https://github.com/tim-hub/pambdajs
Parallelized Lambda, the wrapper to help run lambda/anonymous function in parallel. 🚀
https://github.com/tim-hub/pambdajs
lambda multi-process nodejs parallel-programming
Last synced: 11 months ago
JSON representation
Parallelized Lambda, the wrapper to help run lambda/anonymous function in parallel. 🚀
- Host: GitHub
- URL: https://github.com/tim-hub/pambdajs
- Owner: tim-hub
- License: mit
- Created: 2020-08-23T09:10:50.000Z (almost 6 years ago)
- Default Branch: master
- Last Pushed: 2022-06-10T00:49:00.000Z (almost 4 years ago)
- Last Synced: 2025-05-19T19:33:30.557Z (about 1 year ago)
- Topics: lambda, multi-process, nodejs, parallel-programming
- Language: TypeScript
- Homepage: https://www.npmjs.com/package/pambdajs
- Size: 468 KB
- Stars: 28
- Watchers: 1
- Forks: 0
- Open Issues: 18
-
Metadata Files:
- Readme: README.md
- Contributing: CONTRIBUTING.md
- License: LICENSE
- Code of conduct: code-of-conduct.md
Awesome Lists containing this project
README
# PambdaJS - Multi-Process Has Never Been This Easy
[](https://www.npmjs.com/package/pambdajs)  [](https://codecov.io/gh/tim-hub/pambdajs)
> Parallelized Lambda, a wrapper to help run lambda/anonymous function in parallel easily.
PambdaJS is to orchestrate different child process to finish one heavy iteration work.
- make multi process programming easily
- gain better performance through empowering multi core CPU
**Start to use PambdaJS**
```javascript
import pambda from 'pambdajs';
const p = await pambda.init(5);
data = [1,2,3];
const heavyWork = (x) => {
return x+x;
}
// original single process way
const singleProcess = () => {
data.map(heavyWork);
}
// PambdaJS multi process way
const pambdaProcess = async () => {
await p.map(heavyWork, data);
}
```
## Features
Run your own anonymous functions in Parallel.
- map
- filter
- reduce
## Limit
The lambda function itself has to be a pure function,
and it has to use only node build-in packages.
for example:
```javascript
// good
const heavyWork = (x) => {
return x+x;
}
```
```javascript
// bad
const sumItSelf = (x) => {
return x + x;
}
const heavyWork = (x) => {
return sumItSelf(x)
}
```
## Performance
- PambdaJS can save up to 55% of processing time. (Diagram ↓)
As you can see the above.
- The best case to repeat summing 10k number 100k times,
is to spawn 5 child process, and **it saves more than half of time than single process.**
- Besides, for simple work, multi process does not help at all.
So make sure usig PambdaJS for heavy work only. ([more diagrams](https://pambdajs-performance.vercel.app/))
## FAQ
- Why not use child process directly?
> Yeah, why not. child_process is the native module of Node, if you seek better flexibility, use child_process.
However, if you want to make your life easier, trust me use PambdaJS,
you do not need to worry about the message between child and parent process anymore.
- Does it run on browser as well?
> No, PambdaJS is a tool for Node environment, not for browser,
because browser does not support multi-process,
but you can try [web worker](https://developer.mozilla.org/en-US/docs/Web/API/Web_Workers_API/Using_web_workers) which is multi-thread solution.
>or Gpu.js which is using GPU instead of CPU for getting better performance.
## Reference
- [What is Lambda Function](https://stackoverflow.com/questions/16501/what-is-a-lambda-function)
- Other solutions for better performance
- [Child Process](https://nodejs.org/api/child_process.html)
- [GPU.js](https://gpu.rocks/)
- [Web Worker](https://developer.mozilla.org/en-US/docs/Web/API/Web_Workers_API/Using_web_workers) for multi threading. **We all love MDN!!!**
- Based on Web Worker/ Multi Thread
- [ncpu](https://github.com/zy445566/ncpu)
- microsoft [napajs](https://github.com/microsoft/napajs)
- Relative Repos
- [PambdaJS Playground](https://github.com/tim-hub/pambdajs-playground) - the place to use pambdajs to run some simple sample work
- [PambdaJS Performance Analyse](https://github.com/tim-hub/pambdajs-performance) - using Chart.js to generate the diagrams
- [Code of conduct](./code-of-conduct.md)
- [Contributing](./CONTRIBUTING.md)