ruchess

rust chess

docs

features

• highly compact bitboard representations
• zobrist hashing
• magic attack tables
• move generation & validation
• outcome evaluation (checkmate, stalemate, repetition, insufficient material)
• fen parsing
• immutable
• hella docs
• really really really fast (relatively)

i only support standard chess. it is unlikely i will support other variants.

probably check out shakmaty if you want more features

quickstart

[!CAUTION] this is library is still at version 0.0.X

it is very likely that there will be breaking api changes until version 0.1.0

if you use this library, make sure you pin the version in cargo.toml

cargo add ruchess

or in your cargo.toml

[dependencies]
ruchess = "0.0.4"

use ruchess::game::Game;
use ruchess::square;

// Play 1.e4 from the standard starting position.
let game = Game::new();
let after_e4 = game.mve(square::E2, square::E4).unwrap();
assert!(after_e4.position().board().is_occupied(square::E4));

see: docs for more details

see: examples or ruchess tui for example applications

dependencies

just lazy_static

i could probably get rid of it with some refactoring but i am lazy

roadmap

• benchmarks
  • will move to mutable api if benchmarks are very bad
• parsing
  • fen
  • pgn
  • san
  • uci
• real uci support
• finalise public api

benchmarks

simple and benchmarking using perft to get an estimation of performance

results in benchmark_results/

code in examples/perft.rs

benchmarks were ran on m1 macbook pro

at depth=4 and depth=5, we are at 7ms and 177ms respectively.

for reference, shakmaty and jordanbray/chess are at 0.8 - 1ms for depth=4 and 18.6 - 24.1ms for depth=5. we are currently an order of magnitude slower.

it is worth noting that the shakmaty and jordanbray benchmarks weren't run on my machine but still (sort of) indicative.

license

unlicense