node-worker-threads-pool

node-worker-threads-pool

Simple worker threads pool using Node's worker_threads module. Compatible with ES6+ Promise, Async/Await.

With this library, you can:

• Use StaticPool to create a threads pool with a task from worker file or from task function provided to make use of multi-core processor.
• Use DynamicPool to create a threads pool with different tasks provided each call. Thus you can get more flexibility than StaticPool and make use of multi-core processor.
• Set timeout for the task, thus you won't get a task that running forever.

Notification

1. This module can only run in Node.js.
2. Since Node's worker_threads module is still in stage of Experimental, this module can be accessed only if the --experimental-worker flag is added., if node.js version is above 11.7.0, worker api is exposed by default.

Installation

npm install node-worker-threads-pool --save

API

Class: StaticPool

Instance of StaticPool is a threads pool with static task provided.

new StaticPool(opt)

• opt <Object>
  • size <number> Number of workers in this pool.
  • task <string | function> Static task to do. It can be a absolute path of worker file or a function. Notice: If task is a function, you can not use closure in it! If you do want to use external data in the function, you can use workerData to pass some cloneable data.
  • workerData <any> Cloneable data you want to access in task function.
  • shareEnv <boolean> Set true to enable SHARE_ENV for every threads in pool.

staticPool.exec(param[, timeout])

• param <any> The param your worker script or task function need.
• timeout <number> Timeout in milisecond for limiting the execution time. When timeout, the function will throw a TimeoutError, use isTimeoutError function to detect it.
• Returns: <Promise>

Choose an idle worker in the pool to execute your heavy task with the param you provided. The Promise is resolved with the result.

staticPool.destroy()

Call worker.terminate() for every worker in the pool and release them.

Example (with worker file)

In the worker.js :

// Access the workerData by requiring it.
const { parentPort, workerData } = require("worker_threads");

// Something you shouldn"t run in main thread
// since it will block.
function fib(n) {
  if (n < 2) {
    return n;
  }
  return fib(n - 1) + fib(n - 2);
}

// Main thread will pass the data you need
// through this event listener.
parentPort.on("message", (param) => {
  if (typeof param !== "number") {
    throw new Error("param must be a number.");
  }
  const result = fib(param);

  // Access the workerData.
  console.log("workerData is", workerData);

  // return the result to main thread.
  parentPort.postMessage(result);
});

In the main.js :

const { StaticPool } = require("node-worker-threads-pool");

const filePath = "absolute/path/to/your/worker/script";

const pool = new StaticPool({
  size: 4,
  task: filePath,
  workerData: "workerData!",
});

for (let i = 0; i < 20; i++) {
  (async () => {
    const num = 40 + Math.trunc(10 * Math.random());

    // This will choose one idle worker in the pool
    // to execute your heavy task without blocking
    // the main thread!
    const res = await pool.exec(num);

    console.log(`Fibonacci(${num}) result:`, res);
  })();
}

Example (with task function)

In the main.js :

const { StaticPool } = require("node-worker-threads-pool");

const pool = new StaticPool({
  size: 4,
  task: function(n) {
    const num = this.workerData.num;
    for (let i = 0; i < num; i++) {
      n += i;
    }
    return n;
  },
  workerData: {
    num: 1 << 30,
  },
});

for (let i = 0; i < 20; i++) {
  (async () => {
    const res = await pool.exec(i);
    console.log(`result${i}:`, res);
  })();
}

Class: DynamicPool

Instance of DynamicPool is a threads pool executes different task functions provided every call.

new DynamicPool(size[, opt])

• size <number> Number of workers in this pool.
• opt
  • shareEnv <boolean> Set true to enable SHARE_ENV for every threads in pool.

dynamicPool.exec(opt)

• opt
  • task <function> Function as a task to do. Notice: You can not use closure in task function! If you do want to use external data in the function, you can use workerData to pass some cloneable data.
  • workerData <any> Cloneable data you want to access in task function.
  • timeout <number> Timeout in milisecond for limiting the execution time. When timeout, the function will throw a TimeoutError, use isTimeoutError function to detect it.
• Returns: <Promise>

Choose one idle worker in the pool to execute your task function. The Promise is resolved with the result your task returned.

dynamicPool.destroy()

Call worker.terminate() for every worker in the pool and release them.

Example

In the main.js :

const { DynamicPool } = require("node-worker-threads-pool");

const pool = new DynamicPool(4);

function task1() {
  // something heavy.
}

function task2() {
  // something heavy too.
}

// execute task1
(async () => {
  const res = await pool.exec({
    task: task1,
    workerData: {
      // some data
    },
  });
  console.log(res);
})();

// execute task2
(async () => {
  const res = await pool.exec({
    task: task2,
    workerData: {
      // some data
    },
  });
  console.log(res);
})();

function: isTimeoutError

Detect if a thrown error is TimeoutError.

isTimeoutError(err)

• err <Error> The error you want to detect.
• Returns <boolean> true if the error is a TimeoutError.

Example

const { isTimeoutError } = require("node-worker-threads-pool");

// create pool.
...

// static pool exec with timeout.
const timeout = 1000;
try {
  const res = await staticPool.exec(param, timeout);
} catch (err) {
  if (isTimeoutError(err)) {
    // deal with timeout.
  } else {
    // deal with other errors.
  }
}

// dynamic pool exec with timeout.
const timeout = 1000;
try {
  const res = await dynamicPool.exec({
    task() {
      // your task.
    },
    timeout
  });
} catch (err) {
  if (isTimeoutError(err)) {
    // deal with timeout.
  } else {
    // deal with other errors.
  }
}