Page 7 of 16
Re: IvanHoe compiles testing
Posted: Tue Nov 23, 2010 1:08 pm
by ernest
ernest wrote:Do you agree, and which version did you use?
OK, I found it:
Latest compile from KLO Date 2010.11.20
Quote:
IvanHoe 999950t Beta Analysis/Game Mode combined in one plus Intel and Microsoft. Windows only -- Linux (Intel and GCC) build available upon request.
Which version are you testing: Intel, Microsoft? Default, popcnt?
Is there something to be checked in the "Engine parameters" in order to choose Analysis or Game Mode?
Re: IvanHoe compiles testing
Posted: Tue Nov 23, 2010 2:35 pm
by 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: Kevin’s Huge Ctg-10 moves
Hash: 256
RB and TB: ON
Ponder: OFF
Blitz:10' 0
Code: Select all
1 IvanHoe T49f.x64 3190 +14/-12/=24 52.00% 26.0/50
2 Houdini 1.03a x64 4_CPU 3190 +12/-14/=24 48.00% 24.0/50
Re: IvanHoe compiles testing
Posted: Tue Nov 23, 2010 5:01 pm
by 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: Rybka 4- 10 moves
Hash: 256
RB and TB: ON
Ponder: OFF
Blitz:10' 0
Code: Select all
1 IvanHoe 9.50b x64GA 3190 +25/-14/=61 55.50% 55.5/100
2 Deep Rybka 4 x64 3190 +14/-25/=61 44.50% 44.5/100
KLO version 099950t M compile
Re: IvanHoe compiles testing
Posted: Wed Nov 24, 2010 1:45 am
by robbolito
Intel(R) Core(TM) i5 Q- 750 @2.67 GHz 6.00 GB Memory
Windows 7.x64 Professional (Build 7600)
Fritz Benchmarks:
Speed:16.84
KNS: 8082
GUI: CB Rybka 3
Hash: 256
Book: Rybka 4-10 moves
RB and TB: ON
Ponder: OFF
Blitz:10 0
Code: Select all
1 Ivanhoe B49fAx64 3190 +8/-7/=25 51.25% 20.5/40
2 Deep Rybka 4 x64 3190 +7/-8/=25 48.75% 19.5/40
Re: IvanHoe compiles testing
Posted: Wed Nov 24, 2010 1:55 am
by robbolito
Intel(R) Core(TM) i7 4CPU 960 3.20GHz8x @ 4.005 GHz with 5.99 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: HS-10 moves
Hash: 256
RB and TB: ON
Ponder: OFF
Blitz:10' 0
Code: Select all
1 Ivanhoe B49fAx64p 3190 +6/-2/=32 55.00% 22.0/40
2 Houdini 1.03a x64 4_CPU 3190 +2/-6/=32 45.00% 18.0/40
Re: IvanHoe compiles testing
Posted: Wed Nov 24, 2010 2:25 am
by 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: Rybka 4-10 moves
Hash: 256
RB and TB: ON
Ponder: OFF
Blitz:10' 0
Code: Select all
1 Ivanhoe B49fAx64 3190 +10/-4/=26 57.50% 23.0/40
2 Houdini 1.03a x64 4_CPU 3190 +4/-10/=26 42.50% 17.0/40
Re: IvanHoe compiles testing
Posted: Wed Nov 24, 2010 2:16 pm
by robbolito
Intel(R) Core(TM) i5 Q- 750 @2.67 GHz 6.00 GB Memory
Windows 7.x64 Professional (Build 7600)
Fritz Benchmarks:
Speed:16.84
KNS: 8082
GUI: CB Rybka 3
Hash: 256
Book: Rybka 4-10 moves
RB and TB: ON
Ponder: OFF
Blitz:10 0
Code: Select all
1 Ivanhoe B49fAx64 3190 +9/-2/=29 58.75% 23.5/40
2 DeepSaros.exe ver.2.3c3 3190 +2/-9/=29 41.25% 16.5/40
Re: IvanHoe compiles testing
Posted: Fri Nov 26, 2010 7:46 pm
by kingliveson
Sample IvanHoe 999949j beta games:
Code: Select all
IvanHoe vs Houdini, Blitz 4m+2s 2010
1 IvanHoe 9.49b x64 +14 +9/=34/-7 52.00% 26.0/50
2 Houdini 1.03a x64 POPCNT 4_CP -14 +7/=34/-9 48.00% 24.0/50
Re: IvanHoe compiles testing
Posted: Sat Nov 27, 2010 12:35 am
by BB+
The last sample game shows how IvanHoe and Rybka differ in evaluation (I intentionally do not say "understanding") of rook endgames. The first game shows what happens when White plays 1. h3?!

The fourth game is almost a Father-like game, with the blocked structure. The fifth game shows the high/lows of RobboBases, with a strange reduction (Rxe6?!) into a "won" position that takes longer to win than doing it directly. With the tenth game (IvanHoe's win as Black), I don't think it signifies any real endgame superiority, but just that it so happened the value of passed pawns [at various stages] worked out to Black's favour in this case.
Re: IvanHoe compiles testing
Posted: Sat Nov 27, 2010 1:23 am
by kingliveson
BB+ wrote:The last sample game shows how IvanHoe and Rybka differ in evaluation (I intentionally do not say "understanding") of rook endgames. The first game shows what happens when White plays 1. h3?!

The fourth game is almost a Father-like game, with the blocked structure. The fifth game shows the high/lows of RobboBases, with a strange reduction (Rxe6?!) into a "won" position that takes longer to win than doing it directly. With the tenth game (IvanHoe's win as Black), I don't think it signifies any real endgame superiority, but just that it so happened the value of passed pawns [at various stages] worked out to Black's favour in this case.
Interesting analysis! As for the 1.h3, it comes from a test suite i designed. Some positions are purposely unbalanced. After a little over a thousand games, it might be safe to say Rybka understand it better than IvanHoe series.
The position below is one that IvanHoe tends to outplay Rybka though both won as white in the sample posted.
Yes, that Rxe6 had me scratching my head.