yes default parameters
I have open and shut aquarium a dozen times to ensure hash is cleared with same result - it must be because of aquarium because in fritz 11 gui
i get
New game, Blitz:25' Harlow
2bq1r1k/1pp3bp/r7/p3pp2/2PP2pn/P1N1P3/1BQ2PPP/3RRBK1 b - - 0 1
Analysis by Rybka 4 x64:
1...exd4 2.Nb5 Re8
+/= (0.65) Depth: 1 00:00:00
1...exd4 2.Nb5 Bd7 3.Bxd4 Kg8
+/= (0.69) Depth: 2 00:00:00
1...exd4 2.Nb5 Bd7 3.Bxd4 Bxd4 4.Nxd4 Re8
+/= (0.68) Depth: 3 00:00:00
1...exd4 2.Nb5 Qe7 3.Bxd4 Be5 4.Qb2 Ng6
+/= (0.60) Depth: 4 00:00:00
1...exd4 2.Nb5 Bd7 3.Bxd4 Bxd4 4.Nxd4 c5 5.Nb5 Rb6 6.Nd6
+/- (0.72) Depth: 5 00:00:00
1...exd4 2.Nb5 Bd7 3.Bxd4 Bxd4 4.Nxd4 c5 5.Nb5 Kg8 6.Nd6 Bc6
+/= (0.68) Depth: 6 00:00:00 2kN
1...e4 2.c5 Rh6 3.Bc4 Bd7 4.Qb3 b6 5.Nd5 Rc6 6.cxb6 cxb6 7.Nf4
+/- (0.73) Depth: 7 00:00:00 9kN
1...exd4 2.Nb5 Rg6 3.Bxd4 Bxd4 4.exd4 c6 5.Nc3 Re8
+/- (0.75) Depth: 8 00:00:00 34kN
1...exd4 2.Ne2 Bd7 3.c5 Qe8 4.Bxd4 Bxd4 5.Rxd4 Ra8 6.Nf4 Bc6 7.Red1 Be4
+/- (0.80) Depth: 9 00:00:01 135kN
1...exd4 2.Ne2 Bd7 3.c5 Rh6 4.Bxd4 Qe7 5.Nf4 Bc6 6.Qc3 Ng6 7.Bxg7+ Qxg7 8.Qxg7+ Kxg7
+/- (0.86) Depth: 10 00:00:01 161kN
1...e4 2.d5 Rg8 3.Nb5 Rh6 4.Qd2 c6 5.Bxg7+ Rxg7 6.dxc6 Qxd2 7.Rxd2 Rxc6 8.Rb1 Ng6 9.Nd4 Rc5
+/- (0.72) Depth: 10 00:00:02 184kN
1...e4 2.d5 Rg8 3.Nb5 Rh6 4.Qd2 Nf3+ 5.gxf3 gxf3 6.Kh1 Qh4 7.Bxg7+ Rxg7 8.h3 Qg5
+/= (0.60) Depth: 11 00:00:03 284kN
1...e4 2.d5 Rg8 3.Ne2 Rh6 4.Nf4 Qe7 5.c5 Nf3+ 6.gxf3 gxf3 7.Kh1
+/= (0.65) Depth: 11 00:00:03 299kN
1...e4 2.d5 Rh6 3.Nb5 Bd7 4.c5 Bxb2 5.Qxb2+ Kg8 6.d6
+/= (0.53) Depth: 12 00:00:04 430kN
1...e4 2.d5 Rh6 3.Nb5 Bd7 4.c5 Bxb2 5.Qxb2+ Kg8 6.d6
+/= (0.38) Depth: 12 00:00:08 816kN
1...e4 2.Ne2 Rh6 3.d5 Qd6 4.Bxg7+ Kxg7 5.Ng3 Qxa3 6.Ra1 Qe7 7.Rxa5 Ng6 8.Ra8 Qf6 9.Ne2 Ne5 10.Nf4 Qh4 11.h3
+/= (0.46) Depth: 12 00:00:12 1225kN
1...e4 2.Ne2 Rh6 3.d5 Qd6 4.Bxg7+ Kxg7 5.Ng3 Qxa3 6.Ra1 Qe7 7.Qc3+ Kg8 8.Rxa5 Ng6 9.Ra7 Ne5 10.Rea1 b6 11.Be2 Kg7 12.Ra8 Kg8
+/= (0.46) Depth: 13 00:00:13 1293kN
1...e4 2.Ne2 Rh6 3.d5 Ng6 4.Bxg7+ Kxg7 5.Qc3+ Kg8 6.g3 b6 7.c5 bxc5 8.Qxc5 Ne5 9.Nd4 Rd6 10.Bc4 Re8 11.Rc1 Nxc4 12.Qxc4 Rxd5 13.Qxc7 Qd6 14.Qc3
+/= (0.34) Depth: 14 00:00:19 1898kN
1...e4 2.Ne2 Rh6 3.d5 Ng6 4.Bxg7+ Kxg7 5.Qc3+ Kg8 6.g3 b6 7.c5 bxc5 8.Nf4 Nxf4 9.exf4 Qd6 10.Rb1 a4 11.Rb5 Qf6 12.Qxf6 Rhxf6 13.Rxc5 Rd6
+/= (0.29) Depth: 14 00:00:30 3059kN
1...e4 2.Ne2 Rh6 3.d5 Ng6 4.Bxg7+ Kxg7 5.Qc3+ Kg8 6.g3 b6 7.c5 bxc5 8.Nf4 Nxf4 9.exf4 Bb7 10.Qxc5 Rb6 11.h3 Qd6 12.Qxd6 Rxd6 13.hxg4 Bxd5 14.gxf5
+/= (0.28) Depth: 15 00:00:48 5104kN
1...e4 2.Ne2 Rh6 3.d5 Ng6 4.Bxg7+ Kxg7 5.Qc3+ Kg8 6.g3 b6 7.c5 bxc5 8.Nf4 Nxf4 9.exf4 Ba6 10.Qxc5 Bxf1 11.Rxf1 a4 12.Qc4 Rd6 13.Qxa4 Rxd5 14.Rxd5 Qxd5 15.Rd1 Qa2
+/= (0.32) Depth: 16 00:00:56 5866kN
1...Nf3+ 2.gxf3 gxf3 3.Kh1 Qh4
= (0.16) Depth: 16 00:01:14 8199kN
1...Nf3+ 2.gxf3 gxf3 3.Kh1 Qh4
= (0.01) Depth: 16 00:01:21 9159kN
1...Nf3+ 2.gxf3 gxf3 3.Kh1 Qh4
=/+ (-0.39) Depth: 16 00:01:38 11256kN
1...Nf3+ 2.gxf3 gxf3 3.Kh1 Qh4
-/+ (-1.19) Depth: 16 00:02:25 16722kN
(, Harlow 29.05.2010)
Test Position: Ribli - Lobron, Germany 1992
Moderator: Ras
-
meschle
- Posts: 200
- Joined: Tue Apr 01, 2008 11:53 pm
Re: Test Position: Ribli - Lobron, Germany 1992
Regards Mark
Quad Q6600 3.0 Ghz, Vista 64bit, 4gb Ram
Quad Q6600 3.0 Ghz, Vista 64bit, 4gb Ram
-
yanquis1972
- Posts: 1766
- Joined: Wed Jun 03, 2009 12:14 am
Re: Test Position: Ribli - Lobron, Germany 1992
hi mark --
on my desktop pc i was able to more or less replicate your results with aquarium (except my Rybka 4 x64 found Nf3+ @ d12 instead of 11, but still 4 ply earlier than i thought it 'should'). but get this, after failing to find Nf3+ until d16 twice more in the shredder GUI, i switched to chessbase & it found it once again, at d12. then i went to rybka 3 (which should be the same as fritz 11 i think), & again, Nf3+ was found at d12. my last stab, arena, was another victim of d16 'bug' in which the engines favor e4 for some time.
of course default settings, 512mb hash were used.
so the other interesting thing: like you, i could clear aquarium's hash all day & all day get the same result. but deep rybka 4x64, limited to one core, default parameters, 512mb hash, comes up with an entirely -different- variation involving Qg5 (as you can also see in a below post, although that was SSE4.2). it still hits Nf3+ at d16, of course..
what the hell is going on here?? any ideas? i thought single-core engine find depth was supposed to be identical at each run, that GUIs were (99%, at least) irrelevant to engine speed or performance, & that multicore rybka set to 1 core was the same engine as the one being marketed as Rybka x64 or 32. (though this last is an assumption on my part based on what vas said about rybka 3). what am i missing?
on my desktop pc i was able to more or less replicate your results with aquarium (except my Rybka 4 x64 found Nf3+ @ d12 instead of 11, but still 4 ply earlier than i thought it 'should'). but get this, after failing to find Nf3+ until d16 twice more in the shredder GUI, i switched to chessbase & it found it once again, at d12. then i went to rybka 3 (which should be the same as fritz 11 i think), & again, Nf3+ was found at d12. my last stab, arena, was another victim of d16 'bug' in which the engines favor e4 for some time.
of course default settings, 512mb hash were used.
so the other interesting thing: like you, i could clear aquarium's hash all day & all day get the same result. but deep rybka 4x64, limited to one core, default parameters, 512mb hash, comes up with an entirely -different- variation involving Qg5 (as you can also see in a below post, although that was SSE4.2). it still hits Nf3+ at d16, of course..
what the hell is going on here?? any ideas? i thought single-core engine find depth was supposed to be identical at each run, that GUIs were (99%, at least) irrelevant to engine speed or performance, & that multicore rybka set to 1 core was the same engine as the one being marketed as Rybka x64 or 32. (though this last is an assumption on my part based on what vas said about rybka 3). what am i missing?