNR w - - 0 11"]
[PlyCount "81"]
[EventDate "2020.??.??"]
[TimeControl "3600+30"]

11. f4 {0.38/11 54} c5 {-0.28/35 137s} 12. g4 {0.41/12 124s} b5 {-0.34/37 278s}
13. Nf3 {0.45/13 64s} Nc6 {-0.30/32 33} 14. h4 {0.45/12 100s} a5 {-0.31/34 191s
} 15. h5 {0.43/13 237s} Bb7 {-0.42/34 292s} 16. e4 {0.64/12 73s} a4 {-0.17/33
96s} 17. h6 {0.69/13 64s} gxh6 {-0.40/33 80s} 18. g5 {0.67/14 49} a3 {-0.30/35
298s} 19. Nxa3 {0.61/16 199s} b4 {-0.80/34 95s} 20. Nc4 {0.65/20 33} Qc7 {
-0.75/37 171s} 21. b3 {1.11/18 199s} Bg7+ {0.00/33 31} 22. e5 {1.08/22 31} Nce7
{0.00/36 50} 23. gxh6 {0.85/13 243s} Nxh6 {0.00/36 53} 24. Bg2 {0.61/15 146s}
d5 {0.00/36 61s} 25. Ne3 {0.65/13 55} Nhf5 {0.00/40 186s} 26. Nxf5 {0.60/15 82s
} Nxf5 {0.45/36 49} 27. Bh3 {0.62/16 52} c4 {0.32/42 135s} 28. a4 {-0.05/20
130s} cxb3 {0.00/44 125s} 29. cxb3 {-0.21/21 0.66s} Bc6 {0.00/47 60s} 30. Kb1 {
-0.21/20 0.20s} Bxa4 {6.48/36 48} 31. bxa4 {-0.74/17 140s} Qc6 {7.38/37 56} 32.
Bxf5 {-0.89/19 208s} Rec8 {8.04/38 50} 33. Qf2 {-2.38/15 301s} Qxa4 {9.08/36 48
} 34. Qb2 {-2.45/15 114s} exf5 {9.79/36 62s} 35. Rd2 {-2.52/14 113s} Rc3 {
10.32/36 63s} 36. Rdh2 {-2.55/13 64s} Rb3 {22.50/34 53} 37. Nd4 {-2.59/14 87s}
Qa2+ {67.20/40 53} 38. Kc2 {-2.18/14 0.042s} Rxb2+ {148.78/43 111s} 39. Bxb2 {
-3.25/13 163s} b3+ {148.85/41 53} 40. Kd1 {-3.64/12 259s} Qb1+ {#51/40 49} 41.
Ke2 {-3.86/11 128s} Qxb2+ {#39/65 48} 42. Ke3 {-3.90/10 0.47s} Qc3 {#37/75 49}
43. Rxh7+ {-3.94/9 141s} Kg8 {#35/1 0} 44. Nxf5 {-3.82/8 122s} Bxe5 {#27/62 38}
45. Kf3 {-4.32/8 164s} Qxd3+ {#19/76 50} 46. Kg4 {-4.94/7 46} Qe2+ {#17/87 51}
47. Kg3 {-5.92/7 113s} Bxf4+ {#15/105 50} 48. Kh4 {-8.75/7 97s} Kxh7 {#11/193
45} 49. Nd4 {-12.73/6 90s} Qg2 {#9/245 13} 50. Rg1 {-14.26/4 60s} Qxg1 {
#7/245 0.017s} 51. Kh5 {adjudication -128.00/2 96s, Black wins by adjudication}
0-1[/pgn]