Search found 811 matches

by elcabesa
Sat Apr 20, 2019 3:17 pm
Forum: Computer Chess Club: Programming and Technical Discussions
Topic: No standard specification for Perft
Replies: 22
Views: 4619

Re: No standard specification for Perft

In my engine I use perft and divide to test the correctness of the engine. Whit this goal, I use the same move generator, do/undo move routines and the move selector. I tested my engine in a lot of random positions and checked the result against stockfish. This helped me validating the code. Sometim...
by elcabesa
Sun Apr 14, 2019 7:23 am
Forum: Computer Chess Club: General Topics
Topic: Vajolet2 2.7 release
Replies: 6
Views: 1400

Re: Vajolet2 2.7 release

Wow, after a week of CCRL games, it looks likes that Vajolet2 2.7 has gained 30 Elo point over Vajolet2 2.6 :D

a really pleasant discovery.

Thank you CCRL boys
by elcabesa
Wed Apr 03, 2019 8:44 pm
Forum: Computer Chess Club: General Topics
Topic: Vajolet2 2.7 release
Replies: 6
Views: 1400

Re: Vajolet2 2.7 release

and always thank you all for your testing
by elcabesa
Wed Apr 03, 2019 8:32 pm
Forum: Computer Chess Club: General Topics
Topic: Vajolet2 2.7 release
Replies: 6
Views: 1400

Vajolet2 2.7 release

With pleasure I'd like to announce the release of Vajolet2 2.7 Vajolet can be downloaded on github Release page from my onw testing I hope in an increment of 10, maybe 20 Elo Point below the long release note from github page: release note 2.7 fix: 2.6 crash debug wrong scores when draw is reached s...
by elcabesa
Fri Mar 15, 2019 10:43 am
Forum: Computer Chess Club: Programming and Technical Discussions
Topic: CPW bitscan with reset could someone explain this line?
Replies: 9
Views: 1570

Re: CPW bitscan with reset could someone explain this line?

I understand that it confuse the compiler about keeping the data in cpu register instead that in ram.
Having them in register should be faster
by elcabesa
Mon Feb 18, 2019 3:33 pm
Forum: Computer Chess Club: Programming and Technical Discussions
Topic: On failing high and finding mates
Replies: 14
Views: 2249

Re: On failing high and finding mates

It's not correct. But I don't think this is a very big problem.
You'll get only some glitch with some gui when looking the interface.
by elcabesa
Sun Feb 17, 2019 9:47 pm
Forum: Computer Chess Club: Programming and Technical Discussions
Topic: On failing high and finding mates
Replies: 14
Views: 2249

Re: On failing high and finding mates

I am not sure about your exact problem. An "info" line containing "lowerbound" is not a PV, so there is no need to expect a mate score in it. Maybe I somehow missed your point? in his root search after the first move has been searched it print the UCI output with the word lowerbound, then after the...
by elcabesa
Sun Feb 17, 2019 4:43 pm
Forum: Computer Chess Club: Programming and Technical Discussions
Topic: On failing high and finding mates
Replies: 14
Views: 2249

Re: On failing high and finding mates

in your search futility pruning and history pruning ends with a break; instead of continue;. this is one of the reason why some moves are skipped because sometimes in the search all the non capture moves are skipped even if they are good or check.
by elcabesa
Sat Feb 16, 2019 9:55 pm
Forum: Computer Chess Club: Programming and Technical Discussions
Topic: On failing high and finding mates
Replies: 14
Views: 2249

Re: On failing high and finding mates

another strange thing is that your movesort give sometimes some duplicates moves, if you look at stndard output you can see that a lot of times in startpos your engine search 22, 23 moves in root. This is because for killers you don't check that a move shall not be equal to a previous move , for exa...