Life In 19x19 :: Engine Tournament

Please add DarkGo! It's a neural network at around 1d level.
https://pjreddie.com/darknet/darkgo-go-in-darknet/
How to run in GTP: https://groups.google.com/forum/#!topic ... BNZTkMmUHI

Also you can try Hiratuka 10.38B since you're using GPU anyway.

Pachi can use neural networks as well, if you're not using them already.
https://github.com/pasky/pachi

Thanks for the hints, I had Pachi with neural networks on my list but I wasn't aware of DarkGo.

Unfortunately Hiratuka 10.38B doesn't run on my Linux machine because of a cuda error...

Also are you making sure that the programs are using as close to two hours as possible? For example you can add the argument -t _7200 to pachi to tell it to use two hours. I've noticed that in gogui-twogtp you're giving the engines two hours (i.e. -time 120) but you're only giving ray one hour (i.e. --time 3600). FYI rn-4.6.1 has just been released.

How do you guys gets the Orego 7.08 running?

I downloaded Jar file, unziped, or downloaded the 7.14 version release and javac compiled with no erros. However,
I can not get the software running in linux Ubuntu 14.04. Erros:
java -ea -Xmx1024M edu.lclark.orego.ui.Orego
Error: Could not find or load main class edu.lclark.orego.ui.Orego

No matter how I switched to src folder, or change to localhost.Orego or just Orego, it does not run

Shoreline wrote:

I had problems running later versions because of similar problems. v7.08 works by adding:

Main-Class: orego.ui.Orego

to /META-INF/MANIFEST.MF inside the .jar file.

Then execute the .jar file by:

java -jar /path to/orego-7.08.jar threads=1 grace

If you get later versions to work let me know

Best,
Alex

Thanks, I got it running like you said. However, I found it is very weak bot at 19x19. How come this Orego 7.08 ranks so high against gnugo?
It appears that it is several stones weaker than gnugo.

I wonder if the java version or OS made the difference. I am using a linux machine,
dual core, 8G of memory, quite low end machine. The java verison I am using is:
$ java -version
java version "1.7.0_45"
Java(TM) SE Runtime Environment (build 1.7.0_45-b18)
Java HotSpot(TM) 64-Bit Server VM (build 24.45-b08, mixed mode)

I would say in my setting, the Orego version 7.08 is probably more than 9 stones weaker than gnugo, it probably ranks below 20 kyu in DGS.

Shoreline wrote:

This is my Java version:

java -version
openjdk version "1.8.0_131"
OpenJDK Runtime Environment (build 1.8.0_131-8u131-b11-2ubuntu1.16.04.3-b11)
OpenJDK 64-Bit Server VM (build 25.131-b11, mixed mode)

Does Orego use all the available time? Maybe you use a time setting that Orego doesn't support and it moves faster.

as0770 wrote:

Shoreline wrote:

Ｔhe Orego in my machine moves bleedingly fast, may be 1 second per move. It can play 5 or 10 different games within 10 seconds.

How can I change the time setting in GTP command? Maybe that is the issue.

I use game in x for my testing because this works with most engines. The gtp command is:

time_settings 3600 0 0

for a game in 60 minutes. But Orego uses some more time as allowed.

Please add DarkGo! It's a neural network at around 1d level.
https://pjreddie.com/darknet/darkgo-go-in-darknet/
How to run in GTP: https://groups.google.com/forum/#!topic ... BNZTkMmUHI

Author:	Cyan [ Wed Jul 19, 2017 4:17 pm ]
Post subject:	Re: Engine Tournament
Please add DarkGo! It's a neural network at around 1d level. https://pjreddie.com/darknet/darkgo-go-in-darknet/ How to run in GTP: https://groups.google.com/forum/#!topic ... BNZTkMmUHI Also you can try Hiratuka 10.38B since you're using GPU anyway. Pachi can use neural networks as well, if you're not using them already. https://github.com/pasky/pachi

Author:	as0770 [ Thu Jul 20, 2017 8:00 am ]
Post subject:	Re: Engine Tournament
Cyan wrote: Please add DarkGo! It's a neural network at around 1d level. https://pjreddie.com/darknet/darkgo-go-in-darknet/ How to run in GTP: https://groups.google.com/forum/#!topic ... BNZTkMmUHI Also you can try Hiratuka 10.38B since you're using GPU anyway. Pachi can use neural networks as well, if you're not using them already. https://github.com/pasky/pachi Thanks for the hints, I had Pachi with neural networks on my list but I wasn't aware of DarkGo. Unfortunately Hiratuka 10.38B doesn't run on my Linux machine because of a cuda error...

Author:	Cyan [ Thu Jul 20, 2017 10:44 am ]
Post subject:	Re: Engine Tournament
as0770 wrote: Thanks for the hints, I had Pachi with neural networks on my list but I wasn't aware of DarkGo. Unfortunately Hiratuka 10.38B doesn't run on my Linux machine because of a cuda error... You're welcome! Also are you making sure that the programs are using as close to two hours as possible? For example you can add the argument -t _7200 to pachi to tell it to use two hours. I've noticed that in gogui-twogtp you're giving the engines two hours (i.e. -time 120) but you're only giving ray one hour (i.e. --time 3600). FYI rn-4.6.1 has just been released.

Author:	as0770 [ Thu Jul 20, 2017 11:36 am ]
Post subject:	Re: Engine Tournament
Cyan wrote: Also are you making sure that the programs are using as close to two hours as possible? For example you can add the argument -t _7200 to pachi to tell it to use two hours. I've noticed that in gogui-twogtp you're giving the engines two hours (i.e. -time 120) but you're only giving ray one hour (i.e. --time 3600). FYI rn-4.6.1 has just been released. I use Ray without the --time 3600 argument, it was a typo in the configuration section and I just corrected it. Ray and Pachi seem to work well with the gtp time command, I observe all games also in matter of time consumption. The engines with fixed playouts use in average 48 seconds per move after 50-100 moves. For the games against DarkForrest I measure how much each engine is slowed down by the CPU usage of DarkForrest and give it more time. So I think it is as fair as possible

Author:	as0770 [ Mon Jul 24, 2017 3:15 am ]
Post subject:	Re: Engine Tournament
Update League A with DarkGo, it won 1 against Pachi and Hiratuka, and 3 against DarkForrest, next Update will be Pachi DCNN. League A:with Code: 1. Leela 0.10.0 22/24 2. Rayon 4.6.0 19/24 3. Oakfoam 0.2.1 NG-06 18/24 4. Hiratuka 10.37B (CPU) 10/24 5. DarkForest v2 MCTS 1.0 7/24 6 DarkGo 1.0 5/24 7. Pachi 11.99 3/24 Configuration: 2h/game, pondering off, 4 threads, 2GB on 4 x Intel® Core™ i5-4210H CPU @ 2.90GHz, 8 GiB Ram and GeForce 840M/PCIe/SSE2 DarkForest: taskset -c 0 bash cnn_evaluator.sh 1 /data/local/go and taskset -c 0 th cnnPlayerMCTSV2.lua --num_gpu 1 --num_tree_thread 1 --rollout 2500 --win_rate_thres 0.1 DarkGo: darknet go engine cfg/go.test.cfg go.weights Hiratuka: Hiratuka-19×19.exe -po 175000 Leela: leela_gtp_opencl –gtp –threads 4 –noponder Oakfoam: oakfoam -c nicego-cnn-06.gtp Code: nicego-cnn-06.gtp: param playouts_per_move_max 40000 param thread_count 4 Pachi: pachi -f pachibook.dat threads=4,max_tree_size=2048,pondering=0 Rayon: rayon --thread 4 --no-debug twogtp: TWOGTP=”gogui-twogtp -black \”$BLACK\” -white \”$WHITE\” -games 2 -size 19 -time 120 -sgffile xxxx” gogui -size 19 -program “$TWOGTP” -computer-both -auto League B: Code: 1. Pachi 11.99 23/24 2. MoGo 4.86 19/24 3. deltaGo 1.0.0 14/24 4. Fuego 1.1 13/24 5. Michi C-2 1.4.2 8/24 6. Orego 7.08 5/24 7. GNU Go 3.8 2/24 Configuration: 1h/game, pondering off, 1 thread, 1GB on 4 x Intel® Core™ i5-4210H CPU @ 2.90GHz, 8 GiB Ram and GeForce 840M/PCIe/SSE2 deltaGo: deltaGo.exe Fuego: fuego.exe –config fuego.cfg Code: fuego.cfg: uct_param_search number_threads 1 uct_param_search lock_free 0 uct_max_memory 1024000000 uct_param_player reuse_subtree 1 uct_param_player ponder 0 uct_param_player early_pass 1 GnuGo: gnugo --mode gtp --level 10 --resign-allowed Michi C-2: michi gtp Code: ui.c: init_large_patterns("patterns2.prob", "patterns2.spat"); // Michis pattern files renamed because they have the same name as Pachis files. MoGo: mogo Orego: java -jar /path/to/orego-7.08.jar threads=1 grace Pachi: pachi -f pachibook.dat threads=1,max_tree_size=1024,pondering=0 twogtp: TWOGTP=”gogui-twogtp -black \”$BLACK\” -white \”$WHITE\” -games 2 -size 19 -time 60 -sgffile xxxx” gogui -size 19 -program “$TWOGTP” -computer-both -auto League C: Code: 1. GNU Go 3.8 24/28 2. Hara 0.9 18/28 3. Dariush 3.1.5.7 16/28 4. Indigo 2009 15/28 5. Matilda 1.24 15/28 6. Aya 6.34 11/28 7. Fudo Go 3.0 11/28 8. JrefBot 081016-2022 2/28 Configuration: 1h/game, pondering off, 1 thread, 1GB on 4 x Intel® Core™ i5-4210H CPU @ 2.90GHz, 8 GiB Ram and GeForce 840M/PCIe/SSE2 Aya: Aya.exe –mode gtp –level max Dariush: DarGTP.exe –level 10 Fudo Go: taskset -c 0 fudo –boardsize=19 –komi=6.5 GnuGo: gnugo --mode gtp --level 10 --resign-allowed Hara: hara IndiGo: Indigo.exe -gtp JrefBot: java -jar jrefgo.jar 10000 Matilda: matilda Code: matilda.h: #define BOARD_SIZ 19 #define DEFAULT_UCT_MEMORY 1000 #define DEFAULT_NUM_THREADS 1 twogtp: TWOGTP=”gogui-twogtp -black \”$BLACK\” -white \”$WHITE\” -games 2 -size 19 -time 60 -sgffile xxxx” gogui -size 19 -program “$TWOGTP” -computer-both -auto League D: Code: 1. JrefBot 081016-2022 25/28 2. Iomrascálaí 0.3.2 21/28 3. Crazy Patterns 0008-13 19/28 4. Marcos Go 1.0 18/28 5. AmiGo 1.8 16/28 6. Stop 0.9-005 6/28 7. CopyBot 0.1 4/28 8. Brown 1.0 3/28 Configuration: 1h/game, pondering off, 1 thread, 1GB on 4 x Intel® Core™ i5-4210H CPU @ 2.90GHz, 8 GiB Ram and GeForce 840M/PCIe/SSE2 Amigo: amigogtp Brown: brown.exe Copybot: python /path/to/__main__.py CrazyPatterns: CrazyPatterns.exe Iomrascálaí: taskset -c 0,1 iomrascalai JrefBot: java -jar jrefgo.jar 10000 Marcos Go: marcos_go --patterns /path/to/patterns.txt --cycles_mcts 10000 --threads_mcts 1 Stop: /usr/bin/java -ea -jar /path/to/stop-09-005.jar --mode gtp twogtp: TWOGTP=”gogui-twogtp -black \”$BLACK\” -white \”$WHITE\” -games 2 -size 19 -time 60 -sgffile xxxx” gogui -size 19 -program “$TWOGTP” -computer-both -auto Links: Amigo: https://sourceforge.net/projects/amigogtp/ Aya: http://www.yss-aya.com/ Brown: http://ricoh51.free.fr/go/engineeng.htm Copybot: https://github.com/sirtango/ICopyMoves CrazyPatterns: https://www.remi-coulom.fr/Amsterdam2007/ Dariush: http://ricoh51.free.fr/go/engineeng.htm DarkForest: https://github.com/facebookresearch/darkforestGo DarkGo: https://pjreddie.com/darknet/darkgo-go-in-darknet/ deltaGo: http://home.q00.itscom.net/otsuki/delta.html Fudo Go: http://www.geocities.jp/hideki_katoh/ Fuego: http://fuego.sourceforge.net/ GnuGo: https://www.gnu.org/software/gnugo/devel.html Hara: https://github.com/antoniogarro/Hara Hiratuka: Non GPU version (10.37B): http://www.vector.co.jp/download/file/winnt/game/fh673259.html / GPU version (10.38B): http://www.vector.co.jp/download/file/winnt/game/fh688349.html IndiGo: http://www.math-info.univ-paris5.fr/~bouzy/INDIGO.html Iomrascálaí: https://github.com/ujh/iomrascalai JrefBot: http://ricoh51.free.fr/go/engineeng.htm Leela: https://sjeng.org/leela.html Marcos Go: https://github.com/MarcosPividori/Go-player Matilda: https://github.com/gonmf/matilda Michi C-2 https://github.com/db3108/michi-c2 MoGo: https://lifein19x19.com/forum/viewtopic.php?p=211091#p211091 Oakfoam: https://bitbucket.org/dsmic/oakfoam Orego: https://sites.google.com/a/lclark.edu/drake/research/orego Pachi: http://pachi.or.cz/ Rayon: https://github.com/zakki/Ray Stop: https://www.vanheusden.com/stop/ Best, Alex

Life In 19x19 http://lifein19x19.com/

Engine Tournament http://lifein19x19.com/viewtopic.php?f=18&t=13322	Page 5 of 20

Author:	as0770 [ Sun Jul 30, 2017 10:04 pm ]
Post subject:	Re: Engine Tournament
Update League A with Pachi DCNN. League A: Code: 1. Leela 0.10.0 22/24 2. Rayon 4.6.0 19/24 3. Oakfoam 0.2.1 NG-06 18/24 4. Hiratuka 10.37B (CPU) 9/24 5. DarkForest v2 MCTS 1.0 7/24 6 DarkGo 1.0 5/24 7. Pachi DCNN 11.99 4/24 Configuration: 2h/game, pondering off, 4 threads, 2GB on 4 x Intel® Core™ i5-4210H CPU @ 2.90GHz, 8 GiB Ram and GeForce 840M/PCIe/SSE2 DarkForest: taskset -c 0 bash cnn_evaluator.sh 1 /data/local/go and taskset -c 0 th cnnPlayerMCTSV2.lua --num_gpu 1 --num_tree_thread 1 --rollout 2500 --win_rate_thres 0.1 DarkGo: darknet go engine cfg/go.test.cfg go.weights Hiratuka: Hiratuka-19×19.exe -po 175000 Leela: leela_gtp_opencl –gtp –threads 4 –noponder Oakfoam: oakfoam -c nicego-cnn-06.gtp Code: nicego-cnn-06.gtp: param playouts_per_move_max 40000 param thread_count 4 Pachi: pachidcnn -f pachibook.dat threads=4,max_tree_size=2048,pondering=0 Rayon: rayon --thread 4 --no-debug twogtp: TWOGTP=”gogui-twogtp -black \”$BLACK\” -white \”$WHITE\” -games 2 -size 19 -time 120 -sgffile xxxx” gogui -size 19 -program “$TWOGTP” -computer-both -auto League B: Code: 1. Pachi 11.99 23/24 2. MoGo 4.86 19/24 3. deltaGo 1.0.0 14/24 4. Fuego 1.1 13/24 5. Michi C-2 1.4.2 8/24 6. Orego 7.08 5/24 7. GNU Go 3.8 2/24 Configuration: 1h/game, pondering off, 1 thread, 1GB on 4 x Intel® Core™ i5-4210H CPU @ 2.90GHz, 8 GiB Ram and GeForce 840M/PCIe/SSE2 deltaGo: deltaGo.exe Fuego: fuego.exe –config fuego.cfg Code: fuego.cfg: uct_param_search number_threads 1 uct_param_search lock_free 0 uct_max_memory 1024000000 uct_param_player reuse_subtree 1 uct_param_player ponder 0 uct_param_player early_pass 1 GnuGo: gnugo --mode gtp --level 10 --resign-allowed Michi C-2: michi gtp Code: ui.c: init_large_patterns("patterns2.prob", "patterns2.spat"); // Michis pattern files renamed because they have the same name as Pachis files. MoGo: mogo Orego: java -jar /path/to/orego-7.08.jar threads=1 grace Pachi: pachi -f pachibook.dat threads=1,max_tree_size=1024,pondering=0 twogtp: TWOGTP=”gogui-twogtp -black \”$BLACK\” -white \”$WHITE\” -games 2 -size 19 -time 60 -sgffile xxxx” gogui -size 19 -program “$TWOGTP” -computer-both -auto League C: Code: 1. GNU Go 3.8 24/28 2. Hara 0.9 18/28 3. Dariush 3.1.5.7 16/28 4. Indigo 2009 15/28 5. Matilda 1.24 15/28 6. Aya 6.34 11/28 7. Fudo Go 3.0 11/28 8. JrefBot 081016-2022 2/28 Configuration: 1h/game, pondering off, 1 thread, 1GB on 4 x Intel® Core™ i5-4210H CPU @ 2.90GHz, 8 GiB Ram and GeForce 840M/PCIe/SSE2 Aya: Aya.exe –mode gtp –level max Dariush: DarGTP.exe –level 10 Fudo Go: taskset -c 0 fudo –boardsize=19 –komi=6.5 GnuGo: gnugo --mode gtp --level 10 --resign-allowed Hara: hara IndiGo: Indigo.exe -gtp JrefBot: java -jar jrefgo.jar 10000 Matilda: matilda Code: matilda.h: #define BOARD_SIZ 19 #define DEFAULT_UCT_MEMORY 1000 #define DEFAULT_NUM_THREADS 1 twogtp: TWOGTP=”gogui-twogtp -black \”$BLACK\” -white \”$WHITE\” -games 2 -size 19 -time 60 -sgffile xxxx” gogui -size 19 -program “$TWOGTP” -computer-both -auto League D: Code: 1. JrefBot 081016-2022 25/28 2. Iomrascálaí 0.3.2 21/28 3. Crazy Patterns 0008-13 19/28 4. Marcos Go 1.0 18/28 5. AmiGo 1.8 16/28 6. Stop 0.9-005 6/28 7. CopyBot 0.1 4/28 8. Brown 1.0 3/28 Configuration: 1h/game, pondering off, 1 thread, 1GB on 4 x Intel® Core™ i5-4210H CPU @ 2.90GHz, 8 GiB Ram and GeForce 840M/PCIe/SSE2 Amigo: amigogtp Brown: brown.exe Copybot: python /path/to/__main__.py CrazyPatterns: CrazyPatterns.exe Iomrascálaí: taskset -c 0,1 iomrascalai JrefBot: java -jar jrefgo.jar 10000 Marcos Go: marcos_go --patterns /path/to/patterns.txt --cycles_mcts 10000 --threads_mcts 1 Stop: /usr/bin/java -ea -jar /path/to/stop-09-005.jar --mode gtp twogtp: TWOGTP=”gogui-twogtp -black \”$BLACK\” -white \”$WHITE\” -games 2 -size 19 -time 60 -sgffile xxxx” gogui -size 19 -program “$TWOGTP” -computer-both -auto Links: Amigo: https://sourceforge.net/projects/amigogtp/ Aya: http://www.yss-aya.com/ Brown: http://ricoh51.free.fr/go/engineeng.htm Copybot: https://github.com/sirtango/ICopyMoves CrazyPatterns: https://www.remi-coulom.fr/Amsterdam2007/ Dariush: http://ricoh51.free.fr/go/engineeng.htm DarkForest: https://github.com/facebookresearch/darkforestGo DarkGo: https://pjreddie.com/darknet/darkgo-go-in-darknet/ deltaGo: http://home.q00.itscom.net/otsuki/delta.html Fudo Go: http://www.geocities.jp/hideki_katoh/ Fuego: http://fuego.sourceforge.net/ GnuGo: https://www.gnu.org/software/gnugo/devel.html Hara: https://github.com/antoniogarro/Hara Hiratuka: Non GPU version (10.37B): http://www.vector.co.jp/download/file/winnt/game/fh673259.html / GPU version (10.38B): http://www.vector.co.jp/download/file/winnt/game/fh688349.html IndiGo: http://www.math-info.univ-paris5.fr/~bouzy/INDIGO.html Iomrascálaí: https://github.com/ujh/iomrascalai JrefBot: http://ricoh51.free.fr/go/engineeng.htm Leela: https://sjeng.org/leela.html Marcos Go: https://github.com/MarcosPividori/Go-player Matilda: https://github.com/gonmf/matilda Michi C-2 https://github.com/db3108/michi-c2 MoGo: https://lifein19x19.com/forum/viewtopic.php?p=211091#p211091 Oakfoam: https://bitbucket.org/dsmic/oakfoam Orego: https://sites.google.com/a/lclark.edu/drake/research/orego Pachi: http://pachi.or.cz/ Rayon: https://github.com/zakki/Ray Stop: https://www.vanheusden.com/stop/ Best, Alex

Author:	Shoreline [ Mon Jul 31, 2017 8:19 pm ]
Post subject:	Re: Engine Tournament
How do you guys gets the Orego 7.08 running? I downloaded Jar file, unziped, or downloaded the 7.14 version release and javac compiled with no erros. However, I can not get the software running in linux Ubuntu 14.04. Erros: java -ea -Xmx1024M edu.lclark.orego.ui.Orego Error: Could not find or load main class edu.lclark.orego.ui.Orego No matter how I switched to src folder, or change to localhost.Orego or just Orego, it does not run

Author:	as0770 [ Mon Jul 31, 2017 9:34 pm ]
Post subject:	Re: Engine Tournament
Shoreline wrote: How do you guys gets the Orego 7.08 running? I downloaded Jar file, unziped, or downloaded the 7.14 version release and javac compiled with no erros. However, I can not get the software running in linux Ubuntu 14.04. Erros: java -ea -Xmx1024M edu.lclark.orego.ui.Orego Error: Could not find or load main class edu.lclark.orego.ui.Orego No matter how I switched to src folder, or change to localhost.Orego or just Orego, it does not run I had problems running later versions because of similar problems. v7.08 works by adding: Main-Class: orego.ui.Orego to /META-INF/MANIFEST.MF inside the .jar file. Then execute the .jar file by: java -jar /path to/orego-7.08.jar threads=1 grace If you get later versions to work let me know Best, Alex

Author:	Shoreline [ Tue Aug 01, 2017 5:17 am ]
Post subject:	Re: Engine Tournament
as0770 wrote: Shoreline wrote: How do you guys gets the Orego 7.08 running? I downloaded Jar file, unziped, or downloaded the 7.14 version release and javac compiled with no erros. However, I can not get the software running in linux Ubuntu 14.04. Erros: java -ea -Xmx1024M edu.lclark.orego.ui.Orego Error: Could not find or load main class edu.lclark.orego.ui.Orego No matter how I switched to src folder, or change to localhost.Orego or just Orego, it does not run I had problems running later versions because of similar problems. v7.08 works by adding: Main-Class: orego.ui.Orego to /META-INF/MANIFEST.MF inside the .jar file. Then execute the .jar file by: java -jar /path to/orego-7.08.jar threads=1 grace If you get later versions to work let me know Best, Alex Thanks, I got it running like you said. However, I found it is very weak bot at 19x19. How come this Orego 7.08 ranks so high against gnugo? It appears that it is several stones weaker than gnugo.

Author:	as0770 [ Tue Aug 01, 2017 12:54 pm ]
Post subject:	Re: Engine Tournament
Shoreline wrote: Thanks, I got it running like you said. However, I found it is very weak bot at 19x19. How come this Orego 7.08 ranks so high against gnugo? It appears that it is several stones weaker than gnugo. Here it won 4-0 against GnuGo 3.8, the games where all close. I think one has to play much more games to see which one is stronger.

Author:	Shoreline [ Tue Aug 01, 2017 1:31 pm ]
Post subject:	Re: Engine Tournament
I wonder if the java version or OS made the difference. I am using a linux machine, dual core, 8G of memory, quite low end machine. The java verison I am using is: $ java -version java version "1.7.0_45" Java(TM) SE Runtime Environment (build 1.7.0_45-b18) Java HotSpot(TM) 64-Bit Server VM (build 24.45-b08, mixed mode) I would say in my setting, the Orego version 7.08 is probably more than 9 stones weaker than gnugo, it probably ranks below 20 kyu in DGS.

Author:	Cyan [ Tue Aug 01, 2017 4:58 pm ]
Post subject:	Re: Engine Tournament
Here are estimates of goratings elo of different bots (from various sources and also my own testing). GPU is assumed to be high-end (e.g. 1080 Ti). Good to put things into perspective. Code: AlphaGo ~4800 elo Ke Jie ~3700 elo DeepZenGo ~3600 elo Average pro =3000 elo Ray GPU ~2700 elo Leela GPU ~2500 elo Hira GPU ~2300 elo Leela ~2100 elo Zen 6 ~2100 elo Hira ~1900 elo DarkGo ~1600 elo Pachi ~1300 elo 1k player ~1000 elo

Author:	as0770 [ Wed Aug 02, 2017 7:11 am ]
Post subject:	Re: Engine Tournament
Shoreline wrote: I wonder if the java version or OS made the difference. I am using a linux machine, dual core, 8G of memory, quite low end machine. The java verison I am using is: $ java -version java version "1.7.0_45" Java(TM) SE Runtime Environment (build 1.7.0_45-b18) Java HotSpot(TM) 64-Bit Server VM (build 24.45-b08, mixed mode) I would say in my setting, the Orego version 7.08 is probably more than 9 stones weaker than gnugo, it probably ranks below 20 kyu in DGS. This is my Java version: java -version openjdk version "1.8.0_131" OpenJDK Runtime Environment (build 1.8.0_131-8u131-b11-2ubuntu1.16.04.3-b11) OpenJDK 64-Bit Server VM (build 25.131-b11, mixed mode) Does Orego use all the available time? Maybe you use a time setting that Orego doesn't support and it moves faster.

Author:	Shoreline [ Wed Aug 02, 2017 10:19 am ]
Post subject:	Re: Engine Tournament
as0770 wrote: Shoreline wrote: I wonder if the java version or OS made the difference. I am using a linux machine, dual core, 8G of memory, quite low end machine. The java verison I am using is: $ java -version java version "1.7.0_45" Java(TM) SE Runtime Environment (build 1.7.0_45-b18) Java HotSpot(TM) 64-Bit Server VM (build 24.45-b08, mixed mode) I would say in my setting, the Orego version 7.08 is probably more than 9 stones weaker than gnugo, it probably ranks below 20 kyu in DGS. This is my Java version: java -version openjdk version "1.8.0_131" OpenJDK Runtime Environment (build 1.8.0_131-8u131-b11-2ubuntu1.16.04.3-b11) OpenJDK 64-Bit Server VM (build 25.131-b11, mixed mode) Does Orego use all the available time? Maybe you use a time setting that Orego doesn't support and it moves faster. Ｔhe Orego in my machine moves bleedingly fast, may be 1 second per move. It can play 5 or 10 different games within 10 seconds. How can I change the time setting in GTP command? Maybe that is the issue. Checked GTP manual and list_commands of Orego, found 2 commands is related "time_settings" "time_left". Not sure how to use them in games.

Page 5 of 20	All times are UTC - 8 hours [ DST ]
Powered by phpBB © 2000, 2002, 2005, 2007 phpBB Group http://www.phpbb.com/

Author:	as0770 [ Wed Aug 02, 2017 10:31 am ]
Post subject:	Re: Engine Tournament
Shoreline wrote: as0770 wrote: Shoreline wrote: I wonder if the java version or OS made the difference. I am using a linux machine, dual core, 8G of memory, quite low end machine. The java verison I am using is: $ java -version java version "1.7.0_45" Java(TM) SE Runtime Environment (build 1.7.0_45-b18) Java HotSpot(TM) 64-Bit Server VM (build 24.45-b08, mixed mode) I would say in my setting, the Orego version 7.08 is probably more than 9 stones weaker than gnugo, it probably ranks below 20 kyu in DGS. This is my Java version: java -version openjdk version "1.8.0_131" OpenJDK Runtime Environment (build 1.8.0_131-8u131-b11-2ubuntu1.16.04.3-b11) OpenJDK 64-Bit Server VM (build 25.131-b11, mixed mode) Does Orego use all the available time? Maybe you use a time setting that Orego doesn't support and it moves faster. Ｔhe Orego in my machine moves bleedingly fast, may be 1 second per move. It can play 5 or 10 different games within 10 seconds. How can I change the time setting in GTP command? Maybe that is the issue. I use game in x for my testing because this works with most engines. The gtp command is: time_settings 3600 0 0 for a game in 60 minutes. But Orego uses some more time as allowed.

Author:	Shoreline [ Wed Aug 02, 2017 11:53 am ]
Post subject:	Re: Engine Tournament
as0770 wrote: I use game in x for my testing because this works with most engines. The gtp command is: time_settings 3600 0 0 for a game in 60 minutes. But Orego uses some more time as allowed. THanks for the information. I think you nailed the issue already correctly. I tried your setting and found Orego runs about 40 seconds per move and it is much stronger player. Now I settled down to time_settings 600 60 0 roughly at 10 to 20 seconds per move in my machine. Here is the DGS bot "oregobot" with same setting: http://www.dragongoserver.net/ratinggraph.php?uid=99178 We will see how strong this bot can be. My internal testing says that this bot should at least get teen kyu level.

Author:	as0770 [ Wed Aug 23, 2017 1:45 pm ]
Post subject:	Re: Engine Tournament
Update League B with Pachi DCNN, DarkGo and Ray 9.0.1 League A: Code: 1. Leela 0.10.0 22/24 2. Rayon 4.6.0 19/24 3. Oakfoam 0.2.1 NG-06 18/24 4. Hiratuka 10.37B (CPU) 9/24 5. DarkForest v2 MCTS 1.0 7/24 6 DarkGo 1.0 5/24 7. Pachi DCNN 11.99 4/24 Configuration: 2h/game, pondering off, 4 threads, 2GB on 4 x Intel® Core™ i5-4210H CPU @ 2.90GHz, 8 GiB Ram and GeForce 840M/PCIe/SSE2 DarkForest: taskset -c 0 bash cnn_evaluator.sh 1 /data/local/go and taskset -c 0 th cnnPlayerMCTSV2.lua --num_gpu 1 --num_tree_thread 1 --rollout 2500 --win_rate_thres 0.1 DarkGo: darknet go engine cfg/go.test.cfg go.weights Hiratuka: Hiratuka-19×19.exe -po 175000 Leela: leela_gtp_opencl –gtp –threads 4 –noponder Oakfoam: oakfoam -c nicego-cnn-06.gtp Code: nicego-cnn-06.gtp: param playouts_per_move_max 40000 param thread_count 4 Pachi: pachidcnn -f pachibook.dat threads=4,max_tree_size=2048,pondering=0 Rayon: rayon --thread 4 --no-debug twogtp: TWOGTP=”gogui-twogtp -black \”$BLACK\” -white \”$WHITE\” -games 2 -size 19 -time 120 -sgffile xxxx” gogui -size 19 -program “$TWOGTP” -computer-both -auto League B: Code: 1. DarkGo 1.0 29/32 2. Pachi DCNN 11.99 27/32 3. Ray 9.0.1 25/32 4. MoGo 4.86 19/32 5. deltaGo 1.0.0 16/32 6. Fuego 1.1 13/32 7. Michi C-2 1.4.2 8/32 8. Orego 7.08 5/32 9. GNU Go 3.8 2/32 Configuration: 1h/game, pondering off, 1 thread, 1GB on 4 x Intel® Core™ i5-4210H CPU @ 2.90GHz, 8 GiB Ram and GeForce 840M/PCIe/SSE2 deltaGo: deltaGo.exe Fuego: fuego.exe –config fuego.cfg Code: fuego.cfg: uct_param_search number_threads 1 uct_param_search lock_free 0 uct_max_memory 1024000000 uct_param_player reuse_subtree 1 uct_param_player ponder 0 uct_param_player early_pass 1 GnuGo: gnugo --mode gtp --level 10 --resign-allowed Michi C-2: michi gtp Code: ui.c: init_large_patterns("patterns2.prob", "patterns2.spat"); // Michis pattern files renamed because they have the same name as Pachis files. MoGo: mogo Orego: java -jar /path/to/orego-7.08.jar threads=1 grace Pachi: pachi -f pachibook.dat threads=1,max_tree_size=1024,pondering=0 Ray: ray --time 3600 --thread 1 --no-debug twogtp: TWOGTP=”gogui-twogtp -black \”$BLACK\” -white \”$WHITE\” -games 2 -size 19 -time 60 -sgffile xxxx” gogui -size 19 -program “$TWOGTP” -computer-both -auto League C: Code: 1. GNU Go 3.8 24/28 2. Hara 0.9 18/28 3. Dariush 3.1.5.7 16/28 4. Indigo 2009 15/28 5. Matilda 1.24 15/28 6. Aya 6.34 11/28 7. Fudo Go 3.0 11/28 8. JrefBot 081016-2022 2/28 Configuration: 1h/game, pondering off, 1 thread, 1GB on 4 x Intel® Core™ i5-4210H CPU @ 2.90GHz, 8 GiB Ram and GeForce 840M/PCIe/SSE2 Aya: Aya.exe –mode gtp –level max Dariush: DarGTP.exe –level 10 Fudo Go: taskset -c 0 fudo –boardsize=19 –komi=6.5 GnuGo: gnugo --mode gtp --level 10 --resign-allowed Hara: hara IndiGo: Indigo.exe -gtp JrefBot: java -jar jrefgo.jar 10000 Matilda: matilda Code: matilda.h: #define BOARD_SIZ 19 #define DEFAULT_UCT_MEMORY 1000 #define DEFAULT_NUM_THREADS 1 twogtp: TWOGTP=”gogui-twogtp -black \”$BLACK\” -white \”$WHITE\” -games 2 -size 19 -time 60 -sgffile xxxx” gogui -size 19 -program “$TWOGTP” -computer-both -auto League D: Code: 1. JrefBot 081016-2022 25/28 2. Iomrascálaí 0.3.2 21/28 3. Crazy Patterns 0008-13 19/28 4. Marcos Go 1.0 18/28 5. AmiGo 1.8 16/28 6. Stop 0.9-005 6/28 7. CopyBot 0.1 4/28 8. Brown 1.0 3/28 Configuration: 1h/game, pondering off, 1 thread, 1GB on 4 x Intel® Core™ i5-4210H CPU @ 2.90GHz, 8 GiB Ram and GeForce 840M/PCIe/SSE2 Amigo: amigogtp Brown: brown.exe Copybot: python /path/to/__main__.py CrazyPatterns: CrazyPatterns.exe Iomrascálaí: taskset -c 0,1 iomrascalai JrefBot: java -jar jrefgo.jar 10000 Marcos Go: marcos_go --patterns /path/to/patterns.txt --cycles_mcts 10000 --threads_mcts 1 Stop: /usr/bin/java -ea -jar /path/to/stop-09-005.jar --mode gtp twogtp: TWOGTP=”gogui-twogtp -black \”$BLACK\” -white \”$WHITE\” -games 2 -size 19 -time 60 -sgffile xxxx” gogui -size 19 -program “$TWOGTP” -computer-both -auto Links: Amigo: https://sourceforge.net/projects/amigogtp/ Aya: http://www.yss-aya.com/ Brown: http://ricoh51.free.fr/go/engineeng.htm Copybot: https://github.com/sirtango/ICopyMoves CrazyPatterns: https://www.remi-coulom.fr/Amsterdam2007/ Dariush: http://ricoh51.free.fr/go/engineeng.htm DarkForest: https://github.com/facebookresearch/darkforestGo DarkGo: https://pjreddie.com/darknet/darkgo-go-in-darknet/ deltaGo: http://home.q00.itscom.net/otsuki/delta.html Fudo Go: http://www.geocities.jp/hideki_katoh/ Fuego: http://fuego.sourceforge.net/ GnuGo: https://www.gnu.org/software/gnugo/devel.html Hara: https://github.com/antoniogarro/Hara Hiratuka: Non GPU version (10.37B): http://www.vector.co.jp/download/file/winnt/game/fh673259.html / GPU version (10.38B): http://www.vector.co.jp/download/file/winnt/game/fh688349.html IndiGo: http://www.math-info.univ-paris5.fr/~bouzy/INDIGO.html Iomrascálaí: https://github.com/ujh/iomrascalai JrefBot: http://ricoh51.free.fr/go/engineeng.htm Leela: https://sjeng.org/leela.html Marcos Go: https://github.com/MarcosPividori/Go-player Matilda: https://github.com/gonmf/matilda Michi C-2 https://github.com/db3108/michi-c2 MoGo: https://lifein19x19.com/forum/viewtopic.php?p=211091#p211091 Oakfoam: https://bitbucket.org/dsmic/oakfoam Orego: https://sites.google.com/a/lclark.edu/drake/research/orego Pachi: http://pachi.or.cz/ Rayon: https://github.com/zakki/Ray Ray: https://github.com/kobanium/Ray Stop: https://www.vanheusden.com/stop/ Best, Alex

Author:	anpd [ Mon Aug 28, 2017 11:18 pm ]
Post subject:	Re: Engine Tournament
Is there a version of Pachi with neural networks available for Windows somewhere? Do you use patterns with that build?

Author:	q30 [ Sat Sep 30, 2017 12:07 am ]
Post subject:	Re: Engine Tournament
Cyan wrote: Please add DarkGo! It's a neural network at around 1d level. https://pjreddie.com/darknet/darkgo-go-in-darknet/ How to run in GTP: https://groups.google.com/forum/#!topic ... BNZTkMmUHI I had downloaded DarkGo. I had downloaded also CUDA. Unfortunately CUDA requires NVIDIA driver v.375 or higher, but v.375 driver tells, that with my video card can work drivers v.340 or lower (card has 32 CUDA cores and 256 MB memory)... I had downloaded also go.train and go.weights. And there is a lot of cfg files, but it isn't any description. Can anybody tell me, how is it possible to run darknet in gtp mode with maximum strong (including pondering, neural network), using 2 CPU cores, up to 8 GB memory and up to 2 min / move time?

Author:	as0770 [ Sat Sep 30, 2017 10:46 pm ]
Post subject:	Re: Engine Tournament
Unfortunately AQ can't play my tournament because of some issues: AQ always ponders and it uses the GPU very effectively so that other engines get less GPU power. Beside that Rayon and Oakfoam don't work together with AQ because of a cuda memory error. It was possible to set up a match between Leela and AQ with ponder on for both engines where Leela reaches around 35% of it's playouts with full power. So I gave AQ 2h/game and Leela 4h/game. Still I can't control what is happening on the GPU during the games, but I think the conditions are as fair as possible. BTW DarkGo was able to win one of four games against AQ during my tests. Update with the next Leela version will follow Leela vs. AQ Code: 1. AQ 2.0.2 5/8 2. Leela 0.10.0 3/8 Configuration: 2h/game (Leela 4h/game because Leela gets only ~35% and AQ ~65% GPU time), pondering on (can't be switched off in AQ), 4 threads, 2 (AQ 4) GB on 4 x Intel® Core™ i5-4210H CPU @ 2.90GHz, 8 GiB Ram and GeForce 840M/PCIe/SSE2 AQ: AQ Code: aq_config.txt: -main　time[sec] =7200 -time controll =true -japanese rule =true Leela: leela_gtp_opencl –gtp –threads 4 twogtp: gogui-twogtp -auto -white "leela_gtp_opencl --gtp --threads 4" -black "AQ" -games 4 -size 19 -time 240 -verbose -sgffile xxxx League A: Code: 1. Leela 0.10.0 22/24 2. Rayon 4.6.0 19/24 3. Oakfoam 0.2.1 NG-06 18/24 4. Hiratuka 10.37B (CPU) 9/24 5. DarkForest v2 MCTS 1.0 7/24 6 DarkGo 1.0 5/24 7. Pachi DCNN 11.99 4/24 Configuration: 2h/game, pondering off, 4 threads, 2GB on 4 x Intel® Core™ i5-4210H CPU @ 2.90GHz, 8 GiB Ram and GeForce 840M/PCIe/SSE2 DarkForest: taskset -c 0 bash cnn_evaluator.sh 1 /data/local/go and taskset -c 0 th cnnPlayerMCTSV2.lua --num_gpu 1 --num_tree_thread 1 --rollout 2500 --win_rate_thres 0.1 DarkGo: darknet go engine cfg/go.test.cfg go.weights Hiratuka: Hiratuka-19×19.exe -po 175000 Leela: leela_gtp_opencl –gtp –threads 4 –noponder Oakfoam: oakfoam -c nicego-cnn-06.gtp Code: nicego-cnn-06.gtp: param playouts_per_move_max 40000 param thread_count 4 Pachi: pachidcnn -f pachibook.dat threads=4,max_tree_size=2048,pondering=0 Rayon: rayon --thread 4 --no-debug twogtp: TWOGTP=”gogui-twogtp -black \”$BLACK\” -white \”$WHITE\” -games 2 -size 19 -time 120 -sgffile xxxx” gogui -size 19 -program “$TWOGTP” -computer-both -auto League B: Code: 1. Pachi DCNN 11.99 25/28 2. Ray 9.0.1 25/28 3. MoGo 4.86 18/28 4. deltaGo 1.0.0 16/28 5. Fuego 1.1 13/28 6. Michi C-2 1.4.2 8/28 7. Orego 7.08 5/28 8. GNU Go 3.8 2/28 Configuration: 1h/game, pondering off, 1 thread, 1GB on 4 x Intel® Core™ i5-4210H CPU @ 2.90GHz, 8 GiB Ram and GeForce 840M/PCIe/SSE2 deltaGo: deltaGo.exe Fuego: fuego.exe –config fuego.cfg Code: fuego.cfg: uct_param_search number_threads 1 uct_param_search lock_free 0 uct_max_memory 1024000000 uct_param_player reuse_subtree 1 uct_param_player ponder 0 uct_param_player early_pass 1 GnuGo: gnugo --mode gtp --level 10 --resign-allowed Michi C-2: michi gtp Code: ui.c: init_large_patterns("patterns2.prob", "patterns2.spat"); // Michis pattern files renamed because they have the same name as Pachis files. MoGo: mogo Orego: java -jar /path/to/orego-7.08.jar threads=1 grace Pachi: pachi -f pachibook.dat threads=1,max_tree_size=1024,pondering=0 Ray: ray --time 3600 --thread 1 --no-debug twogtp: TWOGTP=”gogui-twogtp -black \”$BLACK\” -white \”$WHITE\” -games 2 -size 19 -time 60 -sgffile xxxx” gogui -size 19 -program “$TWOGTP” -computer-both -auto League C: Code: 1. GNU Go 3.8 24/28 2. Hara 0.9 18/28 3. Dariush 3.1.5.7 16/28 4. Indigo 2009 15/28 5. Matilda 1.24 15/28 6. Aya 6.34 11/28 7. Fudo Go 3.0 11/28 8. JrefBot 081016-2022 2/28 Configuration: 1h/game, pondering off, 1 thread, 1GB on 4 x Intel® Core™ i5-4210H CPU @ 2.90GHz, 8 GiB Ram and GeForce 840M/PCIe/SSE2 Aya: Aya.exe –mode gtp –level max Dariush: DarGTP.exe –level 10 Fudo Go: taskset -c 0 fudo –boardsize=19 –komi=6.5 GnuGo: gnugo --mode gtp --level 10 --resign-allowed Hara: hara IndiGo: Indigo.exe -gtp JrefBot: java -jar jrefgo.jar 10000 Matilda: matilda Code: matilda.h: #define BOARD_SIZ 19 #define DEFAULT_UCT_MEMORY 1000 #define DEFAULT_NUM_THREADS 1 twogtp: TWOGTP=”gogui-twogtp -black \”$BLACK\” -white \”$WHITE\” -games 2 -size 19 -time 60 -sgffile xxxx” gogui -size 19 -program “$TWOGTP” -computer-both -auto League D: Code: 1. JrefBot 081016-2022 25/28 2. Iomrascálaí 0.3.2 21/28 3. Crazy Patterns 0008-13 19/28 4. Marcos Go 1.0 18/28 5. AmiGo 1.8 16/28 6. Stop 0.9-005 6/28 7. CopyBot 0.1 4/28 8. Brown 1.0 3/28 Configuration: 1h/game, pondering off, 1 thread, 1GB on 4 x Intel® Core™ i5-4210H CPU @ 2.90GHz, 8 GiB Ram and GeForce 840M/PCIe/SSE2 Amigo: amigogtp Brown: brown.exe Copybot: python /path/to/__main__.py CrazyPatterns: CrazyPatterns.exe Iomrascálaí: taskset -c 0,1 iomrascalai JrefBot: java -jar jrefgo.jar 10000 Marcos Go: marcos_go --patterns /path/to/patterns.txt --cycles_mcts 10000 --threads_mcts 1 Stop: /usr/bin/java -ea -jar /path/to/stop-09-005.jar --mode gtp twogtp: TWOGTP=”gogui-twogtp -black \”$BLACK\” -white \”$WHITE\” -games 2 -size 19 -time 60 -sgffile xxxx” gogui -size 19 -program “$TWOGTP” -computer-both -auto Links: Amigo: https://sourceforge.net/projects/amigogtp/ AQ: https://github.com/ymgaq/AQ Aya: http://www.yss-aya.com/ Brown: http://ricoh51.free.fr/go/engineeng.htm Copybot: https://github.com/sirtango/ICopyMoves CrazyPatterns: https://www.remi-coulom.fr/Amsterdam2007/ Dariush: http://ricoh51.free.fr/go/engineeng.htm DarkForest: https://github.com/facebookresearch/darkforestGo DarkGo: https://pjreddie.com/darknet/darkgo-go-in-darknet/ deltaGo: http://home.q00.itscom.net/otsuki/delta.html Fudo Go: http://www.geocities.jp/hideki_katoh/ Fuego: http://fuego.sourceforge.net/ GnuGo: https://www.gnu.org/software/gnugo/devel.html Hara: https://github.com/antoniogarro/Hara Hiratuka: Non GPU version (10.37B): http://www.vector.co.jp/download/file/winnt/game/fh673259.html / GPU version (10.38B): http://www.vector.co.jp/download/file/winnt/game/fh688349.html IndiGo: http://www.math-info.univ-paris5.fr/~bouzy/INDIGO.html Iomrascálaí: https://github.com/ujh/iomrascalai JrefBot: http://ricoh51.free.fr/go/engineeng.htm Leela: https://sjeng.org/leela.html Marcos Go: https://github.com/MarcosPividori/Go-player Matilda: https://github.com/gonmf/matilda Michi C-2 https://github.com/db3108/michi-c2 MoGo: https://lifein19x19.com/forum/viewtopic.php?p=211091#p211091 Oakfoam: https://bitbucket.org/dsmic/oakfoam Orego: https://sites.google.com/a/lclark.edu/drake/research/orego Pachi: http://pachi.or.cz/ Rayon: https://github.com/zakki/Ray Ray: https://github.com/kobanium/Ray Stop: https://www.vanheusden.com/stop/ Best, Alex