Test no.1 (DiscoCheck w32 comparative)

[Sylwy]

S R

[Sylwy]

Conditions :

-TC = 4'+2"
-Hash = 128 MB
-Book = Shredder12.ctg
-GUI = Fritz 11
-CPU = Athlon 64 3800+

Scorpio 2.7.6 JA was linked to Scorpio egbbs.
Crafty 24.1 was linked to Nalimov TBs.



The best Crafty on 32-bit remains the version 23.8:




S R

[carldaman]

Conditions :

-TC = 4'+2"
-Hash = 128 MB
-Book = Shredder12.ctg
-GUI = Fritz 11
-CPU = Athlon 64 3800+

Scorpio 2.7.6 JA was linked to Scorpio egbbs.
Crafty 24.1 was linked to Nalimov TBs.



The best Crafty on 32-bit remains the version 23.8:




S R

Hi there Ruxy,

Yes, it's odd how Crafty has regressed since version 23.8. The newer versions may have introduced a lot of changes and maybe a bug has crept in, too. It could also be a slower compile.

Regards,
CL

[Sylwy]

...........IPON, TCEC & Co. GmbH are extremely interested (from years) to waste their CPU time testing various versions of the same 2 cents engine the old Mr.Ruxy do a very important & serious tournament: he tests the best chess engine programmed in Python: Sunfish_28_09_14 (the newest version of this engine). Never tested by the lords of absurd !!!!

Conditions:

-TC = 5'/40 moves (with repetition) - because of Tiffanys 0.5
-Book = Perfect_2010.abk; Tiffanys 0.5 uses its own incorporated book
-GUI = aReNa 3.0 Hey Miky, how are you ???
-CPU = Athlon 64 3800+



Well, until the tournament will be ready, everybody in galaxy can think to a rating of this super-Sunfish, because:

on CCRL 4/40:

-Talvmenni 0.1.........1066 Elo points
-AdaChess 1.0..........1061
-Tiffanys 0.5.............1026 (after a single game.....so lazzy........!!!!!)
-Minnow....................1023
-Zoe 0.1...................1022

Have the night of your life !!!

S R

[Sylwy]

Hi there Ruxy,

Yes, it's odd how Crafty has regressed since version 23.8. The newer versions may have introduced a lot of changes and maybe a bug has crept in, too. It could also be a slower compile.

Regards,
CL

Hey Carl,

Take a look in CCRL 4/40 (close to my TC = 4'+2"):

http://www.computerchess.org.uk/ccrl/40 ... t_all.html

-Crafty 23.8 x64......................2824 Elo points
-Crafty 24.0 x64......................2800

They used the binaries coming from Mr.Peter Skinner.
The binaries on 32-bit coming from Mr.Daniel José are OK. The benchmark is normal compared to the old compiles of Mr.Jim Ablett !

S R

[jdart]

My understanding is 24.0 mainly had changes to improve speed on large multi-processor systems. So you mayl not see improvements running it on one core, and may even see a regression.

--Jon

[Sylwy]

My understanding is 24.0 mainly had changes to improve speed on large multi-processor systems. So you mayl not see improvements running it on one core, and may even see a regression.

--Jon

Yes, of course ! You are right !

S S

[Sylwy]

Like bonus: IASI in October:



Regards,
S R

[Sylwy]

S R

[Sylwy]

Well, I understand : no CPU time for Sunfish_28_09_2014. All the resources in crisis are dedied for testing various derivatives/clones of this oriental Ippo !

So, here are my final results:



Project Invincible 2.04 has 1189 Elo points (CCRL 4/40).
Talvmenni 01 has 1066 Elo points.

===============
Sunfish_28_09_2014 has around 1130 Elo points !
===============

Also a bit of help for Tiffanys 0.1 (it played only 1 game in CCRL 4/40 - I suspect they don't know how to run this very particular chess engine, or is a bit of lazzines ?).

===============
Tiffanys 0.5 has a bit over 1200 Elo points !
===============

Well, being Saturday two bonus:

1.- a superb victory of Sunfish_28_09_2014 versus the leader (Tiffanys 0.5):

[pgn]
[Event "Sunfish_28_09_2014 Test"]
[Site "SILVIAN"]
[Date "2014.10.11"]
[Round "20"]
[White "Sunfish_28_09_2014"]
[Black "Tiffanys 0.5"]
[Result "1-0"]
[BlackElo "2200"]
[ECO "C45"]
[Opening "Scotch"]
[Time "20:06:44"]
[Variation "4.Nxd4 Nf6 5.Nxc6 bxc6 6.Bd3"]
[WhiteElo "2200"]
[TimeControl "40/300:40/300:40/300"]
[Termination "normal"]
[PlyCount "101"]
[WhiteType "program"]
[BlackType "program"]

1. e4 g6 2. Nc3 e5 3. Nf3 Bd6 4. d4 exd4 5. Qxd4 Nf6 6. Nd5 Be7 7. Bg5 Nc6
8. Nxf6+ Bxf6 9. Qxf6 Qxf6 10. Bxf6 O-O 11. Bd3 d6 12. c3 Rb8 13. O-O h5
14. Nd4 Bd7 15. Nxc6 bxc6 16. c4 Rfc8 17. f4 Rb6 18. Rae1 Rcb8 19. Bc3 R6b7
20. Re3 Be6 21. Rfe1 Rf8 22. a3 Rc8 23. a4 a6 24. a5 Rf8 25. b4 Ra7 26. h3
f5 27. Rg3 Kh7 28. exf5 Bxf5 29. Re7+ Kh6 30. Bg7+ Kh7 31. Bd4+ Kh6 32.
Bxf5 Rxf5 33. Bxa7 d5 34. Rxc7 Rf6 35. Be3 Rf5 36. Rxc6 Kh7 37. Rgxg6 Rf7
38. cxd5 Re7 39. Rge6 Rd7 40. Rh6+ Kg8 41. Rxh5 Re7 42. Re5 Rb7 43. Bc5 Rf7
44. Rg5+ Kh7 45. g3 Rg7 46. Rxg7+ Kxg7 47. d6 Kg8 48. d7 Kh8 49. d8=Q+ Kh7
50. Qd7+ Kg8 51. Rc8# 1-0
[/pgn]

A beautiful victory of this superb Python chess engine !!!

2.-an image with Dobrovat forest (imediately near IASI) two days ago:




Have a nice weekend !
S R

The nest week - two tests:
-the best chess engine from IASI (Freyr) versus the best engine from MONTENEGRO (Zeta Dva);
-Arasan 17.4 test !