Search found 408 matches

by vittyvirus
Sun Jan 13, 2019 5:35 am
Forum: Computer Chess Club: General Topics
Topic: Quality of these computer generated puzzles
Replies: 11
Views: 1985

Quality of these computer generated puzzles

These are extracted from World Blitz Chess Championship 2017 from a program I wrote past year during summer. https://justpaste.it/5aiig I'm not an extremely strong chess player, so I'm unsure of the worth of the puzzles that it generates, but I've tried to make them look like if they were picked by ...
by vittyvirus
Fri Aug 17, 2018 12:47 pm
Forum: Computer Chess Club: General Topics
Topic: Chesser and Yaka by Syed Fahad
Replies: 3
Views: 871

Re: Chesser and Yaka by Syed Fahad

I wanted to inform all of you Yaka is not a complete chess engine and it does not abide by any standard protocol. Besides, it's abandoned for now.
by vittyvirus
Wed Jan 11, 2017 7:32 am
Forum: Computer Chess Club: Programming and Technical Discussions
Topic: Table-driven move generation for Pawns
Replies: 5
Views: 1343

Re: Table-driven move generation for Pawns

AndrewGrant wrote:Are you trying to get these ideas into an actual engine, or just a perft calc?
For now, they're only for people who care about perft speed, or for chess engine authors who don't want to miss that 0.1% optimisation.
by vittyvirus
Wed Jan 11, 2017 6:06 am
Forum: Computer Chess Club: Programming and Technical Discussions
Topic: Table-driven move generation for Pawns
Replies: 5
Views: 1343

Re: Table-driven move generation for Pawns

AndrewGrant wrote:Are you basing the speed up off PERFT tests, or off of actual searches?
Duh, I'm talking about perft speed.
by vittyvirus
Sun Jan 08, 2017 1:58 pm
Forum: Computer Chess Club: Programming and Technical Discussions
Topic: Table-driven move generation for Pawns
Replies: 5
Views: 1343

Table-driven move generation for Pawns

Hello! I did some more work with the table driven move generation thing and managed do some more improvement. Here are the ideas: 1. Pawn attack masks are hashed per two ranks, which means you require 16 bits for two ranks. Captures are excluded because capture attack masks are sparsely populated, a...
by vittyvirus
Wed Jan 04, 2017 1:32 pm
Forum: Computer Chess Club: Programming and Technical Discussions
Topic: Fast table-driven move generation
Replies: 10
Views: 2894

Re: Fast table-driven move generation

The bad part --- How can you score and sort the generated moves in this way? Apart from perft, in normal search you need to score them. Hi Marco, Your question is certainly interesting. Let me answer your question in perspective of Stockfish. I'm using Stockfish 5 as reference, so pardon me if I sa...
by vittyvirus
Sun Jan 01, 2017 1:05 pm
Forum: Computer Chess Club: Programming and Technical Discussions
Topic: Fast table-driven move generation
Replies: 10
Views: 2894

Fast table-driven move generation

Hi! Gerd Isenberg had the of "magically" hashing the attack masks to a precomputed movelist (see this ). Here's the source code he provided: moveTarget = KingAttacks[sq] & ~(ownPieces | attackedSquares); idx = (moveTarget * kingMagic[sq]) >> kingShift[sq]); movelists = kingMoveLists[sq];...
by vittyvirus
Wed Nov 30, 2016 4:10 am
Forum: Computer Chess Club: Programming and Technical Discussions
Topic: Subtraction based Attack Mask Generation
Replies: 6
Views: 2100

Re: Subtraction based Attack Mask Generation

hmm ... is the | 1 really necessary? Lower (PrevSquares[sq]) should always intersect with outer (RankOuterSquares[sq]). Then argument of bsr is always not empty and bsr defined ... assert ( pMask->outer & pMask->lower); occ |= pMask->outer; int bsq = bsr64(occ & pMask->lower); Gerd Hi! I'm sorry fo...
by vittyvirus
Thu Nov 17, 2016 4:25 am
Forum: Computer Chess Club: Programming and Technical Discussions
Topic: Subtraction based Attack Mask Generation
Replies: 6
Views: 2100

Re: Subtraction based Attack Mask Generation

This is really interesting. I guess it performs well too. Maybe it should get put into the chess wiki. https://chessprogramming.wikispaces.com/SBAMG First of all, a big thanks to Gerd Isenberg for explaining it so wonderfully! Apparently I can't edit the wiki. Could someone please update the code? ...
by vittyvirus
Wed Nov 16, 2016 3:58 pm
Forum: Computer Chess Club: Programming and Technical Discussions
Topic: Steven Edwards question
Replies: 6
Views: 2615

Re: Steven Edwards question

Did he compute perft 14?