Search found 406 matches

by brianr
Sat Mar 06, 2010 5:45 am
Forum: Computer Chess Club: Programming and Technical Discussions
Topic: Material tables
Replies: 21
Views: 5673

Re: Material tables

Just as a reminder, there was Sturman's "Beware The Bishop Pair" article from (the then) ICCA Vol 19, No. 2 (June, 1996) pg 83. Some of the data looks like this (winning percentages based on number of pawns [from 0-8] for given material combinations), from Table 3: BvN[9]= {500,512,506,503,495,...
by brianr
Wed Mar 03, 2010 10:39 pm
Forum: Computer Chess Club: Programming and Technical Discussions
Topic: Computer for CC?
Replies: 19
Views: 2327

Re: Computer for CC?

A dual socket system will be quite a bit more expensive, and will not be SMP. Look here for non-OC'd info http://www.tomshardware.com/charts/2009-desktop-cpu-charts-update-1/Fritz-11,1406.html and here for high-end options http://www.jens.tauchclub-krems.at/diverses/Schach/fritz9_benchmarks.html
by brianr
Wed Mar 03, 2010 12:37 am
Forum: Computer Chess Club: Programming and Technical Discussions
Topic: Computer for CC?
Replies: 19
Views: 2327

Re: Computer for CC?

Forgot to address your specific questions. Doubling speed adds about 50-100 elo, depending on the engine. On a quad core system with engines that support parallel search, this means a 4x increase so about 100-200 elo. However, even single core engines benefit. 64 bits is also a significant benefit (...
by brianr
Wed Mar 03, 2010 12:05 am
Forum: Computer Chess Club: Programming and Technical Discussions
Topic: Computer for CC?
Replies: 19
Views: 2327

Re: Computer for CC?

A key question is are you comfortable running an overclocked system. If so, most seem to prefer the i920 these days. In years past, the Q6600 was a good choice. I have been running an OC'd Q6600 (going from 2.4 to just 3.1 GHz) for a couple of years with no problems (granted 3.1 is hardly pushing th...
by brianr
Sun Feb 28, 2010 6:24 pm
Forum: Computer Chess Club: Programming and Technical Discussions
Topic: Mirror.epd Test Positions
Replies: 6
Views: 940

Re: Mirror.epd Test Positions

Thank you. I think it is pretty clear that "mirroring" (same position but switching sides) should be evaluated and scored equally. However, this set (of fours) raises the question: should left/right reflections (for the same side) always be symmetric ? For example, piece square tables may tend to fa...
by brianr
Sun Feb 14, 2010 4:04 pm
Forum: Computer Chess Club: General Topics
Topic: CCT12 time
Replies: 18
Views: 1398

Re: CCT12 time

It would be quite helpful to also post the time here and on the CCT12 web page for the first round as FICS "server time",
then anyone can just log in and confirm their local time relative to the server "date" command (PST).

Thank you,
A Time Zone Challenged
Brian
by brianr
Sat Feb 13, 2010 11:47 pm
Forum: Computer Chess Club: Programming and Technical Discussions
Topic: Mirror.epd Test Positions
Replies: 6
Views: 940

Mirror.epd Test Positions

I have been using a mirror.epd test suite with Tinker since at least Nov, 2003, but foolishly had been using it only when changing things like the evaluation. Well, today I found a SEE bug using this suite and thought I would recommend it here for more general testing. I have looked a bit online and...
by brianr
Sat Feb 13, 2010 11:39 pm
Forum: Computer Chess Club: Programming and Technical Discussions
Topic: Repetition Detection Without Hashing
Replies: 10
Views: 2720

Re: Repetition Detection Without Hashing

Older versions of GNUChess used a hashless repetition approach with some lists of from/to squares, obviously resetting for non-reversible moves. I found some code in version 4.0 in search.c It takes some time to figure out, IIRC inline int repetition () /* Check for draw by threefold repetition. */ ...
by brianr
Fri Jan 29, 2010 11:49 pm
Forum: Computer Chess Club: Programming and Technical Discussions
Topic: request assistance or advice on an obscure bug
Replies: 12
Views: 1615

Re: request assistance or advice on an obscure bug

Does your move ordering not vary with initialization (for example, are killers cleared)?

Do you generate check evasions (perhaps turn this off and see if it matters)?

How do you handle repetition and 50 move draws (especially with no hash--a separate table)?

Do you have a perpetual check test?
by brianr
Sat Jan 23, 2010 2:27 pm
Forum: Computer Chess Club: General Topics
Topic: Question about comptuer speed
Replies: 4
Views: 954

Re: Question about comptuer speed

I also have a Q6600 (but under Vista). Overclocking from the stock 2.4GHz to 3.1 is not hard (and has proved _very_ stable for 2+ years), provided you have an overclockable motherboard and more than the stock cooling fan. As far as tweaking various services and ancillary background processes, just d...