How to Use Worker Threads in Node.js for CPU-Heavy Work

How to Use Worker Threads in Node.js for CPU-Heavy Work

Worker threads let you run CPU-heavy code off the main thread, keeping the event loop free. Here is how to use them.

What Worker Threads Are

Node.js's worker_threads module lets you run JavaScript on separate threads. This is for CPU-heavy work that would block the main thread's event loop.

When to Use Worker Threads

For heavy computation: image processing, cryptographic computation, large data processing, machine learning inference. Anything that takes more than a few milliseconds of synchronous CPU time should be offloaded.

How to Create a Worker

Import { Worker } from 'worker_threads'. Create a new Worker with a file path. The worker runs the file on a separate thread. Communicate via messages using postMessage and on('message').

Passing Data

Workers can receive data via postMessage and send results back. Large data is copied by default, but you can use SharedArrayBuffer for zero-copy sharing between the main thread and workers.

Error Handling

Workers can throw errors. Listen for the 'error' event on the worker to catch them. An uncaught error in a worker crashes the worker but not the main thread.

When Not to Use Worker Threads

For I/O-bound work, async APIs are better. Workers are for CPU-heavy computation. Using workers for I/O adds overhead without benefit, since async I/O already keeps the event loop free.

The Takeaway

Use worker_threads for heavy CPU computation: create a Worker with a file, communicate via postMessage, handle errors with the 'error' event, and use SharedArrayBuffer for zero-copy large data. Do not use them for I/O-bound work.