OliThink 5.3.2 released (Source, Windows, Mac and Linux)

[OliverBr]

One week after 5.3.1 there is a new version of OliThink. A bug was removed. The move generation is now completely splitted.
This seems to improve a lot! Behold: OliThink 5.3.2 makes against Crafty 20.14-32bit at very short time controls about 50%!

-The Java Version is still in work.

-Additional commands "perft" and "divide" added.

-Analyze mode still not implemented, but the winboard command "st" in order to have work-around.

-Thanks in advance to Jim Ablett for his builds. Jim, please do not bother to implement analyze=1 as it is not possible with the current engine.

Download here as usual:

http://home.arcor.de/dreamlike/chess/index.html

Thank you all for your support.

[lucasart]

One week after 5.3.1 there is a new version of OliThink. A bug was removed. The move generation is now completely splitted.
This seems to improve a lot! Behold: OliThink 5.3.2 makes against Crafty 20.14-32bit at very short time controls about 50%!

-The Java Version is still in work.

-Additional commands "perft" and "divide" added.

-Analyze mode still not implemented, but the winboard command "st" in order to have work-around.

-Thanks in advance to Jim Ablett for his builds. Jim, please do not bother to implement analyze=1 as it is not possible with the current engine.

Download here as usual:

http://home.arcor.de/dreamlike/chess/index.html

Thank you all for your support.

Good news. I'll test v5.3.2 asap!

[Jim Ablett]

Scid will ignore 'feature' because it is merged with wb-xboard command (non-standard) instead of in 'protover'. (so setboard will be set to 0)

Code: Select all

int protV2(char* buf) {
		if (!strncmp(buf,"xboard",6)) printf("feature setboard=1 myname=\"OliThink " VER "\" colors=0 analyze=0 done=1\n");
		else if (!strncmp(buf,"quit",4)) return -2;
		else if (!strncmp(buf,"new",3)) return -3; 
		else if (!strncmp(buf,"remove",6)) return -4;
		else if (!strncmp(buf,"force",5)) engine = -1;
		else if (!strncmp(buf,"go",2)) engine = pondering ? onmove^1 : onmove;
		else if (!strncmp(buf,"setboard",8)) _parse_fen(buf+9);
		else if (!strncmp(buf,"undo",4)) undo();
		else if (!strncmp(buf,"easy",4)) ponder = 0;
		else if (!strncmp(buf,"hard",4)) ponder = 1;
		else if (!strncmp(buf,"sd",2)) sscanf(buf+3,"%d",&sd);
		else if (!strncmp(buf,"time",4)) sscanf(buf+5,"%d",&ttime);
		else if (!strncmp(buf,"level",4)) sscanf(buf+6,"%d %d %d",&mps, &base, &inc);
		else if (!strncmp(buf,"post",4)) post = 1;
		else if (!strncmp(buf,"nopost",6)) post = 0;
		else if (!strncmp(buf,"result",6));//result 0-1 {Black mates}
		else if (!strncmp&#40;buf,"otim",4&#41;);//otim <optime>
		else if (!strncmp&#40;buf,"draw",4&#41;);//draw offer 
		else if (!strncmp&#40;buf,"st",2&#41;) sscanf&#40;buf+3,"%d",&st&#41;;
		else if (!strncmp&#40;buf,"bk",2&#41;);
		else if (!strncmp&#40;buf,"hint",4&#41;);
		else if (!strncmp&#40;buf,"computer",8&#41;);
		else if (!strncmp&#40;buf,"protover",8&#41;);//protover 2
		else if (!strncmp&#40;buf,"accepted",8&#41;);//accepted <feature>
		else if (!strncmp&#40;buf,"random",6&#41;);
		else if (!strncmp&#40;buf,"rating",6&#41;);//ICS&#58; rating <myrat> <oprat>
		else if (!strncmp&#40;buf,"name",4&#41;);//ICS&#58; name <opname>
		else if (!strncmp&#40;buf,"perft",5&#41;) &#123;int i; for &#40;i = 1; i <= sd; i++) printf&#40;"Depth&#58; %d Nodes&#58; %llu\n", i, perft&#40;onmove, i, 0&#41;);&#125;
		else if (!strncmp&#40;buf,"divide",5&#41;) perft&#40;onmove, sd, 1&#41;;
		else return -1;
		return 0;
&#125;

Jim.

[OliverBr]

Scid will ignore 'feature' because it is merged with wb-xboard command (non-standard) instead of in 'protover'. (so setboard will be set to 0)

Is it safe to assume that any relevant GUI is using Protocol 2 and thus sending "protover 2" to the engine?

[stevenaaus]

Scid will ignore 'feature' because it is merged with wb-xboard command (non-standard) instead of in 'protover'. (so setboard will be set to 0)

I'm not sure what you're saying here.
Scid (vs. PC) seems to detect 'feature', 'setboard' fine.

BTW, Scid vs. PC's tournament does not work with Olithink under Linux at per-game time controls. I'm not sure of the reason, as xboard tournament support isn't so great.
Cheers.

Code: Select all

Scid-Engine communication log file
Engine: OliThink-5.3.2
Command: olithink5.3.2
Date: Mon Mar 05 06:16:55 EST 2012

This file was automatically generated by Scid.
It is rewritten every time an engine is started in Scid.

NOTE  : Quiet engine (still no output); sending it initial commands.
Scid  : xboard
Scid  : protover 2
Scid  : easy
Scid  : post
Scid  : force
Engine: feature setboard=1 myname="OliThink 5.3.2" colors=0 analyze=0 done=1
NOTE  : First line from engine seen; sending it initial commands now.
Scid  : xboard
Scid  : protover 2
Scid  : easy
Scid  : post
Engine: feature setboard=1 myname="OliThink 5.3.2" colors=0 analyze=0 done=1
Scid  : setboard rnbqkbnr/pppppppp/8/8/3P4/8/PPP1PPPP/RNBQKBNR b KQkq - 0 1
Scid  : new
Scid  : force
Scid  : setboard rnbqkbnr/pppppppp/8/8/3P4/8/PPP1PPPP/RNBQKBNR b KQkq - 0 1
Scid  : black
Scid  : st 120000
Scid  : sd 50
Scid  : post
Scid  : go
Engine: UNKNOWN COMMAND: black
Engine:  1     2      0        25  e7e6 
Engine:  2   -33      0       184  d7d5 c1f4 
Engine:  3     0      0       426  d7d5 c1f4 c8g4 
Engine:  4   -27      0      2801  e7e6 e2e4 d8h4 d1f3 
Engine:  5   -14      3     12235  e7e6 g1f3 g8f6 f3e5 f6e4 
Engine:  6   -10     11     46276  g8f6 e2e3 d7d5 b1c3 c8f5 f1d3 
Engine:  7    -3     29    121199  g8f6 g1f3 d7d6 e2e3 f6e4 f1b5 c7c6 b5c4 
Engine:  8    -9     66    282165  g8f6 c1d2 d7d6 e2e3 f6e4 f2f3 e4d2 b1d2 
Engine:  9   -11    212    950685  g8f6 e2e3 d7d5 f1d3 b8c6 c1d2 e7e5 d4e5 c6e5 
Engine: 10   -19    918   4290628  g8f6 e2e3 d7d6 g1f3 b8c6 f1b5 c8d7 b1d2 f6d5 b5d3 
Engine: 11    -7   1925   9249650  g8f6 g1f3 e7e6 b1c3 b8c6 a2a3 d7d5 c1f4 f6e4 c3b5 f8d6 
Scid  : exit
Scid  : quit

[Jim Ablett]

I'm not sure what you're saying here.
Scid (vs. PC) seems to detect 'feature', 'setboard' fine.

BTW, Scid vs. PC's tournament does not work with Olithink under Linux at per-game time controls. I'm not sure of the reason, as xboard tournament support isn't so great.
Cheers.

Yes, you're quite correct it is working fine with latest update.
It must have been my faulty analysis mode implementation in v5.31. I could only play from standard fen starting position until I separated wb-xboard from wb-protover.

Jim.

[stevenaaus]

Hmmm - i was doing something wrong with the tournament.

http://www.talkchess.com/forum/viewtopi ... 605#453605

[Sylwy]

Hi Mr. Brausch !

All is OK with the Java version ( here I'm very interested).

Thank you a lot !

SilvianR