ThreadSort

A parallel sorting implementation that efficiently processes large datasets through concurrent execution.

Super fast array sort method.

ThreadSort is a super fast array sort method that uses the power of Web Workers to sort large arrays in parallel.

Features

• ⚡ Fast (see benchmark) +(~)50% faster than native Array.sort
• 🐦 Lightweight (~18kb)
• 🩵 Written in TypeScript
• 📦 Zero dependencies
• 🌐 Works in Node.js and browsers

Usage

import { threadSort } from 'threadsort';

const array = [3, 1, 2];

const sortedArray = await threadSort(array);

Benchmark

npm run benchmark (see src/Benchmark.ts)

Sorting arrays of sizes = [1000000, 5000000, 10000000]

threadsort (worker_threads) - size 1000000: 225.787ms
mergeSort (native) - size 1000000: 617.442ms
Array.sort (native) - size 1000000: 227.572ms

threadsort (worker_threads) - size 5000000: 506.851ms
mergeSort (native) - size 5000000: 2.960s
Array.sort (native) - size 5000000: 1.212s

threadsort (worker_threads) - size 10000000: 1.055s
mergeSort (native) - size 10000000: 7.488s
Array.sort (native) - size 10000000: 2.876s

threadsort (worker_threads) - size 100000000: 29.085s
mergeSort (native) - size 100000000: - (to slow 💀)
Array.sort (native) - size 100000000: 59.238s

Key Features

• Distributes sorting workload across multiple threads
• Optimizes performance for large data collections
• Implements merge-sort algorithm in parallel
• Thread-safe execution

Performance

Offers improved sorting speed on multi-core systems, particularly for:

• Large arrays
• Complex data structures
• Memory-intensive sorting operations

Usage Considerations

• Best for datasets larger than 100,000 elements
• Requires proper thread management
• Memory overhead due to concurrent operations
• May not be optimal for small datasets due to threading overhead

Notes

Performance gains depend on:

• Available CPU cores
• Data size
• Memory constraints
• System load

MIT - Cristiancast