Rich-Bench for Node.js

A Tiny Benchmarking Tool for Node.js (this is a port of the original python tool rich-bench by Anthony Shaw)

🚀 Overview

rich-bench is a simple CLI tool that benchmarks JavaScript functions and provides a richly formatted performance comparison.
It measures execution time, computes performance gains, and generates insightful statistics.

📦 Installation

1️⃣ Install Locally for Development

git clone https://github.com/manekinekko/rich-bench-node.git
cd rich-bench
npm install
npm link  # Enable CLI usage globally

🛠 Usage

Basic Benchmark Execution

rich-bench benchmarks/

Run a Specific Benchmark

rich-bench benchmarks/ --benchmark encoding_format

Show Percentage Instead of Multipliers

rich-bench benchmarks/ --percentage

This will show the performance gains as percentages instead of multipliers.

┌────────────┬──────────┬───────────┬───────────┬─────────────────┬─────────────────────┬────────────────────┐
│ Benchmark  │ Min (ms) │  Max (ms) │ Mean (ms) │         Min (+) │             Max (+) │           Mean (+) │
├────────-───┼──────────┼───────────┼───────────┼─────────────────┼─────────────────────┼────────────────────┤
│    #1      │   73.376 │ 19666.000 │  9844.744 │  44.430 (39.4%) │    8351.000 (57.5%) │   4213.139 (57.2%) │
└────────────┴──────────┴───────────┴───────────┴─────────────────┴─────────────────────┴────────────────────┘

Repeat Benchmark Execution

rich-bench benchmarks/ --repeat 100

Run benchmarks N times

rich-bench benchmarks/ --times 100

Note

If the benchmarked functions returns a value, it will be used as the benchmark value and the function will be called 1 time, but still repeated --repeat times. If the benchmarked functions does not return a value, the execution time will be used as the benchmark value. --repeatand --times will be used.

📊 Example Output

┌───────────┬──────────┬───────────┬───────────┬───────────────┬──────────────────┬─────────────────┐
│ Benchmark │ Min (ms) │  Max (ms) │ Mean (ms) │       Min (+) │          Max (+) │        Mean (+) │
├───────────┼──────────┼───────────┼───────────┼───────────────┼──────────────────┼─────────────────┤
│    #1     │   62.825 │ 19651.000 │  9861.410 │ 51.549 (1.2x) │  8351.000 (2.4x) │ 4217.198 (2.3x) │
├───────────┼──────────┼───────────┼───────────┼───────────────┼──────────────────┼─────────────────┤
│    #2     │   47.765 │ 19633.000 │  9852.893 │ 60.384 (0.8x) │ 19611.000 (1.0x) │ 9841.065 (1.0x) │
├───────────┼──────────┼───────────┼───────────┼───────────────┼──────────────────┼─────────────────┤
│    #3     │   44.447 │  8351.000 │  4214.801 │ 45.261 (1.0x) │ 19641.000 (0.4x) │ 9849.066 (0.4x) │
└───────────┴──────────┴───────────┴───────────┴───────────────┴──────────────────┴─────────────────┘

📑 Writing a Benchmark

Create a bench_*.js file inside benchmarks/ and export an array of benchmark functions:

async function encoding_format_float(inputSize) {
    const emb = await openai.embeddings.create({
        model: "text-embedding-ada-002",
        input: generateRandomString(inputSize),
        encoding_format: "float"
    });
    return new Blob([emb.data[0].embedding]).size; // return bytes as the benchmark
}

async function encoding_format_base64(inputSize) {
    const emb = await openai.embeddings.create({
        model: "text-embedding-ada-002",
        input: generateRandomString(inputSize),
        encoding_format: "base64"
    });
    return new Blob([emb.data[0].embedding]).size; // return bytes as the benchmark
}

module.exports.__benchmarks__ = [
    [
        encoding_format_float,
        encoding_format_base64,
        "bench #1"
    ]
];

Then run:

npx rich-bench benchmarks/

Features

✔ Compares two functions side by side
✔ Displays min, max, and mean execution times
✔ Highlights performance gains with colors

Credits

Shoutout to Anthony Shaw for creating the original rich-bench tool in Python.

License

manekinekko/rich-bench-node

ErrorLooks like something went wrong!

About

Resources

License

Uh oh!

Stars

Watchers

Forks

Releases

Packages 0

Uh oh!

Languages

Error
Looks like something went wrong!

Packages