Crate simplebench_runtime 

SimpleBench Runtime - Core library for the SimpleBench microbenchmarking framework.

This crate provides the runtime components for SimpleBench:

• Benchmark registration via the SimpleBench struct and inventory crate
• Timing and measurement with warmup phases
• Statistical analysis of benchmark results
• Baseline storage and regression detection

Usage

This crate is typically used alongside simplebench-macros which provides the #[bench] attribute for easy benchmark registration:

use simplebench_macros::bench;

// Simple benchmark - measures single function calls
#[bench]
fn my_benchmark() {
    // code to benchmark
}

// Setup runs once, benchmark receives reference
#[bench(setup = create_data)]
fn benchmark_with_setup(data: &Data) {
    process(data);
}

// Setup runs before each sample - for mutations/consumption
#[bench(setup_each = || vec![3, 1, 4, 1, 5])]
fn bench_sort(mut data: Vec<i32>) {
    data.sort();
}

The cargo simplebench CLI tool handles compilation and execution of benchmarks.

Re-exports

pub use inventory;

pub use baseline::*;

pub use changepoint::*;

pub use config::*;

pub use cpu_analysis::*;

pub use cpu_monitor::*;

pub use measurement::*;

pub use output::*;

pub use progress::*;

pub use statistics::*;

Modules

baseline

changepoint

Bayesian Online Change Point Detection

config

cpu_analysis

CPU analysis for detecting thermal throttling, frequency variance, and cold starts

cpu_monitor

CPU monitoring for Linux systems

measurement

output

progress

Progress reporting for benchmark execution.

statistics

Statistical functions for baseline comparison and regression detection

Structs

BenchResult

Complete result of a benchmark run.

BenchmarkInfo

Benchmark metadata for JSON listing.

Comparison

Comparison between current benchmark run and baseline.

Percentiles

Percentile statistics for a benchmark run.

SimpleBench

A registered benchmark function.

Statistics

Comprehensive statistics for a benchmark run.

Functions

calculate_statistics

Calculate comprehensive statistics from raw timing samples

list_benchmarks_json

List all registered benchmarks as JSON to stdout

run_and_stream_benchmarks

Run all benchmarks with configuration and stream results

run_single_benchmark_json

Run a single benchmark and output JSON result to stdout