Search found 3963 matches

by sje
Tue Sep 06, 2016 10:01 pm
Forum: Computer Chess Club: Programming and Technical Discussions
Topic: Baffling multithreading scaling behavior
Replies: 41
Views: 6428

Idea: Try a parallel perft()

Idea: Try a parallel perft() It's fairly simple to write a parallel perft() which splits only at the root and where the threads have minimal interaction and where the results are known beforehand and the order of calculation is irrelevant. This should produce a near linear speed-up; if not, at least...
by sje
Sun Sep 04, 2016 9:30 pm
Forum: Computer Chess Club: Programming and Technical Discussions
Topic: Two perft(7) mismatches from work unit 571
Replies: 2
Views: 961

Re: Two perft(7) mismatches from work unit 571

Thank you for your calculations.

----

Symbolic has produced more than 56,100,000 perft(7) results so far, about 58.19% of the 96,400,068 needed.

Work units not yet started (390): 574-963
by sje
Sun Sep 04, 2016 5:00 am
Forum: Computer Chess Club: Programming and Technical Discussions
Topic: Perft(14) Weekly Status Reports for 2016
Replies: 75
Views: 17973

Perft(14) Weekly Status 2016-09-04

Perft(14) Weekly Status 2016-09-04 Symbolic has produced more than 56,100,000 perft(7) results so far, about 58.19% of the 96,400,068 needed. Day count: 760 Estimated remaining day count: 546 Estimated total day count: 1,306 Average throughput: 73,816 results/day Effective frequency: 238.06 GHz Work...
by sje
Sun Sep 04, 2016 1:17 am
Forum: Computer Chess Club: Programming and Technical Discussions
Topic: Two perft(7) mismatches from work unit 571
Replies: 2
Views: 961

Two perft(7) mismatches from work unit 571

Two (adjacent) perft(7) mismatches from work unit 571: 31355,31356c31355,31356 < rnbqkb1r/1pppppp1/p4n1p/1B6/4P3/4Q3/PPPP1PPP/RNB1K1NR b KQkq - 3 4 30 18436657731 553099731930 < rnbqkb1r/1pppppp1/p4n1p/1B6/4PP2/5Q2/PPPP2PP/RNB1K1NR b KQkq - 2 4 30 15669028771 470070863130 --- > rnbqkb1r/1pppppp1/p4n...
by sje
Sat Sep 03, 2016 7:39 pm
Forum: Computer Chess Club: Programming and Technical Discussions
Topic: One perft(7) mismatch for work unit 566
Replies: 4
Views: 1047

Re: One perft(7) mismatch for work unit 566

Thanks to everyone for their verification work. The 128 bit signature program output is again replicated.

----

Symbolic has produced more than 56,000,000 perft(7) results so far, about 58.09% of the 96,400,068 needed.

Work unit #573 is the next to be started.
by sje
Sat Sep 03, 2016 1:24 pm
Forum: Computer Chess Club: Programming and Technical Discussions
Topic: One perft(7) mismatch for work unit 566
Replies: 4
Views: 1047

One perft(7) mismatch for work unit 566

One perft(7) mismatch for work unit 566

Very close!

Code: Select all

85465c85465
< rnbqkbnr/pp2pp1p/3p4/2p3BQ/3PP3/8/PPP2PPP/RN2KBNR b KQkq - 0 4 28 53830194040 1507245433120
---
> rnbqkbnr/pp2pp1p/3p4/2p3BQ/3PP3/8/PPP2PPP/RN2KBNR b KQkq - 0 4 28 53830194113 1507245435164
by sje
Thu Sep 01, 2016 10:52 am
Forum: Computer Chess Club: Programming and Technical Discussions
Topic: One perft(7) mismatch for work unit 568
Replies: 4
Views: 1212

Re: One perft(7) mismatch for work unit 568

Sharper v0.17 by Albert Bertilsson setboard r1bqkbnr/ppppp2p/2n5/5pp1/1P6/B7/P1PPPPPP/RNQ1KBNR b KQkq - 0 4 perft 7 Nodes: 4466138745, Time: 7954 cs Thanks for your calculation. ---- Symbolic has produced more than 55,900,000 perft(7) results so far, about 57.99% of the 96,400,068 needed. The next ...
by sje
Thu Sep 01, 2016 9:51 am
Forum: Computer Chess Club: Programming and Technical Discussions
Topic: One perft(7) mismatch for work unit 568
Replies: 4
Views: 1212

Re: One perft(7) mismatch for work unit 568

Thanks. The 128 bit signature output is once again verified.
by sje
Thu Sep 01, 2016 6:47 am
Forum: Computer Chess Club: Programming and Technical Discussions
Topic: One perft(7) mismatch for work unit 568
Replies: 4
Views: 1212

One perft(7) mismatch for work unit 568

One perft(7) mismatch for work unit 568

Code: Select all

80801c80801
< r1bqkbnr/ppppp2p/2n5/5pp1/1P6/B7/P1PPPPPP/RNQ1KBNR b KQkq - 0 4 30 4466139808 133984194240
---
> r1bqkbnr/ppppp2p/2n5/5pp1/1P6/B7/P1PPPPPP/RNQ1KBNR b KQkq - 0 4 30 4466138745 133984162350
by sje
Thu Sep 01, 2016 1:13 am
Forum: Computer Chess Club: Programming and Technical Discussions
Topic: Perft(14) Weekly Status Reports for 2016
Replies: 75
Views: 17973

Re: Castling does not reset the half move counter

I suppose that detecting and recording those metrics could help debug a chess program. But otherwise they won't help much with move selection, so if they're included in a program, then they should be conditionally compiled so as to not slow the program during regular play. ---- For those wanting to ...