ChessUSA.com TalkChess.com
Hosted by Your Move Chess & Games
 
 FAQFAQ   SearchSearch   MemberlistMemberlist   UsergroupsUsergroups   RegisterRegister 
 ProfileProfile   Log in to check your private messagesLog in to check your private messages   Log inLog in 

Djinn 1.006
Post new topic    TalkChess.com Forum Index -> Computer Chess Club: General Topics Flat
View previous topic :: View next topic  
Author Message
Tom Likens



Joined: 28 Apr 2012
Posts: 302
Location: Austin, TX

PostPost subject: Djinn 1.006    Posted: Tue Dec 24, 2013 1:33 am Reply to topic Reply with quote

Djinn 1.006 is now available at:

http://webpages.charter.net/tlikens/index.html

Don't make to much out of the transition to 1.xxx, since the numbering for Djinn is a little odd. Basically, I use the changes in the history file to set the version number (I have a python script that parses it). So the transition to 1.xxx basically means I need to go outside more!

As with 0.992 this is mainly a bug fix with a small strength increase (approx. 20 elo I would estiimate from my testing). Here are the changes since the last version was released:

Code:
17-DEC-2013 (ver. 1.006)
 ------------------------
 E  o Modified the way the penalty/bonus for a Queen vs. 2 Rooks trade
      was being scored.  I also ran this under CLOP to get optimal (or
      at least not "sub-optimal") scaling values.

 16-DEC-2013
 -----------
 E  o I changed the way the pawns were being scored in front of the king.
      Previously, the pawns occupying the six squares in front of the
      king's actual location were used to access a "Shelter[64]" array. 
      But using the king's actual position would occasionally cause the
      program to move the king forward (in an unsafe manner) so that it
      would be closer to any kingside pawns that had advanced.  Now the
      program now uses the canonical kingside square, (either G1/G8 or
      C1/C8).  This seems to result in less unsafe king forays.   

 07-DEC-2013
 -----------
 S  o Removed the "Botvinnik-Markoff" extension.  Extensive testing
      showed no real elo gain from it and removing it simplifies the code.
 E  o Clopped the rook mobility tables.

 04-DEC-2013
 -----------
 E  o Added safe king mobility to the evaluation function.  For some odd
      reason this was left out when I added the mobility for all the
      other pieces to the evaluation function.  Clopped both the penalty
      and mobility multiplier portions.

 20-NOV-2013
 -----------
 S  o Changed the criteria for beta pruning, so that it is only applied
      if the hash move is empty.  It's also skipped if either alpha or
      beta is a mate score, (worth about 5 elo).
 B  o Potential bug fix for the Windows version.  I realized that the
      "unsigned long" type on Windows is only 4 bytes, (under Linux it's
      8 bytes).  It's possible that some of my time management and node
      counting variables were under/overflowing.  So rather than worry
      about it I simply changed everything to "unsigned long long"
      which is 8 bytes on both operating systems.

 17-NOV-2013
 -----------
 I  o Dusted off Andrei Fortuna's "Chant" utility and started making
      modifications to both Djinn and the Chant viewer (Andrei was kind
      enough to give me the source code) so that more information is
      shown.  I'm also planning to speed up the loading of the database
      file so that the viewing of large files isn't so slow.

 13-NOV-2013
 -----------
 O  o Micro-Optimized to the swapoff routine that separates winning,
      even and losing captures. Since the captures are generated in
      MVV/LVA order and stable sorted, we don't need the full SEE score. 
      Instead, we only need to know if the SEE score is >= 0.  This is
      always satisfied if the piece capturing is of lesser or equal
      value to the piece being captured.  If that is true we skip the
      (relatively) expensive call to the full swapoff routine.

 05-NOV-2013
 -----------
 B  o Found a minor bug in the "toolbox::convert_string_to_double()"
      function which could actually cause an interface crash if the
      input was pathological.

 03-NOV-2013
 -----------
 B  o While going over the evaluation routines, I ran across a bug.
      Basically, the king attacks were being left out of the attack
      bitmaps if the endgame evaluation function exited early,
      (specifically, if it was a king/pawn-only endgame).  Probably,
      not a big elo hit, but certainly not the intent.
 E  o Added a new function that calculates the attack bitmaps from
      scratch.  This is a useful debug routine.  This is now used to
      restrict the definition of "safe" squares that the queen and
      rook can move to (i.e. their mobility).

 27-OCT-2013
 -----------
 B  o Fixed a memory overrun bug found by Valgrind in the hash routines.
      It was possible for the slot pointers to read or write past the
      end of the allocated memory if the initial hash value into the
      table was the last (or 2nd to last) slot.
 B  o Fixed a buffer underflow read into "ThreatMove" array (found by
      the LLVM Clang compiler using the '-fsanitize=address' option).


regards,
--tom
Back to top
View user's profile Send private message Visit poster's website
Display posts from previous:   
Subject Author Date/Time
Djinn 1.006 Tom Likens Tue Dec 24, 2013 1:33 am
      Re: Djinn 1.006 Tony Mokonen Tue Dec 24, 2013 6:33 am
            Re: Djinn 1.006 Tom Likens Tue Dec 24, 2013 2:36 pm
      Re: Djinn 1.006 Ed Schroder Tue Dec 24, 2013 7:38 am
            Re: Djinn 1.006 Tom Likens Tue Dec 24, 2013 2:35 pm
                  Re: Djinn 1.006 Ed Schroder Wed Dec 25, 2013 5:20 am
      Re: Djinn 1.006 Gabor Szots Tue Dec 24, 2013 8:21 am
      Re: Djinn 1.006 Gabor Szots Tue Dec 24, 2013 11:24 am
            Re: Djinn 1.006 Tom Likens Tue Dec 24, 2013 1:25 pm
                  Re: Djinn 1.006 Tom Likens Tue Dec 24, 2013 1:59 pm
                        Re: Djinn 1.006 Gabor Szots Tue Dec 24, 2013 2:44 pm
                              Re: Djinn 1.006 Tom Likens Tue Dec 24, 2013 3:43 pm
                                    Re: Djinn 1.006 Geert Maeckelbergh Fri Dec 27, 2013 2:18 pm
                                          Re: Djinn 1.006 Tom Likens Fri Dec 27, 2013 2:26 pm
            Re: Djinn 1.006 Tony Mokonen Tue Dec 24, 2013 10:52 pm
                  Re: Djinn 1.006 Tom Likens Wed Dec 25, 2013 12:28 am
                  Re: Djinn 1.006 Tom Likens Wed Dec 25, 2013 3:38 pm
                        Re: Djinn 1.006 Graham Banks Fri Dec 27, 2013 9:24 am
                              Re: Djinn 1.006 Tom Likens Fri Dec 27, 2013 12:42 pm
                                    Re: Djinn 1.006 Gabor Szots Fri Dec 27, 2013 1:28 pm
                                          Re: Djinn 1.006 Tom Likens Fri Dec 27, 2013 2:25 pm
      Re: Djinn 1.006 Graham Banks Tue Dec 24, 2013 11:09 pm
Post new topic    TalkChess.com Forum Index -> Computer Chess Club: General Topics

 
Jump to:  
You cannot post new topics in this forum
You cannot reply to topics in this forum
You cannot edit your posts in this forum
You cannot delete your posts in this forum
You cannot vote in polls in this forum




Powered by phpBB © 2001, 2005 phpBB Group
Enhanced with Moby Threads