A UCI chess engine written in Rust. Please take a look at the UCI Documentation for usage instructions.

Binaries for Windows, Mac OS and Linux are available on the releases page.

• Alpha-beta search with iterative deepening and aspiration windows
• Lazy SMP multi-threaded search (configurable via UCI Threads option)
• Transposition table, killer moves, history heuristic, countermove heuristic
• Null move pruning, late move reductions, late move pruning, reverse futility pruning
• Singular extensions, probcut, multi-cut, SEE pruning
• Quiescence search with static exchange evaluation
• SPSA-tuned evaluation and search parameters

1. Build the engine:

The engine can be about 50% faster if compiled on the machine on which it will be run in order to take advantage of cpu-specific instructions. To compile locally, it requires that Rust be installed.

Linux/macOS:

RUSTFLAGS="-C target-cpu=native" RUST_MIN_STACK=4097152 cargo build --release

Windows (PowerShell):

$env:RUSTFLAGS="-C target-cpu=native"; $env:RUST_MIN_STACK=4097152; cargo build --release

Windows (cmd.exe):

set RUSTFLAGS=-C target-cpu=native
set RUST_MIN_STACK=4097152
cargo build --release

Note: .cargo/config.toml sets -C target-cpu=native for local builds by default to maximize performance. This makes the binaries non-portable across different CPUs. The release workflow overrides this to keep published binaries portable. 2. Store the executable in engines directory:

mkdir -p engines/v020-my-feature
cp target/release/rusty-rival engines/v020-my-feature/
git tag v020-my-feature

For engine vs engine matches with Elo tracking, see the separate chess-compete repository.

The scripts/benchmark.py script measures search performance (nodes per second) across a diverse set of test positions.

# Benchmark a single engine at depth 12
./scripts/benchmark.py v020-my-feature --depth 12

# Compare multiple engines
./scripts/benchmark.py v019-baseline v020-my-feature --depth 12

# Quick test with lower depth
./scripts/benchmark.py v020-my-feature --depth 8

Output shows NPS for each position and overall comparison when benchmarking multiple engines.

Use perft, to determine the total number of positions encountered while playing through every move and every response to a certain depth.

position fen 8/2p5/3p4/KP5r/1R3p1k/8/4P1P1/8 w - - 0 1
READY
go perft 7
a5a4: 14,139,786  68,974,565 nps
a5a6: 16,022,983  74,845,580 nps
b4b1: 19,481,757  77,087,773 nps
b4b2: 12,755,330  77,515,349 nps
b4b3: 15,482,610  77,882,466 nps
b4f4: 3,069,955  78,064,051 nps
b4e4: 14,187,097  78,047,184 nps
b4d4: 15,996,777  78,099,996 nps
b4c4: 17,400,108  78,280,391 nps
b4a4: 11,996,400  78,291,255 nps
g2g3: 4,190,119  78,313,269 nps
g2g4: 13,629,805  78,198,877 nps
e2e3: 11,427,551  78,095,804 nps
e2e4: 8,853,383  78,005,965 nps
Time elapsed in perft is: 2.290587309s
178633661 nodes 78005965.50218341 nps