Search found 1392 matches

by Houdini
Fri Jul 16, 2010 8:34 pm
Forum: Computer Chess Club: General Topics
Topic: Houdini 1.03 is available
Replies: 101
Views: 33857

Re: Houdini 1.03 is available

Osipov Jury wrote:
Alex Lobanov wrote:Robert Houdart
Why in the engine made a distortion evaluation?

:?:
Very interesting question.
Not sure that I entirely grasp the question, but you may find the answer in this interesting topic Larry Kaufman recently started: http://www.talkchess.com/forum/viewtopic.php?t=35458 .

Robert
by Houdini
Fri Jul 16, 2010 12:30 pm
Forum: Computer Chess Club: General Topics
Topic: Houdini 1.03 is available
Replies: 101
Views: 33857

Re: Houdini 1.03 is available

By the way, people with 4 or 8 cores should experiment with the Split_Depth UCI parameter to achieve the best performance on their hardware.
The default value is 10 (tuned on my Core i5-750 @2.66 GHz), but on some systems it may be advantageous to increase this to 12 or 13.

Robert
by Houdini
Fri Jul 16, 2010 9:51 am
Forum: Computer Chess Club: General Topics
Topic: Houdini 1.03 is available
Replies: 101
Views: 33857

Re: Houdini 1.03 is available

Lonnie, here's a link to the old Houdini 1.02 w32 2CPU.

Robert
by Houdini
Fri Jul 16, 2010 3:14 am
Forum: Computer Chess Club: General Topics
Topic: Houdini 1.03 is available
Replies: 101
Views: 33857

Re: Houdini 1.03 is available

LJC wrote:I need 32 bit version 2CPU 1.02 version
What exactly do you mean by "it will not work"? Do you get an error message when you double-click the 1.03 executable in Windows Explorer?
On what hardware are you running?
by Houdini
Fri Jul 16, 2010 3:03 am
Forum: Computer Chess Club: General Topics
Topic: Houdini 1.03 is available
Replies: 101
Views: 33857

Re: Houdini 1.03 is available

Could you tell me how I can get Houdini 1.02, I just tried installing Houdini 1.03 and it will not work on my machine, Houdini 1.02 worked fine. I overwrote 1.2 and now I have no Houdini, I used it alot on ICC, very strong engine What exactly do you mean by "it will not work"? Which version do you ...
by Houdini
Fri Jul 16, 2010 2:58 am
Forum: Computer Chess Club: General Topics
Topic: Houdini 1.03 is available
Replies: 101
Views: 33857

Re: Houdini 1.03 is available

mhalstern wrote:Robert thanks,

Without enabling large pages, how many elo points stronger is it than version 1,02?
In fast time controls about 20 Elo, in long time controls probably somewhat less.

Robert
by Houdini
Fri Jul 16, 2010 12:01 am
Forum: Computer Chess Club: General Topics
Topic: Mate in 17 revisited
Replies: 18
Views: 3475

Re: Mate in 17 revisited

A problem by Joseph Halumbriek (Mate in in 17). [D]6k1/8/5Q2/8/8/8/3p1pp1/3Kbrrb w - - 0 1 is only solved by YACE and of course by ChestUCI Houdini 1.03 contains some improvement in zugzwang detection and will find a mate fairly rapidly: FEN: 6k1/8/5Q2/8/8/8/3p1pp1/3Kbrrb w - - 0 1 Houdini 1.03: .....
by Houdini
Thu Jul 15, 2010 9:57 pm
Forum: Computer Chess Club: General Topics
Topic: Houdini 1.03 is available
Replies: 101
Views: 33857

Houdini 1.03 is available

The new Houdini 1.03 is available. Download from http://www.cruxis.com/chess/houdini.htm . Houdini 1.03 extends the UCI features of the engine with multi-PV analysis, search moves and large memory page support. The web page contains some explanation of how to use large page support on Windows Vista/...
by Houdini
Mon Jul 12, 2010 8:14 am
Forum: Computer Chess Club: Programming and Technical Discussions
Topic: Komodo 1.2 crash in KPK ending
Replies: 4
Views: 994

Re: Komodo 1.2 crash in KPK ending

Here's the latest occurrence, another drawn KPK: [White "Komodo_1.2"] [Black "Houdini_CDFRP"] [Result "0-1"] [FEN "r1bq1rk1/ppp1bppp/3p1n2/8/3QPB2/2N5/PPP1BPPP/R3K2R w KQ -"] [TimeControl "40/5"] 1. O-O {+0.74/9 0.092s} Be6 {-0.36/12 0.19s} 2. h3 {+0.76/8 0.075s} Nd7 {-0.30/12 0.13s} 3. Nd5 {+0.59/9...
by Houdini
Sun Jul 11, 2010 1:31 pm
Forum: Computer Chess Club: Programming and Technical Discussions
Topic: Stockfish 1.8 tweaks
Replies: 53
Views: 11464

Re: Stockfish 1.8 tweaks

Doesn't the current formula for R (and also the proposed modification) rely on the value of OnePly being 2?
In other words, instead of

Code: Select all

int R = 3 + (depth >= 5 * OnePly ? depth / 8 : 0);
I suggest

Code: Select all

int R = 3 + (depth >= 5 * OnePly ? depth / (4 * OnePly) : 0); 
Robert