Discussion of computer chess matches and engine tournaments.
Moderator: Ras
Werner
Posts: 3017Joined: Wed Mar 08, 2006 10:09 pmLocation: GermanyFull name: Werner Schüle
Post by Werner Tue Mar 03, 2009 5:51 pm
I started to test some "fruits" for CEGT 40/20 (eng-eng, ponder off).
After round 10 we have following interim results:
Code: Select all
Naum 3 w32 1CPU - Toga II 1.4.3JDbeta19a 1CPU 5.5 - 4.5 +3/-2/=5 55.00%
Naum 3 w32 1CPU - Grapefruit 1.0 1CPU 7.5 - 2.5 +6/-1/=3 75.00%
Naum 3 w32 1CPU - Fruit 2.3.5m p15 w32 1CPU 4.0 - 6.0 +3/-5/=2 40.00%
Naum 3 w32 1CPU - Cyclone 3.4 1CPU 6.0 - 4.0 +4/-2/=4 60.00%Code: Select all
Rybka 1.1.w32 2200 - TogaII 1.4.3JD b19a 1CPU 2200 5.5 - 4.5 +3/-2/=5 55.00%
Rybka 1.1.w32 2200 - Grapefruit 1.0 1CPU 2200 4.0 - 6.0 +2/-4/=4 40.00%
Rybka 1.1.w32 2200 - Fruit 2.3.5m p15 w32 1CPU xxxx 4.5 - 5.5 +2/-3/=5 45.00%
Rybka 1.1.w32 2200 - Cyclone 3.4 1CPU 2200 5.5 - 4.5 +3/-2/=5 55.00%Werner
Werner
Posts: 3017Joined: Wed Mar 08, 2006 10:09 pmLocation: GermanyFull name: Werner Schüle
Post by Werner Wed Mar 04, 2009 1:48 pm
I started to test some "fruits" for CEGT 40/20 (eng-eng, ponder off).
After round 20 we have following interim results:
Code: Select all
Naum 3 w32 1CPU - Toga II 1.4.3JDbeta19a 1CPU 10.5 - 9.5 52.50%
Naum 3 w32 1CPU - Grapefruit 1.0 1CPU 12.5 - 7.5 62.50%
Naum 3 w32 1CPU - Fruit 2.3.5m p15 w32 1CPU 9.5 - 10.5 47.50%
Naum 3 w32 1CPU - Cyclone 3.4 1CPU 11.5 - 8.5 57.050%Code: Select all
Rybka 1.1.w32 2200 - TogaII 1.4.3JD b19a 1CPU 2200 9.5 - 10.5 47.50%
Rybka 1.1.w32 2200 - Grapefruit 1.0 1CPU 2200 9.0 - 11.0 45.00%
Rybka 1.1.w32 2200 - Fruit 2.3.5m p15 w32 1CPU xxxx 10.5 - 9.5 52.50%
Rybka 1.1.w32 2200 - Cyclone 3.4 1CPU 2200 11.0 - 9.0 55.00%Werner
Denis P. Mendoza
Posts: 415Joined: Fri Dec 15, 2006 9:46 pmLocation: Philippines
Post by Denis P. Mendoza Wed Mar 04, 2009 8:26 pm
Werner wrote: I started to test some "fruits" for CEGT 40/20 (eng-eng, ponder off).
After round 20 we have following interim results:
Code: Select all
Naum 3 w32 1CPU - Toga II 1.4.3JDbeta19a 1CPU 10.5 - 9.5 52.50%
Naum 3 w32 1CPU - Grapefruit 1.0 1CPU 12.5 - 7.5 62.50%
Naum 3 w32 1CPU - Fruit 2.3.5m p15 w32 1CPU 9.5 - 10.5 47.50%
Naum 3 w32 1CPU - Cyclone 3.4 1CPU 11.5 - 8.5 57.050%Code: Select all
Rybka 1.1.w32 2200 - TogaII 1.4.3JD b19a 1CPU 2200 9.5 - 10.5 47.50%
Rybka 1.1.w32 2200 - Grapefruit 1.0 1CPU 2200 9.0 - 11.0 45.00%
Rybka 1.1.w32 2200 - Fruit 2.3.5m p15 w32 1CPU xxxx 10.5 - 9.5 52.50%
Rybka 1.1.w32 2200 - Cyclone 3.4 1CPU 2200 11.0 - 9.0 55.00%
Thanks Werner. I see Jerry's new Toga is doing well.
But we have an issue when using 'beta19a' on quad-core. Shaun mentioned that it's using only around 25% cpu as quoted here:
it does not seem to use 4 threads (25% CPU on a quad) and does not play nicely with F11 GUI e.g. load errors, also if you use for analysis it seems to ignore the stop as the engine continues to think?
If you mind, could you verify this just get some third party opinion?
Thanks.
Denis
PS> I also made some small gauntlet for comparison:
Code: Select all
Toga143jdbeta19a
Rank Engine Score To
1 TogaII143b19a 51.5/100 · ·· ·· ·· ·· ·· ·· ·· ·· ·· ·· ·· ·· ·· ·· ·· ·· ·· ·· ·· ·· ·
2 SP-x64-Inert---Thinker54A 27.0/50 0=11010==1==0=01=1=11==1=00=110100=0=0=101=101=1=1
3 Glaurung22_x64_ja 21.5/50 ==0==00000=0=1=00=1=011=00==110=0110110101==100==0
100 games played / Tournament is finished
Tournament start: 2009.02.25, 06:09:13
Latest update: 2009.02.26, 21:59:40
Site/ Country: SUPERCHESS, United States
Level: Tournament 40/5
Hardware: Dual AMD Athlon(tm) 64 X2 Dual Core Processor 3800+ 2009 MHz with 1,920 MB Memory
Operating system: Microsoft Windows XP 64 Bit Professional Service Pack 2 (Build 3790)
PGN-File: Toga143jdbeta19a.pgn
Table created with: Arena 1.99beta4
Werner
Posts: 3017Joined: Wed Mar 08, 2006 10:09 pmLocation: GermanyFull name: Werner Schüle
Post by Werner Thu Mar 05, 2009 8:58 am
"If you mind, could you verify this just get some third party opinion? "
Yes I can confirm this - and the engine crashes using more than 1CPU
Werner
Werner
Posts: 3017Joined: Wed Mar 08, 2006 10:09 pmLocation: GermanyFull name: Werner Schüle
Post by Werner Thu Mar 05, 2009 11:09 am
Werner wrote: "If you mind, could you verify this just get some third party opinion? "
Yes I can confirm this - and the engine crashes using more than 1CPU
both compiles:
Toga1.4.3JDbeta19a_C2D.exe from 21. feb and
TogaII143b19a.exe from Jerry (05. febr)
Werner
Werner
Posts: 3017Joined: Wed Mar 08, 2006 10:09 pmLocation: GermanyFull name: Werner Schüle
Post by Werner Thu Mar 05, 2009 5:41 pm
Code: Select all
Naum 3 w32 1CPU - Toga II 1.4.3JDbeta19a 1CPU 22 - 18 55%
Naum 3 w32 1CPU - Grapefruit 1.0 1CPU 24 - 16 60%
Naum 3 w32 1CPU - Fruit 2.3.5m p15 w32 1CPU 19.5 - 20.5 48,75%
Naum 3 w32 1CPU - Cyclone 3.4 1CPU 24 -16 60% Code: Select all
Rybka 1.1.w32 2200 - TogaII 1.4.3JD b19a 1CPU 2200 18 - 22 45%
Rybka 1.1.w32 2200 - Grapefruit 1.0 1CPU 2200 23 - 17 57,5%
Rybka 1.1.w32 2200 - Fruit 2.3.5m p15 w32 1CPU xxxx 24,5 - 15.5 61,25%
Rybka 1.1.w32 2200 - Cyclone 3.4 1CPU 2200 21,5 - 18,5 53,75%Werner
Jouni
Posts: 3810Joined: Wed Mar 08, 2006 8:15 pmFull name: Jouni Uski
Post by Jouni Thu Mar 05, 2009 6:17 pm
So combined:
Denis P. Mendoza
Posts: 415Joined: Fri Dec 15, 2006 9:46 pmLocation: Philippines
Post by Denis P. Mendoza Fri Mar 06, 2009 4:18 am
Jouni wrote: So combined:
http://www.computerchess.info/tdbb/phpBB3/index.php
You'll see some unofficial releases there. But all official releases are most likely be posted at CCRL or TDDB public forum. The latest development engines are still private. Jerry wants to release 'beta19a', but because of the mentioned problems >1 cpu it will be delayed.
Werner
Posts: 3017Joined: Wed Mar 08, 2006 10:09 pmLocation: GermanyFull name: Werner Schüle
Post by Werner Fri Mar 06, 2009 12:54 pm
2 Matches finished:
Code: Select all
Rybka 1.1 w32 2200 - Cyclone 3.4 1CPU 2200 26.5 - 23.5 +12/-9/=29 53.00%
Rybka 1.1 w32 2200 - Fruit 2.3.5m p15 w32 1CPU 2200 28.5 - 21.5 +16/-9/=25 57.00%
Rybka 1.1 w32 2200 - Grapefruit 1.0 1CPU 2200 30.5 - 19.5 +19/-8/=23 61.00%
Rybka 1.1 w32 2200 - TogaII 1.4.3JD b19a 1CPU 2200 23.0 - 27.0 +13/-17/=20 46.00%Code: Select all
Naum 3.0 w32 1CPU - TogaII 1.4.3JD b19a 1CPU 27.5 - 22.5 +17/-12/=21 55.00%
Naum 3.0 w32 1CPU - Grapefruit 1.0 1CPU 29.0 - 21.0 +20/-12/=18 58.00%
Naum 3.0 w32 1CPU - Fruit 2.3.5m p15 w32 1CPU 23.5 - 26.5 +13/-16/=21 47.00%
Naum 3.0 w32 1CPU - Cyclone 3.4 1CPU 29.5 - 20.5 +17/-8/=25 59.00%
Werner
now running Zappa Mexico II x64 1CPU and Deep Sjeng WC 2008 w32 1CPU against these 4 engines.
Werner
Werner
Posts: 3017Joined: Wed Mar 08, 2006 10:09 pmLocation: GermanyFull name: Werner Schüle
Post by Werner Tue Mar 10, 2009 9:32 am
Code: Select all
DeepSjeng WC2008 w32 1CPU - Cyclone 3.4 1CPU 2200 25.5 - 24.5 +20/-19/=11 51.00%
DeepSjeng WC2008 w32 1CPU - Fruit 2.3.5m p15 w32 1CPU 2200 25.5 - 24.5 +18/-17/=15 51.00%
DeepSjeng WC2008 w32 1CPU - Grapefruit 1.0 1CPU 2200 22.5 - 27.5 +12/-17/=21 45.00%
DeepSjeng WC2008 w32 1CPU - Toga II 1.4.3JDbeta19a1C 2200 27.5 - 22.5 +18/-13/=19 55.00%Code: Select all
Zappa Mexico II x64 1CPU - Cyclone 3.4 1CPU 26.0 - 24.0 +12/-10/=28 52.00%
Zappa Mexico II x64 1CPU - Fruit 2.3.5m p15 w32 1CPU 25.5 - 24.5 +12/-11/=27 51.00%
Zappa Mexico II x64 1CPU - Grapefruit 1.0 1CPU 30.5 - 19.5 +20/-9/=21 61.00%
Zappa Mexico II x64 1CPU - Toga II 1.4.3JDbeta19a1C 28.5 - 21.5 +20/-13/=17 57.00%Werner
