Benchy

Benchy is a Rust crate for benchmarking long-running tasks. Unlike other benchmarking libraries such as Criterion, which are optimized for high-frequency, nanosecond-level performance, Benchy is designed for tasks that take a significant amount of time to execute. It provides a flexible and customizable environment, allowing you to set custom iteration counts and even measure memory usage metrics.

Features

• Custom Iteration Counts: Run your benchmarks as many or as few times as you need.

• Memory Usage Metrics: Get insights into how much memory your code is using.

• Environment Variable Configuration: Customize your benchmarks on the fly using environment variables.

• JSON Output: Easily export your benchmark results to JSON for further analysis.

Installation

cargo add benchy

Quick Start

benches/bench.rs:

use benchy::{benchmark, BenchmarkRun};

#[benchmark]
fn fibonacci_single(b: &mut BenchmarkRun) {
    let mut x = 0;
    let mut y = 1;
    b.run(|| {
        for _ in 0..1_000_000 {
            let temp = x;
            x = y;
            y = temp + y;
        }
    });
}

#[benchmark("Fibonacci", [
    ("1 million iterations", 1_000_000),
    ("2 million iterations", 2_000_000),
])]
fn fibonacci_parametrized(b: &mut BenchmarkRun, iterations: usize) {
    let mut x = 0;
    let mut y = 1;
    b.run(|| {
        for _ in 0..iterations {
            let temp = x;
            x = y;
            y = temp + y;
        }
    });
}

benchy::main!(fibonacci_single, fibonacci_parametrized);

Cargo.toml:

[[bench]]
name = "bench"
harness = false

For more advanced usage, check the zk-bench repository that utilizes this crate, or refer to the documentation.

Environment variables

• BENCHY_QUICK (default: false) - if true, runs only the first parameter of each benchmark.
• BENCHY_OUTPUT_DIR (default: None) - directory to output the JSON benchmark results to.
• BENCHY_MAX_DEFAULT_ITERATIONS_DURATION (default: 10s) - the maximum total duration for the default (10) iterations of a single benchmark.