CCC has serious hardware update!

[Pedro]

With such powerful gpus, as well as the cpu, it would be cool if the chess.com team improved Leela's networks to the next level of excellence. Or the Komodo neural network, since it is owned by chess.com.

[Nay Lin Tun]

With such powerful gpus, as well as the cpu, it would be cool if the chess.com team improved Leela's networks to the next level of excellence. Or the Komodo neural network, since it is owned by chess.com.

It is not due to programming fault for not being able to utilize additional resources. It is due to incredibly smart features ( highly selectively search features ) of current top engines that doesnt require much resources to get the right choice.


As average branching factor of selective search of Stockfish is around 2.2 with aggressive pruning, around 16 cores of Stockfish is generally enough to play high quality chess. If you have 256 cores, SF SMP will use it for non priority trees. Therefore, elo gain is really diminished.

On the other hand, 200+ cores to ancient engines without pruning like deep blue will be massive elo gain, compared to Stockfish.

[mwyoung]

With such powerful gpus, as well as the cpu, it would be cool if the chess.com team improved Leela's networks to the next level of excellence. Or the Komodo neural network, since it is owned by chess.com.

It is not due to programming fault for not being able to utilize additional resources. It is due to incredibly smart features ( highly selectively search features ) of current top engines that doesnt require much resources to get the right choice.


As average branching factor of selective search of Stockfish is around 2.2 with aggressive pruning, around 16 cores of Stockfish is generally enough to play high quality chess. If you have 256 cores, SF SMP will use it for non priority trees. Therefore, elo gain is really diminished.

On the other hand, 200+ cores to ancient engines without pruning like deep blue will be massive elo gain, compared to Stockfish.

Correct. This is a type B search issue. That is why they are called fast and dumb.

And it seems you need even less then 16 cores to play high quality chess with NNUE. Current standings with a 10x NPS advantage and a 2 core vs. 32 thread advantage. The Elo difference right now is only 35 ELO!! after 30 games. And this would be about the best case. Since Stockfish is only splitting the search into 2 threads.

[AndrewGrant]

And it seems you need even less then 16 cores to play high quality chess with NNUE. Current standings with a 10x NPS advantage and a 2 core vs. 32 thread advantage. The Elo difference right now is only 35 ELO!! after 30 games. And this would be about the best case. Since Stockfish is only splitting the search into 2 threads.

Garbage results, which lead to poor understanding of how engines scale. Misinformation at best, deceit at worst.
Unless you are saying that you are giving the 2 cores 10x the thinking time? Which if you are, ignore my entire post.

Typically speaking, a core doubling from 1 to 2 starts at about +70 elo for an engine like Stockfish, using a balanced opening book. As you continue the doubling, that begins to drop. I am not sure exactly of the rate, but it appears logarithmic. However, there was a test with 192 cores vs 384 threads. Note that this is not a typical core doubling, but making use of the hyperthreads on a core. IE, it is not as good as doubling the number of cores. This resulted in about +20 elo after a few thousand games. This test is on fishtest somewhere, and is public record.

So you start your initial core doubling at +70 per doubling, and by the end a doubling is worth at least 20 elo. log2(256) = 8. Hyperthreads are being used. So it can be derived from this knowledge that 1 core SF vs 256 thread SF is roughly 7 * ((70 + 20) / 2) + 20, which is 300+ elo.

Will you see this over the board? Probably not. Elo differences at those extremes are poorly defined by the elo curve, which is well adjusted for similarly skilled opponents, but fails at the extremes due to the nature of the games it is employed in. The issue in this case being that SF is already so strong that, if there were a skill cap, SF is closer to it than any other entity. The result is more draws, which dampen your ability to exploit hundreds of elo advantages, and compresses the elo curve.

[Chessqueen]

And it seems you need even less then 16 cores to play high quality chess with NNUE. Current standings with a 10x NPS advantage and a 2 core vs. 32 thread advantage. The Elo difference right now is only 35 ELO!! after 30 games. And this would be about the best case. Since Stockfish is only splitting the search into 2 threads.

Garbage results, which lead to poor understanding of how engines scale. Misinformation at best, deceit at worst.

Typically speaking, a core doubling from 1 to 2 starts at about +70 elo for an engine like Stockfish, using a balanced opening book. As you continue the doubling, that begins to drop. I am not sure exactly of the rate, but it appears logarithmic. However, there was a test with 192 cores vs 384 threads. Note that this is not a typical core doubling, but making use of the hyperthreads on a core. IE, it is not as good as doubling the number of cores. This resulted in about +20 elo after a few thousand games. This test is on fishtest somewhere, and is public record.

So you start your initial core doubling at +70 per doubling, and by the end a doubling is worth at least 20 elo. log2(256) = 8. Hyperthreads are being used. So it can be derived from this knowledge that 1 core SF vs 256 thread SF is roughly 7 * ((70 + 20) / 2) + 20, which is 300+ elo.

Will you see this over the board? Probably not. Elo differences at those extremes are poorly defined by the elo curve, which is well adjusted for similarly skilled opponents, but fails at the extremes due to the nature of the games it is employed in. The issue in this case being that SF is already so strong that, if there were a skill cap, SF is closer to it than any other entity. The result is more draws, which dampen your ability to exploit hundreds of elo advantages, and compresses the elo curve.

Is the Ethereal version that is playing on TCEC S20 different than version 12.75 ?

[AndrewGrant]

Is the Ethereal version that is playing on TCEC S20 different than version 12.75 ?

Yes. If you have questions, PM me. This is not an Ethereal thread

[Pedro]

With such powerful gpus, as well as the cpu, it would be cool if the chess.com team improved Leela's networks to the next level of excellence. Or the Komodo neural network, since it is owned by chess.com.

It is not due to programming fault for not being able to utilize additional resources. It is due to incredibly smart features ( highly selectively search features ) of current top engines that doesnt require much resources to get the right choice.


As average branching factor of selective search of Stockfish is around 2.2 with aggressive pruning, around 16 cores of Stockfish is generally enough to play high quality chess. If you have 256 cores, SF SMP will use it for non priority trees. Therefore, elo gain is really diminished.

On the other hand, 200+ cores to ancient engines without pruning like deep blue will be massive elo gain, compared to Stockfish.

But it is that DeepMind took only 24 hours to make the Alphazero very powerful, and it was used during this training a monster hardware from Google. The question of the layman I have is whether, with this new hardware from chess.com, it would not be possible to further increase the strength of Lc0, if the chess.com team were interested in Leela.

[AndrewGrant]

But it is that DeepMind took only 24 hours to make the Alphazero very powerful, and it was used during this training a monster hardware from Google. The question of the layman I have is whether, with this new hardware from chess.com, it would not be possible to further increase the strength of Lc0, if the chess.com team were interested in Leela.

Both AlphaZero, and Leela (Not Zero) have far more processing power than what you see running CCC's events. Google had hundreds (thousands?) of TPUs. Leela team, back when people were invested heavily, had hundreds of GPU contributors.

[mwyoung]

And it seems you need even less then 16 cores to play high quality chess with NNUE. Current standings with a 10x NPS advantage and a 2 core vs. 32 thread advantage. The Elo difference right now is only 35 ELO!! after 30 games. And this would be about the best case. Since Stockfish is only splitting the search into 2 threads.

Garbage results, which lead to poor understanding of how engines scale. Misinformation at best, deceit at worst.
Unless you are saying that you are giving the 2 cores 10x the thinking time? Which if you are, ignore my entire post.

Typically speaking, a core doubling from 1 to 2 starts at about +70 elo for an engine like Stockfish, using a balanced opening book. As you continue the doubling, that begins to drop. I am not sure exactly of the rate, but it appears logarithmic. However, there was a test with 192 cores vs 384 threads. Note that this is not a typical core doubling, but making use of the hyperthreads on a core. IE, it is not as good as doubling the number of cores. This resulted in about +20 elo after a few thousand games. This test is on fishtest somewhere, and is public record.

So you start your initial core doubling at +70 per doubling, and by the end a doubling is worth at least 20 elo. log2(256) = 8. Hyperthreads are being used. So it can be derived from this knowledge that 1 core SF vs 256 thread SF is roughly 7 * ((70 + 20) / 2) + 20, which is 300+ elo.

Will you see this over the board? Probably not. Elo differences at those extremes are poorly defined by the elo curve, which is well adjusted for similarly skilled opponents, but fails at the extremes due to the nature of the games it is employed in. The issue in this case being that SF is already so strong that, if there were a skill cap, SF is closer to it than any other entity. The result is more draws, which dampen your ability to exploit hundreds of elo advantages, and compresses the elo curve.

The results are correct! Keep up to date. This is current testing with NNUE. Not typically Speaking.

SF 280920 (2 core, No TB) just smashed Fritz 17 4 Core by over 300 Elo

And playing Dragon 32 threads is only -24 Elo after 44 games.

Live Stream:

"Misinformation at best, deceit at worst." Feel free to play along at home with your own system. And then let us know about the deceit you are finding. That is why I stream live.

[mwyoung]

And it seems you need even less then 16 cores to play high quality chess with NNUE. Current standings with a 10x NPS advantage and a 2 core vs. 32 thread advantage. The Elo difference right now is only 35 ELO!! after 30 games. And this would be about the best case. Since Stockfish is only splitting the search into 2 threads.

Garbage results, which lead to poor understanding of how engines scale. Misinformation at best, deceit at worst.
Unless you are saying that you are giving the 2 cores 10x the thinking time? Which if you are, ignore my entire post.

Typically speaking, a core doubling from 1 to 2 starts at about +70 elo for an engine like Stockfish, using a balanced opening book. As you continue the doubling, that begins to drop. I am not sure exactly of the rate, but it appears logarithmic. However, there was a test with 192 cores vs 384 threads. Note that this is not a typical core doubling, but making use of the hyperthreads on a core. IE, it is not as good as doubling the number of cores. This resulted in about +20 elo after a few thousand games. This test is on fishtest somewhere, and is public record.

So you start your initial core doubling at +70 per doubling, and by the end a doubling is worth at least 20 elo. log2(256) = 8. Hyperthreads are being used. So it can be derived from this knowledge that 1 core SF vs 256 thread SF is roughly 7 * ((70 + 20) / 2) + 20, which is 300+ elo.

Will you see this over the board? Probably not. Elo differences at those extremes are poorly defined by the elo curve, which is well adjusted for similarly skilled opponents, but fails at the extremes due to the nature of the games it is employed in. The issue in this case being that SF is already so strong that, if there were a skill cap, SF is closer to it than any other entity. The result is more draws, which dampen your ability to exploit hundreds of elo advantages, and compresses the elo curve.

Your post is B.S. The data is here for all to see. And Just not from myself, but other testers. You just refuse to accept what is obvious. With NNUE you can ignore your scaling formula. Because it does not hold water.

You can trash me, CCRL, and others. Because you do not like the facts. But the facts are not going away with how NNUE scales.

Lets look not at my data, but CCRL blitz data. And lets see how NNUE behaves to a typical A/B engine like Ethereal.

1. Ethereal 12.75 64-bit 8CPU 3537 +20 −20 39.9% +61.4 57.2% 683
2. Ethereal 12.75 64-bit 3377 +22 −22 64.8% −86.2 56.5% 600

CCRL rating difference for 1 to 8 cores. 160 Elo

1. Stockfish 12 64-bit 8CPU 3692 +15 −14 71.5% −147.4 54.5% 1532
2. Stockfish 12 64-bit 3639 +15 −14 79.0% −226.6 36.1% 2054

CCRL rating difference from 1 to 8 cores. 53 Elo

"So you start your initial core doubling at +70 per doubling"