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

KataGo 22946962 playouts
http://www.lifein19x19.com/viewtopic.php?f=18&t=17319		 Page 1 of 1
Author:  goame [ Sat Mar 14, 2020 3:30 am ]
Post subject:  KataGo 22946962 playouts
2x RTX 2080 Ti
~60 out of 64 GB RAM
~3 hours
g170-b30c320x2-s1287828224-d525929064.bin.gz 193 MB

Attachments:
Go, Baduk, Weiqi 2.jpg
Go, Baduk, Weiqi 2.jpg [ 1.96 MiB | Viewed 15253 times ]
Author:  SoDesuNe [ Sat Mar 14, 2020 3:43 am ]
Post subject:  Re: KataGo 22946962 playouts
Looks... conventional? :study:
Author:  go4thewin [ Sat Mar 14, 2020 4:35 am ]
Post subject:  Re: KataGo 22946962 playouts
I wonder what would be the result with 5.5 komi. I like games where black has the advantage
Author:  Limeztone [ Sat Mar 14, 2020 6:50 am ]
Post subject:  Re: KataGo 22946962 playouts
Rules?
Komi?
Threads?
Author:  Bill Spight [ Sat Mar 14, 2020 8:09 am ]
Post subject:  Re: KataGo 22946962 playouts
SoDesuNe wrote:
Looks... conventional? :study:


Maybe so. But :w6: has no hits on Waltheri (ps.waltheri.net). And AlphaGoTeach plays the taisha in the top right for :b7:.

I suppose that the sequence up to :w10: is the principal variation. If so, how many rollouts does :w10: get?
Author:  ez4u [ Sat Mar 14, 2020 5:44 pm ]
Post subject:  Re: KataGo 22946962 playouts
Excellent example of the limitations of bots? Even with 23 million playouts, katago calculates one corner is superior to the other three (only a little, but blue is clear)!

It would have been interesting to immediately click through the variation shown and see how many of the 23M playouts were used in following the main line beyond :w10: and how many were used on alternatives along the way.
Author:  Bill Spight [ Sat Mar 14, 2020 6:17 pm ]
Post subject:  Re: KataGo 22946962 playouts
ez4u wrote:
It would have been interesting to immediately click through the variation shown and see how many of the 23M playouts were used in following the main line beyond :w10: and how many were used on alternatives along the way.


Yes. And with 6 million playouts in the main line, if we follow the example of the Elf commentaries and extend the main line until the next play in the main line would have fewer than 1500 playouts, the main line would be very long, I expect. :) Also very interesting. :)
Author:  goame [ Sat Mar 14, 2020 11:39 pm ]
Post subject:  Re: KataGo 22946962 playouts
go4thewin wrote:
I wonder what would be the result with 5.5 komi.


That's a very good question.
Maybe this would be a fair play.
Is it possible to change the komi for Lizzie/KataGo?
And how?
Author:  goame [ Sat Mar 14, 2020 11:44 pm ]
Post subject:  Re: KataGo 22946962 playouts
Limeztone wrote:
Rules?
Komi?
Threads?


Japanese
Komi I think 6.5
What do you mean by threads?

https://lifein19x19.com/viewtopic.php?f=18&t=17317
Tuning with 50000 visits:

Z:\>LG0\Lizzie\katago\katago.exe genconfig -model \LG0\Lizzie\katago\g170-b30c32
0x2-s1287828224-d525929064.bin.gz -output gtp_custom.cfg

=========================================================================
RULES

What rules should KataGo use by default for play and analysis?
(chinese, japanese, korean, tromp-taylor, aga, chinese-ogs, new-zealand, bga, st
one-scoring, aga-button):
japanese

=========================================================================
SEARCH LIMITS

When playing games, KataGo will always obey the time controls given by the GUI/t
ournament/match/online server.
But you can specify an additional limit to make KataGo move much faster. This do
es NOT affect analysis/review,
only affects playing games. Add a limit? (y/n) (default n):
n

NOTE: No limits configured for KataGo. KataGo will obey time controls provided b
y the GUI or server or match script
but if they don't specify any, when playing games KataGo may think forever witho
ut moving. (press enter to continue)


When playing games, KataGo can optionally ponder during the opponent's turn. Thi
s gives faster/stronger play
in real games but should NOT be enabled if you are running tests with fixed limi
ts (pondering may exceed those
limits), or to avoid stealing the opponent's compute time when testing two bots
on the same machine.
Enable pondering? (y/n, default n):y

Specify max num seconds KataGo should ponder during the opponent's turn. Leave b
lank for no limit:


=========================================================================
GPUS AND RAM

Finding available GPU-like devices...
Found CUDA device 0: GeForce RTX 2080 Ti
Found CUDA device 1: GeForce RTX 2080 Ti

Specify devices/GPUs to use (for example "0,1,2" to use devices 0, 1, and 2). Le
ave blank for good default:
"0,1"
could not parse int: "0
Specify devices/GPUs to use (for example "0,1,2" to use devices 0, 1, and 2). Le
ave blank for good default:
0,1

By default, KataGo will cache up to about 3GB of positions in memory (RAM), in a
ddition to
whatever the current search is using. Specify a max in GB or leave blank for def
ault:
60

=========================================================================
PERFORMANCE TUNING

Specify number of visits to use test/tune performance with, leave blank for defa
ult based on GPU speed.
Use large number for more accurate results, small if your GPU is old and this is
taking forever:
50000

Specify number of seconds/move to optimize performance for (default 5), leave bl
ank for default:

2020-03-12 22:55:26+0100: Loading model and initializing benchmark...

=========================================================================
TUNING NOW
Tuning using 50000 visits.
Automatically trying different numbers of threads to home in on the best:

2020-03-12 22:55:26+0100: nnRandSeed0 = 2369906978592220054
2020-03-12 22:55:26+0100: After dedups: nnModelFile0 = \LG0\Lizzie\katago\g170-b
30c320x2-s1287828224-d525929064.bin.gz useFP16 auto useNHWC auto
2020-03-12 22:55:28+0100: Cuda backend: Found GPU GeForce RTX 2080 Ti memory 118
11160064 compute capability major 7 minor 5
2020-03-12 22:55:28+0100: Cuda backend: Found GPU GeForce RTX 2080 Ti memory 118
11160064 compute capability major 7 minor 5
2020-03-12 22:55:28+0100: Cuda backend: Model version 8 useFP16 = true useNHWC =
true
2020-03-12 22:55:28+0100: Cuda backend: Model name: g170-b30c320x2-s1287828224-d
525929064
2020-03-12 22:55:28+0100: Cuda backend: Model version 8 useFP16 = true useNHWC =
true
2020-03-12 22:55:28+0100: Cuda backend: Model name: g170-b30c320x2-s1287828224-d
525929064


Possible numbers of threads to test: 1, 2, 3, 4, 5, 6, 8, 10, 12, 16, 20, 24, 32
,

numSearchThreads = 5: 10 / 10 positions, visits/s = 533.10 nnEvals/s = 350.16 n
nBatches/s = 213.88 avgBatchSize = 1.64 (938.0 secs)
numSearchThreads = 12: 10 / 10 positions, visits/s = 1131.75 nnEvals/s = 769.38
nnBatches/s = 198.99 avgBatchSize = 3.87 (441.9 secs)
numSearchThreads = 10: 10 / 10 positions, visits/s = 964.41 nnEvals/s = 649.12 n
nBatches/s = 204.31 avgBatchSize = 3.18 (518.5 secs)
numSearchThreads = 20: 10 / 10 positions, visits/s = 1520.41 nnEvals/s = 1003.61
nnBatches/s = 152.46 avgBatchSize = 6.58 (329.0 secs)
numSearchThreads = 16: 10 / 10 positions, visits/s = 1387.92 nnEvals/s = 932.16
nnBatches/s = 178.77 avgBatchSize = 5.21 (360.4 secs)
numSearchThreads = 24: 10 / 10 positions, visits/s = 1624.20 nnEvals/s = 1089.80
nnBatches/s = 136.46 avgBatchSize = 7.99 (308.0 secs)
numSearchThreads = 32: 10 / 10 positions, visits/s = 1796.26 nnEvals/s = 1201.35
nnBatches/s = 113.86 avgBatchSize = 10.55 (278.5 secs)


Optimal number of threads is fairly high, tripling the search limit and trying a
gain.

2020-03-12 23:49:10+0100: nnRandSeed0 = 6506758374797114957
2020-03-12 23:49:10+0100: After dedups: nnModelFile0 = \LG0\Lizzie\katago\g170-b
30c320x2-s1287828224-d525929064.bin.gz useFP16 auto useNHWC auto
2020-03-12 23:49:13+0100: Cuda backend: Found GPU GeForce RTX 2080 Ti memory 118
11160064 compute capability major 7 minor 5
2020-03-12 23:49:13+0100: Cuda backend: Found GPU GeForce RTX 2080 Ti memory 118
11160064 compute capability major 7 minor 5
2020-03-12 23:49:13+0100: Cuda backend: Model version 8 useFP16 = true useNHWC =
true
2020-03-12 23:49:13+0100: Cuda backend: Model name: g170-b30c320x2-s1287828224-d
525929064
2020-03-12 23:49:13+0100: Cuda backend: Model version 8 useFP16 = true useNHWC =
true
2020-03-12 23:49:13+0100: Cuda backend: Model name: g170-b30c320x2-s1287828224-d
525929064


Possible numbers of threads to test: 1, 2, 3, 4, 5, 6, 8, 10, 12, 16, 20, 24, 32
, 40, 48, 64, 80, 96,

numSearchThreads = 6: 10 / 10 positions, visits/s = 626.73 nnEvals/s = 407.14 n
nBatches/s = 209.06 avgBatchSize = 1.95 (797.9 secs)
numSearchThreads = 48: 10 / 10 positions, visits/s = 2214.93 nnEvals/s = 1421.03
nnBatches/s = 93.34 avgBatchSize = 15.22 (226.0 secs)
numSearchThreads = 64: 10 / 10 positions, visits/s = 2301.42 nnEvals/s = 1500.58
nnBatches/s = 77.43 avgBatchSize = 19.38 (217.5 secs)
numSearchThreads = 80: 10 / 10 positions, visits/s = 2322.34 nnEvals/s = 1543.88
nnBatches/s = 65.55 avgBatchSize = 23.55 (215.6 secs)
numSearchThreads = 40: 10 / 10 positions, visits/s = 1983.09 nnEvals/s = 1353.57
nnBatches/s = 104.84 avgBatchSize = 12.91 (252.3 secs)


Ordered summary of results:

numSearchThreads = 5: 10 / 10 positions, visits/s = 533.10 nnEvals/s = 350.16 n
nBatches/s = 213.88 avgBatchSize = 1.64 (938.0 secs) (EloDiff baseline)
numSearchThreads = 6: 10 / 10 positions, visits/s = 626.73 nnEvals/s = 407.14 n
nBatches/s = 209.06 avgBatchSize = 1.95 (797.9 secs) (EloDiff +57)
numSearchThreads = 10: 10 / 10 positions, visits/s = 964.41 nnEvals/s = 649.12 n
nBatches/s = 204.31 avgBatchSize = 3.18 (518.5 secs) (EloDiff +208)
numSearchThreads = 12: 10 / 10 positions, visits/s = 1131.75 nnEvals/s = 769.38
nnBatches/s = 198.99 avgBatchSize = 3.87 (441.9 secs) (EloDiff +264)
numSearchThreads = 16: 10 / 10 positions, visits/s = 1387.92 nnEvals/s = 932.16
nnBatches/s = 178.77 avgBatchSize = 5.21 (360.4 secs) (EloDiff +334)
numSearchThreads = 20: 10 / 10 positions, visits/s = 1520.41 nnEvals/s = 1003.61
nnBatches/s = 152.46 avgBatchSize = 6.58 (329.0 secs) (EloDiff +362)
numSearchThreads = 24: 10 / 10 positions, visits/s = 1624.20 nnEvals/s = 1089.80
nnBatches/s = 136.46 avgBatchSize = 7.99 (308.0 secs) (EloDiff +381)
numSearchThreads = 32: 10 / 10 positions, visits/s = 1796.26 nnEvals/s = 1201.35
nnBatches/s = 113.86 avgBatchSize = 10.55 (278.5 secs) (EloDiff +408)
numSearchThreads = 40: 10 / 10 positions, visits/s = 1983.09 nnEvals/s = 1353.57
nnBatches/s = 104.84 avgBatchSize = 12.91 (252.3 secs) (EloDiff +436)
numSearchThreads = 48: 10 / 10 positions, visits/s = 2214.93 nnEvals/s = 1421.03
nnBatches/s = 93.34 avgBatchSize = 15.22 (226.0 secs) (EloDiff +471)
numSearchThreads = 64: 10 / 10 positions, visits/s = 2301.42 nnEvals/s = 1500.58
nnBatches/s = 77.43 avgBatchSize = 19.38 (217.5 secs) (EloDiff +467)
numSearchThreads = 80: 10 / 10 positions, visits/s = 2322.34 nnEvals/s = 1543.88
nnBatches/s = 65.55 avgBatchSize = 23.55 (215.6 secs) (EloDiff +451)


Based on some test data, each speed doubling gains perhaps ~250 Elo by searching
deeper.
Based on some test data, each thread costs perhaps 7 Elo if using 800 visits, an
d 2 Elo if using 5000 visits (by making MCTS worse).
So APPROXIMATELY based on this benchmark, if you intend to do a 5 second search:

numSearchThreads = 5: (baseline)
numSearchThreads = 6: +57 Elo
numSearchThreads = 10: +208 Elo
numSearchThreads = 12: +264 Elo
numSearchThreads = 16: +334 Elo
numSearchThreads = 20: +362 Elo
numSearchThreads = 24: +381 Elo
numSearchThreads = 32: +408 Elo
numSearchThreads = 40: +436 Elo
numSearchThreads = 48: +471 Elo (recommended)
numSearchThreads = 64: +467 Elo
numSearchThreads = 80: +451 Elo

Using 48 numSearchThreads!

=========================================================================
DONE

Writing new config file to gtp_custom.cfg
You should be now able to run KataGo with this config via something like:
LG0\Lizzie\katago\katago.exe gtp -model '\LG0\Lizzie\katago\g170-b30c320x2-s1287
828224-d525929064.bin.gz' -config 'gtp_custom.cfg'

Feel free to look at and edit the above config file further by hand in a txt edi
tor.
For more detailed notes about performance and what options in the config do, see
:
https://github.com/lightvector/KataGo/b ... xample.cfg
Author:  goame [ Sat Mar 14, 2020 11:49 pm ]
Post subject:  Re: KataGo 22946962 playouts
Bill Spight wrote:
SoDesuNe wrote:
Looks... conventional? :study:


Maybe so. But :w6: has no hits on Waltheri (ps.waltheri.net). And AlphaGoTeach plays the taisha in the top right for :b7:.

I suppose that the sequence up to :w10: is the principal variation. If so, how many rollouts does :w10: get?



-no hits, because this game is way to difficult to find every move, even in the beginning.
-ps.walther.net is a "nice to have" and not the "truth about Go".
Author:  goame [ Sat Mar 14, 2020 11:53 pm ]
Post subject:  Re: KataGo 22946962 playouts
ez4u wrote:
Excellent example of the limitations of bots? Even with 23 million playouts, katago calculates one corner is superior to the other three (only a little, but blue is clear)!


Is this good or bad?

I think you mean that it would be better, if KataGo will focus on one corner and then copy and paste the results to the other corners.
Author:  goame [ Sun Mar 15, 2020 12:02 am ]
Post subject:  Re: KataGo 22946962 playouts
I need help:

How to kill the circles on the picture???

I want the have only the 3 circles from every corner.

I don't need this SPACE INVADERS version.

You should know that this SPACE INVADERS attack at the beginning of the game is the most pleasant one.
Author:  goame [ Sun Mar 15, 2020 1:06 am ]
Post subject:  Re: KataGo 22946962 playouts
Bug or feature?

I see every time at the start of an analyse the visits per second.

But after some time I see the visits every two seconds and the other second I see 0 visits.
And after some time I see the visits every three seconds and this is inreasing.

But it's not a loss of visits.
It's more like:
Every 1 second = 4000 visits
Every 2 seconds = 8000 visits
Every 3 seconds = 12000 visits
And this is increasing.

On my picture at the top you also see 0 visits, but I haven't stopped the analysis.
Author:  lightvector [ Sun Mar 15, 2020 5:23 am ]
Post subject:  Re: KataGo 22946962 playouts
Try turning off the ownership calculation unless you specifically need it. It hurts performance and slows KataGo's reporting down, particularly when the search tree grows large. See https://github.com/lightvector/KataGo/issues/155

In Lizzie, turning it off is the "." button, or you can do it in one of the menus at the top.
Author:  goame [ Sun Mar 15, 2020 6:00 am ]
Post subject:  Re: KataGo 22946962 playouts
lightvector wrote:
Try turning off the ownership calculation unless you specifically need it. It hurts performance and slows KataGo's reporting down, particularly when the search tree grows large. See https://github.com/lightvector/KataGo/issues/155

In Lizzie, turning it off is the "." button, or you can do it in one of the menus at the top.


Thx. The button works.
For what is the ownership calculation?
Author:  lightvector [ Sun Mar 15, 2020 6:02 am ]
Post subject:  Re: KataGo 22946962 playouts
The ownership estimation is what Lizzie shows as the black and white overlaid squares that indicate what areas will become owned in the future by black or white. Try going to a non-empty board position and pressing "." to toggle it back and forth to take a look.

Calculating and tracking ownership predictions during the search can be expensive, which is why turning it off improves performance.
Page 1 of 1 All times are UTC - 8 hours [ DST ]
Powered by phpBB © 2000, 2002, 2005, 2007 phpBB Group
http://www.phpbb.com/