the digital rebirth of Deep Sjeng

Discussion of computer chess matches and engine tournaments.

Moderators: bob, hgm, Harvey Williamson

Forum rules
This textbox is used to restore diagrams posted with the [d] tag before the upgrade.
User avatar
Guenther
Posts: 3054
Joined: Wed Oct 01, 2008 4:33 am
Location: Regensburg, Germany
Full name: Guenther Simon
Contact:

Re: the digital rebirth of Deep Sjeng

Post by Guenther » Fri May 03, 2019 2:07 pm

ThatsIt wrote:
Fri May 03, 2019 2:01 pm
Hi Günther,

i've made an update of the CEGT-Tester-Page:

http://www.cegt.net/testers/testers.html

Some details are still missing but will be added soon.

Best wishes,
G.S.
(CEGT team)
Thanks Gerhard!
(BTW the links page is also very very outdated, but it's not too important.)
Current foe list count : [99]
http://rwbc-chess.de/chronology.htm

ThatsIt
Posts: 924
Joined: Thu Mar 09, 2006 1:11 pm
Contact:

Re: the digital rebirth of Deep Sjeng

Post by ThatsIt » Fri May 03, 2019 2:27 pm

Guenther wrote:
Fri May 03, 2019 2:07 pm
(BTW the links page is also very very outdated, but it's not too important.)
Hi Günther,

That's right
I think i will delete the link-page ;-)

Best wishes,
G.S.
(CEGT team)

User avatar
Werner
Posts: 2419
Joined: Wed Mar 08, 2006 9:09 pm

Re: the digital rebirth of Deep Sjeng

Post by Werner » Sun May 05, 2019 9:19 pm

Werner wrote:
Fri May 03, 2019 6:20 am
Result for 40/20:

1 StoofVlees II a9 +143 +54/=31/-15 69.50% 69.5/100
2 Ethereal 11.25 x64 1CPU -143 +15/=31/-54 30.50% 30.5/100
now against Komodo 12.2.2 x64 1CPU

1 StoofVlees II a9 2200 +3 +25/=51/-24 50.50% 50.5/100
2 Komodo 12.2.2 x64 1CPU 2200 -3 +24/=51/-25 49.50% 49.5/100

now against Deep Shredder 13 x64 1CPU
Werner

User avatar
Werner
Posts: 2419
Joined: Wed Mar 08, 2006 9:09 pm

Re: the digital rebirth of Deep Sjeng

Post by Werner » Wed May 08, 2019 3:24 pm

Result against DS13 not so good:

1 StoofVlees II a9 +111 +50/=31/-19 65.50% 65.5/100
2 Deep Shredder 13 x64 1CPU -111 +19/=31/-50 34.50% 34.5/100
Werner

chessfurby
Posts: 49
Joined: Thu Apr 26, 2007 11:41 am
Location: Germany, Bavaria

Re: the digital rebirth of Deep Sjeng

Post by chessfurby » Thu Jun 27, 2019 7:42 am

Stoofvlees has quite successfully made it through the qualification stage in current CCC 9.

It has progressed into the quarterfinals. It might be the end of the story for now, as it had
a bad start in the quarters but moving up the ranks slowly.... but unlikely to get into the top 4.

Still not so bad for an all new enghine which is not related to the existing NN engine LC0 (well aside
from the fact that GCP did write Leela go in the first place of course).

supersharp77
Posts: 754
Joined: Sat Jul 05, 2014 5:54 am
Location: Southwest USA

Re:Stoofvlees IIa12-digital rebirth of Deep Sjeng

Post by supersharp77 » Sun Nov 17, 2019 4:30 am

Werner wrote:
Wed May 01, 2019 10:00 am
On May 2018 I asked Gian Carlo: is it possible to write a Deep Sjeng running on a GPU?
Now I got the answer: done!
"It is a new design, not related to either program. I mostly wanted to try a bunch of ideas that never made it to Leela (Zero Chess)"
He called it Stoofvlees. Maybe when it is ready it will be Deep Sjeng 4.0 :-)
The engine has an own net and only works on a Nvidia GPU. It also runs on CPU - but about 300 points weaker.

The newest version is named StoofVlees II a9 and we started to test it for CEGT 40/20 and 40/4.
At the moment I have 30 games for 40/20
StoofVlees II a9 - Ethereal 11.25 x64 1CPU : 20,5/30 16-5-9

and I made 100 games for 40/4

1 StoofVlees II a9 +10 +38/=27/-35 51.50% 51.5/100
2 Xiphos 0.5 x64 1CPU -10 +35/=27/-38 48.50% 48.5/100
24.09.19
TCEC Superfinal Stockfish-AllieStein

TCEC sezóna 16, Prémiová divize, skončila velkým triumfem Stockfishe, který přes nešťastnou ztrátu dvou bodů turnaj vyhrál
před moderními NN motory.

1 Stockfish 190826 26.5
2 AllieStein v0.5 26.0
3 LCZero v0.22.0 25.5
4 Komodo 2381.00 22.0
5 Stoofvlees IIa12 19.0
6 Houdini 6.03 18.5
7 KomodoMCTS 2381 16.5
8 ScorpioNN v3.0.1 14.0

Congratulations to Stoofvlees IIa12......(Deep Sjeng v4.0?) :D :wink:

Post Reply