IvanHoe T52 testing

[kingliveson]

Code: Select all

Rybka vs IvanHoe, Blitz:15'+10  2010                      

1   Rybka 4 w32        +24/=54/-22 51.00%   51.0/100
2   IvanHoe 9.52b w32  +22/=54/-24 49.00%   49.0/100

† 50 Positions (test suite)
† 3/4 Nalimov/RobboBase
† 64 MB Hash
† Ponder Off
† AMD Turion(tm) 64 Mobile Technology ML-37
† 1 CPU
† 15+10 Incrementing Time Control

IvanHoe
PawnsHash 8 MB
TimeImitateOpponent
TimeMoreWhenLosing
TimeMoreWhenWinning

Code: Select all

Ivanhoe v Rybka R0001, 4'/40+4'/40+4'/4  2010            

1   Deep Rybka 4 x64   +76/=160/-64 52.00%  156.0/300
2   IvanHoe 9.52b x64  +64/=160/-76 48.00%  144.0/300

† 150 Positions (test suite)
† 3-4-5 Nalimov/RobboBase
† 128 MB Hash
† Ponder Off
† Athlon 64 X2 6000+ @ 3100 MHz
† 2 Cores
† 40/4 Repeating Time Control

IvanHoe
PawnsHash 16 MB


Game data available: http://chess.cygnitec.com/pgn/

[robbolito]

AMD Phenom(tm) II X6 1090T Processor6x @ 3.2 GHz 4,096 MB Memory
Windows 7 Home Premium Edition (Build 7600)
Fritz Benchmark:
Speed: 23.33
KNS: 11196
GUI: CB Rybka 3
Book: HS 10 moves
Hash: 256
RB and TB: ON
Ponder: OFF

Blitz:10' 0

Code: Select all

                                   
1   Ivanhoe 52cU-x64         3190  +4/-3/=13 52.50%   10.5/20
2   Houdini 1.03a x64 8_CPU  3190  +3/-4/=13 47.50%    9.5/20

[DaveD]

AMD Phenom(tm) II X6 1090T Processor6x @ 3.2 GHz 4,096 MB Memory
Windows 7 Home Premium Edition (Build 7600)
Fritz Benchmark:
Speed: 23.33
KNS: 11196
GUI: CB Rybka 3
Book: HS 10 moves
Hash: 256
RB and TB: ON
Ponder: OFF

Blitz:10' 0

Code: Select all

                                   
1   Ivanhoe 52cU-x64         3190  +4/-3/=13 52.50%   10.5/20
2   Houdini 1.03a x64 8_CPU  3190  +3/-4/=13 47.50%    9.5/20

Ivanhoe 52cU is not a familiar version to me. Where can it be downloaded?

[kingliveson]

The head-to-head match up tells the whole story.

Third World War  2010
                                          

1 Houdini 1.03a x64 POPCNT 4_CPU  ***********  22.5 - 27.5  24.0 - 26.0  31.0 - 19.0  77.5/150
2 Deep Rybka 4 SSE42 x64          27.5 - 22.5  ***********  22.0 - 28.0  27.0 - 23.0  76.5/150
3 IvanHoe 9.52b x64               26.0 - 24.0  28.0 - 22.0  ***********  20.0 - 30.0  74.0/150
4 FireBird 1.1 x64                19.0 - 31.0  23.0 - 27.0  30.0 - 20.0  ***********  72.0/150

† 25 Positions (test suite)
† EGTB Off
† Ponder Off
† 128 MB Hash
† Phenom II 940
† 4 Cores
† 2+2 Incrementing Time Control

FireBird
Pawn Hash 64 MB*
Pawn Value 120
Rook Value 600
Queen Value 1170
Bishop Pair Value 70
Verification Reduction 8

IvanHoe
Pawn Hash 16 MB

Code: Select all

Rybka vs IvanHoe, 10'/40+10'/40+10'/40  2010
                  
1   IvanHoe 9.52b w32  +12/=27/-11 51.00%   25.5/50
2   Rybka 4 w32        +11/=27/-12 49.00%   24.5/50

† 25 Positions (test suite)
† 3/4 Nalimov/RobbBase
† 64 MB Hash
† Ponder Off
† AMD Turion(tm) 64 Mobile Technology ML-37
† 1 CPU
† 40/10 Repeating Time Control

IvanHoe
PawnsHash 8 MB


Game data available: http://chess.cygnitec.com/pgn/

Note: IvanHoe used new time management settings that prevents it from losing on time or getting in time trouble.
* FireBird Pawn Hash is way too huge for such time control.

[robbolito]

Intel(R) Core(TM) i7 4CPU 960 3.20GHz8x @ 4.005 GHz with 4,096 MB Memory
Microsoft Windows XP 64 Bit Professional Service Pack 2 (Build 3790)
Fritz Benchmark:
Speed: 24.98
KNS: 11988
GUI: CB Rybka 3
Book: Perfect 2009-10 moves
Hash: 256
RB and TB: ON
Ponder: OFF


Blitz:10' 0

Code: Select all

1   IvanHoe T52E.x64POPCNT         3190  +10/-2/=8 70.00%   14.0/20
2   Houdini 1.03a x64 POPCNT 4_CP  3190  +2/-10/=8 30.00%    6.0/20

Games also included.

[robbolito]

Intel(R) Core(TM)2 Quad Q9550 2.83GHz4x @3.5 GHz 4,096 MB Memory
Microsoft Windows XP 64 Bit Professional Service Pack 2 (Build 3790)
Fritz Benchmark:
Speed: 20.90
KNS: 10032
GUI: CB Rybka 3
Book: Perfect 2009-10 moves
Hash: 256
RB and TB: ON
Ponder: OFF


Blitz:10' 0

Code: Select all

1   IvanHoe T52E.x64         3190  +10/-6/=34 54.00%   27.0/50
2   Houdini 1.03a x64 4_CPU  3190  +6/-10/=34 46.00%   23.0/50

[Gerold]

Why not run 200+ matches. With these top engine matches 20-50 games mean little as for as which is the strongest.

Best,
Gerold.

[robbolito]

Gerold,
easier said then done.
To play 200 games at 10 min per engine in theory would take about 66 hours but practically we are talking more than day and a half.
And with hope that there are no exceptions to stop the games in the middle of the night.
So ,as you see it is not easy to play 200 games but going slowly one can reach to decent number of games to make relative evaluation of the engine strength.
Any volunteers?

[robbolito]

AMD Phenom(tm) II X6 1090T Processor6x @ 3.2 GHz 4,096 MB Memory
Windows 7 Home Premium Edition (Build 7600)
Fritz Benchmark:
Speed: 23.33
KNS: 11196
GUI: CB Rybka 3
Book: Perfect 2009- 10 moves
Hash: 256
RB and TB: ON
Ponder: OFF

Blitz:10' 0

Code: Select all

1   Ivanhoe 52dU-x64         3190  +9/-4/=31 55.68%   24.5/44
2   Houdini 1.03a x64 8_CPU  3190  +4/-9/=31 44.32%   19.5/44

[alexandros]

On my computer, AMD X64 double core, Win XP32 SP3 , the best IvanHoe was T52D. I matched it against T52E at 400 msec + 40 msec time control, 16,034 games (I got bored testing more)



They are virtually equal within +/-4 Elo points interval with 95% confidence. On my comp the order is the following:

Houdini 1.03a > IvanHoe T52D/E > Rybka 4 > Fire 1.31 with >99% confidence, after several long tests at very short time controls. Some of the other IvanHoes are somewhere in between T52D/E and Rybka 4.