Difference between revisions of "Ethereal"

From Chessprogramming wiki
Jump to: navigation, search
(→‎2020 ...)
 
(14 intermediate revisions by one other user not shown)
Line 5: Line 5:
 
'''Ethereal''',<br/>
 
'''Ethereal''',<br/>
 
an [[UCI]] compliant [[:Category:Open Source|open source chess engine]] written by [[Andrew Grant]] in [[C]], licensed under the [[Free Software Foundation#GPL|GNU GPL]] and first officially released in June 2016 <ref>[http://kirill-kryukov.com/chess/discussion-board/viewtopic.php?f=7&t=8645 Ethereal - Official Release] by [[Andrew Grant]], [[Computer Chess Forums|CCRL Discussion Board]], June 25, 2016</ref>. Ethereal is greatly influenced from [[Crafty]], [[Stockfish]], [[TSCP]], [[MadChess]], and [[Fruit]] <ref>[https://github.com/AndyGrant/Ethereal/blob/master/README.md Ethereal/README.md at master · AndyGrant/Ethereal · GitHub]</ref>.  
 
an [[UCI]] compliant [[:Category:Open Source|open source chess engine]] written by [[Andrew Grant]] in [[C]], licensed under the [[Free Software Foundation#GPL|GNU GPL]] and first officially released in June 2016 <ref>[http://kirill-kryukov.com/chess/discussion-board/viewtopic.php?f=7&t=8645 Ethereal - Official Release] by [[Andrew Grant]], [[Computer Chess Forums|CCRL Discussion Board]], June 25, 2016</ref>. Ethereal is greatly influenced from [[Crafty]], [[Stockfish]], [[TSCP]], [[MadChess]], and [[Fruit]] <ref>[https://github.com/AndyGrant/Ethereal/blob/master/README.md Ethereal/README.md at master · AndyGrant/Ethereal · GitHub]</ref>.  
−
On October 09, 2020, Andrew Grant announced Ethereal's end of development with the final release of '''V12.75''' <ref>[http://www.talkchess.com/forum3/viewtopic.php?f=2&t=75335 Final Release of Ethereal, V12.75] by [[Andrew Grant]], [[CCC]], October 09, 2020</ref>, as well as Ethereal '''12.75 SF-NNUE''' <ref>[http://www.talkchess.com/forum3/viewtopic.php?f=2&t=75335&start=91 Re: Final Release of Ethereal, V12.75] by [[Andrew Grant]], [[CCC]], October 09, 2020</ref>.  
+
On October 09, 2020, Andrew Grant initially announced his withdrawal from Ethereal's development with the releases of Ethereal '''V12.75'''  
 +
<ref>[http://www.talkchess.com/forum3/viewtopic.php?f=2&t=75335 Final Release of Ethereal, V12.75] by [[Andrew Grant]], [[CCC]], October 09, 2020</ref>
 +
and Ethereal '''12.75 SF-NNUE''', the latter a [[NNUE]] implementation based on [[Stockfish NNUE]], to deliberately demonstrate how everybody may considerably improve the [[Playing Strength|playing strength]] of their engines without much effort
 +
<ref>[http://www.talkchess.com/forum3/viewtopic.php?f=2&t=75335&start=91 Re: Final Release of Ethereal, V12.75] by [[Andrew Grant]], [[CCC]], October 09, 2020</ref>.
 +
However, Andrew Grant voted for another direction, and announced a commercial release of Ethereal '''13.00''' (NNUE) ,
 +
the free standard version still available on [https://en.wikipedia.org/wiki/GitHub Github] <ref name="Ethereal13NNUE">[http://www.talkchess.com/forum3/viewtopic.php?f=2&t=77438 Commercial Release of Ethereal 13.00 (NNUE) for AVX2 Systems] by [[Andrew Grant]], [[CCC]], June 04, 2021</ref>.
 +
 
 +
=Ethereal 13 (NNUE)=
 +
'''Ethereal 13 (NNUE)''', the commercial Ethereal version for [[AVX2]] systems was released on June 04, 2021. The program comes with two [[NNUE|NNUEs]] for  [[Evaluation|evaluation]],
 +
one for standard chess, and the secondary network trained exclusively for [[Chess960]]. The NNUEs are not derived from, trained on, nor duplicated from the works of the [[Stockfish]] team  <ref name="Ethereal13NNUE"/>. Andrew Grant on his NNUE approach <ref>[http://www.talkchess.com/forum3/viewtopic.php?f=2&t=77438&start=17 Re: Commercial Release of Ethereal 13.00 (NNUE) for AVX2 Systems] by [[Andrew Grant]], [[CCC]], June 04, 2021</ref>:
 +
 
 +
Ethereal is using the [[Stockfish NNUE#NNUE Structure|HalfKP]] paradigm, with a 40960x256 -> 512x32x32x1 Network. This is the textbook approach, but with some changes. Firstly, not all weights are quantized to int8 / int16 for the input layer. Instead, the network goes like this: int16_t => int16_t => (int32_t -> [[Float|float_t]]) => float_t => float_t. This approach allows us to never have to pack the data downwards, saving many operations, and also lets us take a slightly more expensive approach to the later layers in exchange for massively increased precision. If I eventually add support for [[AVX]] (not avx2) machines, it will be a significant gain as AVX does not have 256-bit vector support for integer types in a meaningful way.
 +
 
 +
During training the Network actually has 43850 input parameters, using a few factorization of the board to aid in training without having tens of billions of positions. In practice, each Net was trained somewhere between 2 and 4 billion positions total, evaluated by Ethereal / Ethereal NNUE. The networks are trained using a modified form of the [https://en.wikipedia.org/wiki/Stochastic_gradient_descent#Adam Adam] optimizer, which allows better performance for datasets with extremely sparse input fields. For example, with a Batch Size of 16384, only about 50% of the 43,850 parameters are used on average.
 +
 
 +
Data generation for a given network takes about 3 weeks, completed on a 104 core machine. From there, processing that data down into a list of [[Forsyth-Edwards Notation|FENs]] and then into the format used by Ethereal's NNTrainer takes another 12 hours or so. Finally, training the actual Network can take a few days, with many stops and starts to drop the learning rate and find a global optima.
 +
 
 +
The trainer itself is a fully original work, written in [[C]] and making use of all 104 threads. It includes some AVX2 and even AVX512 code for use in updating the network parameters. This toolkit was used in training the Halogen networks as well. It is fairly flexible and trying things like HalfKA, changing layer sizes, adding layers, changing activation functions, or adding more factorizers is only a few minutes of effort in the code. It rivals speeds of [[GPU]] based trainers, by leveraging massive SMP and efficient implementations.
  
 
=Features=
 
=Features=
Line 36: Line 53:
 
** [[Piece-Square Tables]]  
 
** [[Piece-Square Tables]]  
 
==[[Evaluation]]==
 
==[[Evaluation]]==
−
* [[NNUE]] (Ethereal 12.75 SF-NNUE)
+
* [[NNUE]] (Ethereal 13.00 NNUE)
 
* [[Tapered Eval]]
 
* [[Tapered Eval]]
 
* [[Material]]
 
* [[Material]]
Line 89: Line 106:
 
* [http://www.talkchess.com/forum3/viewtopic.php?f=2&t=75335 Final Release of Ethereal, V12.75] by [[Andrew Grant]], [[CCC]], October 09, 2020
 
* [http://www.talkchess.com/forum3/viewtopic.php?f=2&t=75335 Final Release of Ethereal, V12.75] by [[Andrew Grant]], [[CCC]], October 09, 2020
 
: [http://www.talkchess.com/forum3/viewtopic.php?f=2&t=75335&start=91 Re: Final Release of Ethereal, V12.75] by [[Andrew Grant]], [[CCC]], October 09, 2020
 
: [http://www.talkchess.com/forum3/viewtopic.php?f=2&t=75335&start=91 Re: Final Release of Ethereal, V12.75] by [[Andrew Grant]], [[CCC]], October 09, 2020
 +
: [http://www.talkchess.com/forum3/viewtopic.php?f=2&t=75335&start=134 Re: Final Release of Ethereal, V12.75] by [[Andrew Grant]], [[CCC]], November 12, 2020
 
* [http://www.talkchess.com/forum3/viewtopic.php?f=7&t=75345 Request for someone to train an NNUE for Ethereal] by [[Andrew Grant]], [[CCC]], October 09, 2020 » [[NNUE]]
 
* [http://www.talkchess.com/forum3/viewtopic.php?f=7&t=75345 Request for someone to train an NNUE for Ethereal] by [[Andrew Grant]], [[CCC]], October 09, 2020 » [[NNUE]]
 
* [http://www.talkchess.com/forum3/viewtopic.php?f=7&t=75350 Ethereal Tuning - Data Dump] by [[Andrew Grant]], [[CCC]], October 10, 2020
 
* [http://www.talkchess.com/forum3/viewtopic.php?f=7&t=75350 Ethereal Tuning - Data Dump] by [[Andrew Grant]], [[CCC]], October 10, 2020
 +
* [http://www.talkchess.com/forum3/viewtopic.php?f=2&t=76049 Ethereal questions] by [[Gabor Szots]], [[CCC]], December 12, 2020
 +
'''2021'''
 +
* [http://www.talkchess.com/forum3/viewtopic.php?f=2&t=77438 Commercial Release of Ethereal 13.00 (NNUE) for AVX2 Systems] by [[Andrew Grant]], [[CCC]], June 04, 2021
 +
: [http://www.talkchess.com/forum3/viewtopic.php?f=2&t=77438&start=17 Re: Commercial Release of Ethereal 13.00 (NNUE) for AVX2 Systems] by [[Andrew Grant]], [[CCC]], June 04, 2021
 +
* [http://www.talkchess.com/forum3/viewtopic.php?f=2&t=77503&start=55 Re: will Tcec allow Stockfish with a Leela net to play?] by [[Connor McMonigle]], [[CCC]], June 17, 2021 » [[NNUE]]
 +
* [http://www.talkchess.com/forum3/viewtopic.php?f=2&t=77571 I declare that HCE is dead...] by [[Andrew Grant]], [[CCC]], June 29, 2021 » [[Evaluation|HCE]], [[NNUE]]
 +
'''2022'''
 +
* [https://www.talkchess.com/forum3/viewtopic.php?f=7&t=79160 Resolving once in a trillion crashes] by [[Andrew Grant]], [[CCC]], January 18, 2022 » [[Debugging]]
 +
* [https://www.talkchess.com/forum3/viewtopic.php?f=2&t=79179 Commercial Release of Ethereal 13.50] by [[Andrew Grant]], [[CCC]], January 20, 2022
 +
'''2023'''
 +
* [https://talkchess.com/forum3/viewtopic.php?f=2&t=82750 Commercial Release of Ethereal 14.25] by [[Andrew Grant]], [[CCC]], October 21, 2023
  
 
=External Links=
 
=External Links=
 
==Chess Engine==
 
==Chess Engine==
 +
* [http://chess.grantnet.us/Ethereal/ Ethereal 13.00 (NNUE)]
 
* [https://github.com/AndyGrant/Ethereal GitHub - AndyGrant/Ethereal: Ethereal UCI Chess Engine]
 
* [https://github.com/AndyGrant/Ethereal GitHub - AndyGrant/Ethereal: Ethereal UCI Chess Engine]
−
* [http://www.computerchess.org.uk/ccrl/404/cgi/compare_engines.cgi?family=Ethereal&print=Rating+list&print=Results+table&print=LOS+table&print=Ponder+hit+table&print=Eval+difference+table&print=Comopp+gamenum+table&print=Overlap+table&print=Score+with+common+opponents Ethereal] in [[CCRL|CCRL 40/4]]
+
* [http://www.computerchess.org.uk/ccrl/404/cgi/compare_engines.cgi?family=Ethereal&print=Rating+list&print=Results+table&print=LOS+table&print=Ponder+hit+table&print=Eval+difference+table&print=Comopp+gamenum+table&print=Overlap+table&print=Score+with+common+opponents Ethereal] in [[CCRL|CCRL Blitz]]
 +
* [http://www.computerchess.org.uk/ccrl/4040/cgi/compare_engines.cgi?family=Ethereal&print=Rating+list&print=Results+table&print=LOS+table&print=Ponder+hit+table&print=Eval+difference+table&print=Comopp+gamenum+table&print=Overlap+table&print=Score+with+common+opponents Ethereal] in [[CCRL|CCRL 40/15]]
 
* [http://www.chessdom.com/ethereal-wins-tcec-division-4/ Ethereal wins TCEC Division 4], Interview with [[Andrew Grant]], [[Chessdom]], April 30, 2018 » [[TCEC Season 12]]
 
* [http://www.chessdom.com/ethereal-wins-tcec-division-4/ Ethereal wins TCEC Division 4], Interview with [[Andrew Grant]], [[Chessdom]], April 30, 2018 » [[TCEC Season 12]]
 
* [http://www.chessdom.com/ethereal-chess-engine-leads-tcec-div-3-convincingly/ Ethereal chess engine leads TCEC Div 3 convincingly], [[Chessdom]], August 12, 2018 » [[TCEC Season 13]]
 
* [http://www.chessdom.com/ethereal-chess-engine-leads-tcec-div-3-convincingly/ Ethereal chess engine leads TCEC Div 3 convincingly], [[Chessdom]], August 12, 2018 » [[TCEC Season 13]]
 
* [http://www.chessdom.com/ethereal-chess-engine-wins-the-gold-at-tcec-div-3/ Ethereal chess engine wins the gold at TCEC Div 3], [[Chessdom]], August 17, 2018
 
* [http://www.chessdom.com/ethereal-chess-engine-wins-the-gold-at-tcec-div-3/ Ethereal chess engine wins the gold at TCEC Div 3], [[Chessdom]], August 17, 2018
−
 
 
==Misc==
 
==Misc==
 
* [https://en.wiktionary.org/wiki/ethereal ethereal - Wiktionary]
 
* [https://en.wiktionary.org/wiki/ethereal ethereal - Wiktionary]
Line 108: Line 138:
 
* [https://www.truth-is-beauty.com/blog/style-identities-the-ethereal The Ethereal] from [https://www.truth-is-beauty.com/home.html Truth is Beauty], July 07, 2015
 
* [https://www.truth-is-beauty.com/blog/style-identities-the-ethereal The Ethereal] from [https://www.truth-is-beauty.com/home.html Truth is Beauty], July 07, 2015
 
* [[:Category:Jean-Luc Ponty|Jean-Luc Ponty]] - [https://en.wikipedia.org/wiki/Cosmic_Messenger Ethereal Mood] (1978), [https://en.wikipedia.org/wiki/YouTube YouTube] Video
 
* [[:Category:Jean-Luc Ponty|Jean-Luc Ponty]] - [https://en.wikipedia.org/wiki/Cosmic_Messenger Ethereal Mood] (1978), [https://en.wikipedia.org/wiki/YouTube YouTube] Video
−
: {{#evu:https://www.youtube.com/watch?v=JwibRL7Eqq4|alignment=left|valignment=top}}
+
: {{#evu:https://www.youtube.com/watch?v=Va5_Y4QS3Eo|alignment=left|valignment=top}}
  
 
=References=  
 
=References=  
Line 121: Line 151:
 
[[Category:Linux]]
 
[[Category:Linux]]
 
[[Category:Windows]]
 
[[Category:Windows]]
 +
[[Category:Chess960]]
 +
[[Category:OpenBench]]
 +
[[Category:Commercial]]
 
[[Category:NNUE]]
 
[[Category:NNUE]]
 
[[Category:Jean-Luc Ponty]]
 
[[Category:Jean-Luc Ponty]]
 
[[Category:Astronomy]]
 
[[Category:Astronomy]]

Latest revision as of 09:37, 21 October 2023

Home * Engines * Ethereal

Ethereal night sky [1]

Ethereal,
an UCI compliant open source chess engine written by Andrew Grant in C, licensed under the GNU GPL and first officially released in June 2016 [2]. Ethereal is greatly influenced from Crafty, Stockfish, TSCP, MadChess, and Fruit [3]. On October 09, 2020, Andrew Grant initially announced his withdrawal from Ethereal's development with the releases of Ethereal V12.75 [4] and Ethereal 12.75 SF-NNUE, the latter a NNUE implementation based on Stockfish NNUE, to deliberately demonstrate how everybody may considerably improve the playing strength of their engines without much effort [5]. However, Andrew Grant voted for another direction, and announced a commercial release of Ethereal 13.00 (NNUE) , the free standard version still available on Github [6].

Ethereal 13 (NNUE)

Ethereal 13 (NNUE), the commercial Ethereal version for AVX2 systems was released on June 04, 2021. The program comes with two NNUEs for evaluation, one for standard chess, and the secondary network trained exclusively for Chess960. The NNUEs are not derived from, trained on, nor duplicated from the works of the Stockfish team [6]. Andrew Grant on his NNUE approach [7]: 

Ethereal is using the HalfKP paradigm, with a 40960x256 -> 512x32x32x1 Network. This is the textbook approach, but with some changes. Firstly, not all weights are quantized to int8 / int16 for the input layer. Instead, the network goes like this: int16_t => int16_t => (int32_t -> float_t) => float_t => float_t. This approach allows us to never have to pack the data downwards, saving many operations, and also lets us take a slightly more expensive approach to the later layers in exchange for massively increased precision. If I eventually add support for AVX (not avx2) machines, it will be a significant gain as AVX does not have 256-bit vector support for integer types in a meaningful way.

During training the Network actually has 43850 input parameters, using a few factorization of the board to aid in training without having tens of billions of positions. In practice, each Net was trained somewhere between 2 and 4 billion positions total, evaluated by Ethereal / Ethereal NNUE. The networks are trained using a modified form of the Adam optimizer, which allows better performance for datasets with extremely sparse input fields. For example, with a Batch Size of 16384, only about 50% of the 43,850 parameters are used on average.

Data generation for a given network takes about 3 weeks, completed on a 104 core machine. From there, processing that data down into a list of FENs and then into the format used by Ethereal's NNTrainer takes another 12 hours or so. Finally, training the actual Network can take a few days, with many stops and starts to drop the learning rate and find a global optima.

The trainer itself is a fully original work, written in C and making use of all 104 threads. It includes some AVX2 and even AVX512 code for use in updating the network parameters. This toolkit was used in training the Halogen networks as well. It is fairly flexible and trying things like HalfKA, changing layer sizes, adding layers, changing activation functions, or adding more factorizers is only a few minutes of effort in the code. It rivals speeds of GPU based trainers, by leveraging massive SMP and efficient implementations.

Features

[8]

Board Representation

Search

Evaluation

Misc

Publications

Forum Posts

2016 ...

Re: How to speed up my engine by Andrew Grant, CCC, May 03, 2016

2017

2018

2019

2020 ...

Re: Final Release of Ethereal, V12.75 by Andrew Grant, CCC, October 09, 2020
Re: Final Release of Ethereal, V12.75 by Andrew Grant, CCC, November 12, 2020

2021

Re: Commercial Release of Ethereal 13.00 (NNUE) for AVX2 Systems by Andrew Grant, CCC, June 04, 2021

2022

2023

External Links

Chess Engine

Misc

References

  1. ↑ A total eclipse of the Moon is an impressive spectacle. But it also provides another viewing opportunity: a dark, moonlight-free starry sky. At Cerro Paranal in the Chilean Atacama Desert, one of the most remote places in the world, the distance from sources of light pollution makes the night sky all the more remarkable during a total lunar eclipse. This panoramic photo, taken by ESO Photo Ambassador Yuri Beletsky, shows the view of the starry sky from the site of ESO’s Very Large Telescope (VLT) at Cerro Paranal during the total lunar eclipse of December 21, 2010. The reddish disc of the Moon is seen on the right of the image, while the Milky Way arches across the heavens in all its beauty. Another faint glow of light is also visible, surrounding the brilliant planet Venus in the bottom left corner of the picture. This phenomenon, known as zodiacal light, is produced by sunlight reflecting off dust in the plane of the planets. It is so faint that it’s normally obscured by moonlight or light pollution. Image by ESO, Yuri Beletsky, CC BY 4.0, Wikimedia Commons, Beauty from Wikipedia
  2. ↑ Ethereal - Official Release by Andrew Grant, CCRL Discussion Board, June 25, 2016
  3. ↑ Ethereal/README.md at master · AndyGrant/Ethereal · GitHub
  4. ↑ Final Release of Ethereal, V12.75 by Andrew Grant, CCC, October 09, 2020
  5. ↑ Re: Final Release of Ethereal, V12.75 by Andrew Grant, CCC, October 09, 2020
  6. ↑ 6.0 6.1 Commercial Release of Ethereal 13.00 (NNUE) for AVX2 Systems by Andrew Grant, CCC, June 04, 2021
  7. ↑ Re: Commercial Release of Ethereal 13.00 (NNUE) for AVX2 Systems by Andrew Grant, CCC, June 04, 2021
  8. ↑ Features based on GitHub - AndyGrant/Ethereal: Ethereal UCI Chess Engine
  9. ↑ Ethereal's 8.60 Lazy SMP inspired by Demolito by Lucas Braesch, see Release of Ethereal 8.60 SMP by Andrew Grant, CCC, December 10, 2017
  10. ↑ Evaluation & Tuning in Chess Engines by Andrew Grant, CCC, August 24, 2020
  11. ↑ Ethereal Tuning - Data Dump by Andrew Grant, CCC, October 10, 2020
  12. ↑ Evaluation & Tuning in Chess Engines by Andrew Grant, CCC, August 24, 2020
  13. ↑ Ethereal Tuning - Data Dump by Andrew Grant, CCC, October 10, 2020

Up one Level

Retrieved from "https://www.chessprogramming.org/index.php?title=Ethereal&oldid=26830"