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

zen7 30 second : leela-zero 30 second, Zen7 won consistently
https://lifein19x19.com/viewtopic.php?f=18&t=15599
Page 2 of 3

Author:  tchan001 [ Thu Apr 12, 2018 6:27 pm ]
Post subject:  Re: zen7 30 second : leela-zero 30 second, Zen7 won consiste

How can it be wrong when he is only reporting the results of his tests on his own system. As he provides details of the configuration of the computer and of the settings he uses, you are responsible for making your own judgments on how useful the results are for yourself. Of course, you are free to offer results from tests done on your own system and settings which may differ significantly from what he achieved.

Author:  kyotosato [ Thu Apr 12, 2018 6:43 pm ]
Post subject:  Re: zen7 30 second : leela-zero 30 second, Zen7 won consiste

comments noted, it is still a CPU/GPU prone problem/discussion, as i said result(s) is inclusive. anyone use LZ plays with in realtime with real 9dan players (pro or ama?) i don't mean online LZ bots. I did a lot games with 8-9 dans in fox and temgen.
for among go computer programs, my test is just to test strength of them for my own use. As an AI practitioner, my interests are NN, optimization, algorithm, subjective statistics among others..

Author:  Vargo [ Thu Apr 12, 2018 9:02 pm ]
Post subject:  Re: zen7 30 second : leela-zero 30 second, Zen7 won consiste

Three handicap games between LZ and Zen6 (Sabaki 0.33.4, win 10, 2x1080Ti, i9-12 core, 10" per move, no pondering)

H2, H3, and H4 : gtp4zen6 (B) v LZ13_5d6d9 (W)

Results : LZ wins at H2 and H3, and resigns after only 52 moves at H4... which is inconclusive I think. It mostly shows that LZ doesn't handle high handicaps well.

command line for gtp4zen :
-z 6 -t 12 -T 10 -s 40000

command line for LZ13 :
--gtp -w 5d6d9.txt --noponder -r 10 -t 4 --gpu 0 --gpu 1
time_settings 0 10 1

-komi set at 0.5

-It seems that Zen6 doesn't ponder : high usage of CPU and no GPU usage during zen's time, followed by high GPU usage and not much CPU by LZ.

-I don't have Zen7, but it's probably at least 1 stone stronger than Zen6.

- Only thing missing... gtp4CSDL , because CSDL has a very nice and precise kyu/dan setting

The games : H2 ,H3,H4

Author:  kyotosato [ Fri Apr 13, 2018 3:02 am ]
Post subject:  Re: zen7 30 second : leela-zero 30 second, Zen7 won consiste

From my experience, Zen7 is about 2 stones or more stronger than Zen6. Two reasons: 1. end game has significantly improved, 2. opening has significantly improved. the opening recognizes mini-chinese--variation, and knows how to deal with it, zen6 doesn't. I used zen6 to play with 8dan in wildfox, 7 out of 10 lost because of opponents employing mini-chinese-variation.
30 second setting is a better setting, less than 30 seconds may have undershooting problem, and more than 30 seconds is ok, but if zen6 or zen7 uses 120 seconds may have overshooting problem. if you use the way i mentioned in my previous post to play Zen6 or Zen7 with opponents, you will notice that suggested spots/moves may change often after 30 seconds or more. zen6 7dan (or zen7 9dan) setting takes about 45s +-5s on my XP/pentium4 box.
the vision/GPU is a way to let AI has the ability to jump out of local maximum if needed. human players have that ability on behalf of vision, can easily have global view. opening of a game may benefit the most from vision. This is why i feel gpu matters.

Author:  Vargo [ Fri Apr 13, 2018 3:29 am ]
Post subject:  Re: zen7 30 second : leela-zero 30 second, Zen7 won consiste

That's interesting! You have zen6 and zen7, I would be very much interested in a zen6 v zen7 match, it should be easy with gtp4zen.
As for the time, 10 sec per move with 12-core, rapid access and a lot of ram is probably at least as much as 30" on an older computer. I'm not sure but I think I read somewhere that zen6 or 7 couldn't use more than 10 cores (?)
Anyway, I'm looking forward to the future versions of LZ, Zen, AQ, etc

Author:  kyotosato [ Fri Apr 13, 2018 3:45 am ]
Post subject:  Re: zen7 30 second : leela-zero 30 second, Zen7 won consiste

under same setting, same box, Zen6 vs. Zen7 i tried few times, Zen6 has no match (early in the game zen6 winrate drops steadily, then it is bloodshed.) zen6 always loss. My post is to see if any AI go accessible (by me for free or commercially available) could beat Zen7 or future zen. Current CrazyStone computer version is 7dan max, i have it and it is no match to zen7 (9dan max) or zen6 (7dan max.) i am looking forward CrazyStone new version commercial.
why i am so interested in Leela-zero, one is it is freely available, the other is because sooner or later LZ will surpass Zen7. they are two different animals. one is improving daily, and the other is fixed (discretely fixed). i have no objection of this possible outcome, and alphago master against alphago-zero matches have already proven that. the fact also proves that human heuristic experience is no match to AI heuristic experience. that means human guided search may give way to machine less-guided search. I do know why late Hawkins and Martin Heidegger both against AI development, however I do feel AI must continue but with caution.

Author:  Mike Novack [ Fri Apr 13, 2018 6:26 am ]
Post subject:  Re: zen7 30 second : leela-zero 30 second, Zen7 won consiste

tchan001 wrote:
How can it be wrong when he is only reporting the results of his tests on his own system. As he provides details of the configuration of the computer and of the settings he uses, you are responsible for making your own judgments on how useful the results are for yourself. Of course, you are free to offer results from tests done on your own system and settings which may differ significantly from what he achieved.


Well yes, BUT ..... these programs have minimum hardware requirements for adequate performance. Just saying "equal" does not actually cut it if the hardware requirements are different. Imagine a short race between a Tour de France competitor and a seven year old child on EQUAL bikes (both sized for that seven year old child)

Author:  Vargo [ Fri Apr 13, 2018 7:19 am ]
Post subject:  Re: zen7 30 second : leela-zero 30 second, Zen7 won consiste

I've just run another H3 game between LZ13 and Zen6 (Sabaki 0.33.4, win 10, 2x1080Ti, i9-12 core, no pondering, this time with 30" per move)

LZ wins after taking the 4 corners.

command line for gtp4zen :
-z 6 -t 12 -T 30 -s 80000

command line for LZ13 :
--gtp -w 5d6d9.txt --noponder -r 10 -t 4 --gpu 0 --gpu 1
time_settings 0 30 1


THE GAME

I am almost to the point where I'll buy Zen7 just to test the difference in strength between V6 and V7 ;-)

Author:  Javaness2 [ Fri Apr 13, 2018 7:51 am ]
Post subject:  Re: zen7 30 second : leela-zero 30 second, Zen7 won consiste

black passed mid game...

Author:  Vargo [ Fri Apr 13, 2018 8:09 am ]
Post subject:  Re: zen7 30 second : leela-zero 30 second, Zen7 won consiste

Yes, at move 114... that's weird... :o :scratch:

Author:  kyotosato [ Fri Apr 13, 2018 2:53 pm ]
Post subject:  Re: zen7 30 second : leela-zero 30 second, Zen7 won consiste

hydrogenpi7 wrote:
...Then you use a weak CPU with no GPU... which come on, is not a fair test... everybody has at least a gtx 1080ti these days. Any test or benchmark that doesn't include a >= gtx 1080 isn't a real world test imho

My very first post is not a test on a pentium box.
--- Acer One 10 ; cpu: Intel® Atom™ x5-Z8350 (32 bit, NOT 64 bit) 1.4 GHz; Quad-core; OS: windows 10 ---
This box does have a gpu card GPU: HD Graphics (Cherry Trail); and core: BLAS Nehalem. So the run is not so bias toward zen7. however, the box is an inferior one.

Author:  kyotosato [ Fri Apr 13, 2018 2:57 pm ]
Post subject:  Re: zen7 30 second : leela-zero 30 second, Zen7 won consiste

Javaness2 wrote:
black passed mid game...

in sabaki, after resigned, the last move is recorded as "passed" i noticed.

Author:  kyotosato [ Fri Apr 13, 2018 3:16 pm ]
Post subject:  Re: zen7 30 second : leela-zero 30 second, Zen7 won consiste

Mike Novack wrote:
Imagine a short race between a Tour de France competitor and a seven year old child on EQUAL bikes (both sized for that seven year old child)

That is true. there was once i played at temgen website with an opponent who employed zen7 with cloud computing CPU against me with zen7 on my pentium box, my opponent won gradually. and took less computing time, more interesting is that some of my opponent's moves were not or with low winrate on my hotspots lists. not on my hotspots list bothers me more. better CPU means fast iteration less time consumed for zen series.
Vargo wrote:
As for the time, 10 sec per move with 12-core, rapid access and a lot of ram is probably at least as much as 30" on an older computer.

I agree

Author:  kyotosato [ Fri Apr 13, 2018 3:41 pm ]
Post subject:  Re: zen7 30 second : leela-zero 30 second, Zen7 won consiste

Vargo wrote:
command line for gtp4zen :
-z 6 -t 12 -T 30 -s 80000
command line for LZ13 :
--gtp -w 5d6d9.txt --noponder -r 10 -t 4 --gpu 0 --gpu 1
time_settings 0 30 1

the author of gtp4zen suggested higher -s and -T value will increase zen's strength. I used -s 200000 on my prior test run, you may try higher -s setting, see if the match outcome changes.

Author:  Vargo [ Fri Apr 13, 2018 10:49 pm ]
Post subject:  Re: zen7 30 second : leela-zero 30 second, Zen7 won consiste

New H3 game between LZ13 and gtp4zen(v6) with two differences :
-new network c0cb60.txt for LZ13
-parameter -s 300000 for gtp4zen

no weird passing this time, but same outcome...
Very nice game, I think I'll run more of these games, but with 10" per move, to see if LZ wins at H3 consistently or not.
2-3 games is not enough to have an opinion.

I'm more and more tempted to buy z7 to test the difference ;-)

THE GAME

Author:  as0770 [ Fri Apr 13, 2018 11:02 pm ]
Post subject:  Re: zen7 30 second : leela-zero 30 second, Zen7 won consiste

kyotosato wrote:
Vargo wrote:
command line for gtp4zen :
-z 6 -t 12 -T 30 -s 80000
command line for LZ13 :
--gtp -w 5d6d9.txt --noponder -r 10 -t 4 --gpu 0 --gpu 1
time_settings 0 30 1

the author of gtp4zen suggested higher -s and -T value will increase zen's strength. I used -s 200000 on my prior test run, you may try higher -s setting, see if the match outcome changes.


What exactly does the -s value, and how high can it be set?

Author:  Vargo [ Sat Apr 14, 2018 12:19 am ]
Post subject:  Re: zen7 30 second : leela-zero 30 second, Zen7 won consiste

Quote:
What exactly does the -s value, and how high can it be set?
s is for strength (default =10000), it must be the number of playouts, or visits or such

I made a simple test with gtp4zen (v6):
-T 1 -s 300000 (time=1s per move, 300000 playouts) z6 plays at 1sec. per move
-T 100 -s 1000 z6 plays instantly, although T=100sec per move

So, I think when you set T=a and S=b, z6 plays as soon as it reaches a or b, whichever comes first.

About the number of cores supported by Zen, in another thread, pookpooi said that Zen7 supports 10 cores, and zen6 only 8.

Author:  pnprog [ Sat Apr 14, 2018 1:57 am ]
Post subject:  Re: zen7 30 second : leela-zero 30 second, Zen7 won consiste

Hi!
kyotosato wrote:
pnprog:
Go GridMaster is just what I need for my android devices. thanks to your creation, a decent and nice product. i just wonder why you didn't continue to go further to AI 19x19?

What do you mean exactly?

Author:  pnprog [ Sat Apr 14, 2018 2:50 am ]
Post subject:  Re: zen7 30 second : leela-zero 30 second, Zen7 won consiste

hydrogenpi7 wrote:
everybody has at least a gtx 1080ti these days. Any test or benchmark that doesn't include a >= gtx 1080 isn't a real world test imho

:roll: :roll: :roll: :roll: :roll: :roll:

Well, my most powerful computer is equipped with an Nvidia integrated GPU, the GeForce GT 720M which was probably already a joke at the time it was released.

So here is a comparison benchmark: Nvidia GTX 1080-Ti vs GeForce GT 720M, the Nvidia-GTX-1080-Ti is apparently 3000% faster than my GPU :mrgreen:

Now, I think a good share of Go players use old computers or laptops, at least not gamers computers. The average age of the Go player population (in the east) might be on of the factor. Also, running a low computer is an excellent incentive to play Go, as this game does not require a gamer computer.

So for people like me, comparative tests on low end computers are quite interesting.

Author:  kyotosato [ Sat Apr 14, 2018 3:18 am ]
Post subject:  Re: zen7 30 second : leela-zero 30 second, Zen7 won consiste

pnprog wrote:
Hi!
What do you mean exactly?

my mistake, i thought you're Go GridMaster's author.
pnprog wrote:
Now, I think a good share of Go players use old computers or laptops, at least not gamers computers. The average age of the Go player population (in the east) might be on of the factor. Also, running a low computer is an excellent incentive to play Go, as this game does not require a gamer computer.

exactly! most members of my small go club/circle do not interested in computer or don't have one, therefore an android AQ or android Crazystone could be very valuable to them. age is a factor.

but why do you think it is an incentive by employing a low end computer (not require a gamer computer is for sure, but a high end computer certainly would help.)
pnprog wrote:
So for people like me, comparative tests on low end computers are quite interesting.

What is the point? as a game reviewer's point of view.

Page 2 of 3 All times are UTC - 8 hours [ DST ]
Powered by phpBB © 2000, 2002, 2005, 2007 phpBB Group
http://www.phpbb.com/