Ana stands for Analysis
For analysis I configured the engine like this:
Never clear hash
Contempt = 0
Split Depth 12
8 GByte hash
The system has 16GByte memory, so I can use GB large pages
The Split Deth is otimized for the use of 8 GB hash
Fascinating chess position
Moderator: Ras
-
zullil
- Posts: 6442
- Joined: Tue Jan 09, 2007 12:31 am
- Location: PA USA
- Full name: Louis Zulli
Re: Fascinating chess position
Is a hash table from earlier searches being loaded automatically?Waschbaer wrote:Ana stands for Analysis
For analysis I configured the engine like this:
Never clear hash
-
Waschbaer
- Posts: 68
- Joined: Mon Dec 12, 2011 11:27 pm
Re: Fascinating chess position
No,
in the config there are buttons "Load/save hashfile" and a field to file the name of the hashfile. I didn't use that feature.
in the config there are buttons "Load/save hashfile" and a field to file the name of the hashfile. I didn't use that feature.
-
ernest
- Posts: 2059
- Joined: Wed Mar 08, 2006 8:30 pm
Re: Fascinating chess position
Thanks Günther,Waschbaer wrote:Ana stands for Analysis
I still get only (0.40) eval, long after depth 34...
And if you repeat with clear hash, that is by reloading Houdini (no need to reboot...), it still picks up at depth 34?...
-
ernest
- Posts: 2059
- Joined: Wed Mar 08, 2006 8:30 pm
Re: Fascinating chess position
OK, got it for the 1st time, depth 40 (as Robert H mentioned)
Core2 Duo @3GHz, 1024 MB Hash, Contempt 0, Split Depth 12 (in order to somewhat imitate Günther
)
New game
8/8/2k1pn2/5p2/p1pP1Pp1/2P1K1P1/P7/4B3 b - - 0 1
Analysis by Houdini 2.0c w32 Nlm:
............
1...Ne4 2.Ke2 Kc7 3.Ke3 a3 4.Ke2 Kd8 5.Ke3 Kd7 6.Ke2 Kc6 7.Ke3 Kd6 8.Ke2 Ke7 9.Ke3 Nf6 10.Ke2 Kf8 11.Ke3 Nd5+ 12.Ke2 Ke7 13.Bd2 Kd6 14.Kf1 Nf6 15.Ke2 Kd7 16.Be1 Ne4 17.Ke3 Ke7 18.Ke2 Ke8 19.Ke3 Kf7 20.Ke2 Kg7 21.Ke3 Kf6 22.Ke2 Ke7 23.Ke3
=/+ (-0.40) Depth: 38/75 00:59:54 11910mN, tb=2491
1...Ne4 2.Ke2 Kc7 3.Ke3 a3 4.Ke2 Kd8 5.Ke3 Kd7 6.Ke2 Kc6 7.Ke3 Kd6 8.Ke2 Ke7 9.Ke3 Nf6 10.Ke2 Kf8 11.Ke3 Nd5+ 12.Ke2 Ke7 13.Bd2 Kd6 14.Kf1 Nf6 15.Ke2 Kd5 16.Be1 Kc6 17.Kf1 Kd6 18.Ke2 Ke7 19.Kd1 Ne4 20.Ke2 Ke8 21.Ke3 Kf7 22.Ke2 Kg7 23.Ke3 Kf6 24.Ke2 Ke7 25.Ke3
=/+ (-0.40) Depth: 39/75 01:50:21 22327mN, tb=6309
1...Ne4
=/+ (-0.45 !) Depth: 40/81 03:20:31 41340mN, tb=9729
1...Ne4
=/+ (-0.58 !) Depth: 40/81 03:32:36 44017mN, tb=10228
1...Ne4
-/+ (-0.85 !) Depth: 40/86 03:51:51 48196mN, tb=10675
Core2 Duo @3GHz, 1024 MB Hash, Contempt 0, Split Depth 12 (in order to somewhat imitate Günther
New game
8/8/2k1pn2/5p2/p1pP1Pp1/2P1K1P1/P7/4B3 b - - 0 1
Analysis by Houdini 2.0c w32 Nlm:
............
1...Ne4 2.Ke2 Kc7 3.Ke3 a3 4.Ke2 Kd8 5.Ke3 Kd7 6.Ke2 Kc6 7.Ke3 Kd6 8.Ke2 Ke7 9.Ke3 Nf6 10.Ke2 Kf8 11.Ke3 Nd5+ 12.Ke2 Ke7 13.Bd2 Kd6 14.Kf1 Nf6 15.Ke2 Kd7 16.Be1 Ne4 17.Ke3 Ke7 18.Ke2 Ke8 19.Ke3 Kf7 20.Ke2 Kg7 21.Ke3 Kf6 22.Ke2 Ke7 23.Ke3
=/+ (-0.40) Depth: 38/75 00:59:54 11910mN, tb=2491
1...Ne4 2.Ke2 Kc7 3.Ke3 a3 4.Ke2 Kd8 5.Ke3 Kd7 6.Ke2 Kc6 7.Ke3 Kd6 8.Ke2 Ke7 9.Ke3 Nf6 10.Ke2 Kf8 11.Ke3 Nd5+ 12.Ke2 Ke7 13.Bd2 Kd6 14.Kf1 Nf6 15.Ke2 Kd5 16.Be1 Kc6 17.Kf1 Kd6 18.Ke2 Ke7 19.Kd1 Ne4 20.Ke2 Ke8 21.Ke3 Kf7 22.Ke2 Kg7 23.Ke3 Kf6 24.Ke2 Ke7 25.Ke3
=/+ (-0.40) Depth: 39/75 01:50:21 22327mN, tb=6309
1...Ne4
=/+ (-0.45 !) Depth: 40/81 03:20:31 41340mN, tb=9729
1...Ne4
=/+ (-0.58 !) Depth: 40/81 03:32:36 44017mN, tb=10228
1...Ne4
-/+ (-0.85 !) Depth: 40/86 03:51:51 48196mN, tb=10675
-
zullil
- Posts: 6442
- Joined: Tue Jan 09, 2007 12:31 am
- Location: PA USA
- Full name: Louis Zulli
Re: Fascinating chess position
Split Depth 12 using 2 threads seems like a very bad idea. But who knows.ernest wrote: Core2 Duo Split Depth 12 (in order to somewhat imitate Günther)
-
Waschbaer
- Posts: 68
- Joined: Mon Dec 12, 2011 11:27 pm
Re: Fascinating chess position
You should find the best Split Depth for your System by yourself.
How to do?
Best is reading the Houdinis User's Manual Section 3.4
http://www.cruxis.com/chess/manual/index.html
Have fun!
How to do?
Best is reading the Houdinis User's Manual Section 3.4
http://www.cruxis.com/chess/manual/index.html
Have fun!
-
ernest
- Posts: 2059
- Joined: Wed Mar 08, 2006 8:30 pm
Re: Fascinating chess position
OK here is another try, this time with my default valueszullil wrote:Split Depth 12 using 2 threads seems like a very bad idea. But who knows.
(Core2 Duo @3GHz, 512 MB Hash, Contempt 1, Split Depth 10)
Got it at depth 39!
I think Günther's depth 34 is not reproducible...
New game
8/8/2k1pn2/5p2/p1pP1Pp1/2P1K1P1/P7/4B3 b - - 0 1
Analysis by Houdini 2.0c w32 Nlm:
......................
1...Ne4 2.Ke2 a3 3.Ke3 Kd5 4.Ke2 Kd6 5.Ke3 Kd7 6.Ke2 Kc7 7.Ke3 Kd6 8.Ke2 Kd5 9.Ke3 Kc6 10.Ke2 Kd7 11.Ke3 Nf6 12.Ke2 Ke7 13.Bd2 Nh5 14.Be1 Kd7 15.Kd1 Kc7 16.Kd2 Nf6 17.Ke2 Kd6 18.Ke3 Kd7 19.Ke2
=/+ (-0.40) Depth: 36/74 00:28:45 6062mN, tb=683
1...Ne4 2.Ke2 a3 3.Ke3 Kd5 4.Ke2 Kd6 5.Ke3 Kd7 6.Ke2 Kc7 7.Ke3 Kd6 8.Ke2 Kd5 9.Ke3 Kc6 10.Ke2 Kd7 11.Ke3 Nf6 12.Ke2 Ke7 13.Bd2 Nh5 14.Be1 Kd7 15.Kd1 Kc7 16.Kd2 Nf6 17.Ke2 Kd6 18.Ke3 Kd7 19.Ke2
=/+ (-0.40) Depth: 37/75 01:17:02 16586mN, tb=3505
1...Ne4 2.Ke2 a3 3.Ke3 Kd5 4.Ke2 Kd6 5.Ke3 Kd7 6.Ke2 Kc7 7.Ke3 Kd6 8.Ke2 Kd5 9.Ke3 Kc6 10.Ke2 Kd7 11.Ke3 Nf6 12.Ke2 Ke7 13.Bd2 Nh5 14.Be1 Kd7 15.Kd1 Kc7 16.Kd2 Nf6 17.Ke2 Kd6 18.Ke3 Kd7 19.Ke2
=/+ (-0.40) Depth: 38/75 01:49:18 23459mN, tb=3986
1...Ne4
=/+ (-0.45 !) Depth: 39/82 02:22:50 30829mN, tb=6724
1...Ne4
=/+ (-0.58 !) Depth: 39/82 02:34:38 33471mN, tb=6795
1...Ne4
-/+ (-0.85 !) Depth: 39/82 02:54:14 37861mN, tb=6951