Chrome: timeouts/interval suspended in background tabs?


I was testing the accuracy of setTimeout using this test. Now I noticed that (as expected) setTimeout is not very accurate but for most appliances not dramatically inaccurate. Now if I run the test in in Chrome and let it run in a background tab (so, switching to another tab and browse on there), returning to the test and inspecting te results (if the test finished) they are dramatically changed. It looks like the timeouts have been running a lot slower. Tested in FF4 or IE9 this didn't occur.

So it looks like Chrome suspends or at least slows down javascript execution in a tab that has no focus. Couldn't find much on the net on the subject. It would mean that we can't run background tasks, like for example checking periodically on a server using XHR calls and setInterval (I suspect to see the same behavior for setInterval, will write a test if time is with me).

Has anyone encountered this? Would there be there a workaround for this suspension/slowing down? Would you call it a bug and should I file it as such?

I recently asked about this and it is behaviour by design. When a tab is inactive, only at a maximum of once per second the function is called. Here is the code change.

Perhaps this will help: How can I make setInterval also work when a tab is inactive in Chrome?

TL;DR: use Web Workers.


There is a solution to use Web Workers, because they run in separate process and are not slowed down

I've written a tiny script that can be used without changes to your code - it simply overrides functions setTimeout, clearTimeout, setInterval, clearInterval

Just include it before all your code

http://github.com/turuslan/HackTimer


Playing an ~empty sound forces the browser to retain the performance - I discovered it after reading this comment: How to make JavaScript run at normal speed in Chrome even when tab is not active?

I need unlimited performance on-demand for a browser game that uses WebSockets, so I know from experience that using WebSockets doesn't ensure unlimited performance, but from tests, playing an audio file seems to ensure it

Here's 2 empty audio loops I created for this purpose, you can use them freely, commercially: http://adventure.land/sounds/loops/empty_loop_for_js_performance.ogg http://adventure.land/sounds/loops/empty_loop_for_js_performance.wav

(They include -58db noise, -60db doesn't work)

I play them, on user-demand, with Howler.js: https://github.com/goldfire/howler.js

function performance_trick()
{
    if(sounds.empty) return sounds.empty.play();
    sounds.empty = new Howl({
        src: ['/sounds/loops/empty_loop_for_js_performance.ogg','/sounds/loops/empty_loop_for_js_performance.wav'],
        volume:0.5,
        autoplay: true, loop: true,
    });
}

It's sad that there is no built-in method to turn full javascript performance on/off by default, yet, crypto miners can hijack all your computing threads using Web Workers without any prompt :|


I have released worker-interval npm package which setInterval and clearInterval implementation with using Web-Workers to keep up and running on inactive tabs for Chrome, Firefox and IE.

Most of the modern browsers (Chrome, Firefox and IE), intervals (window timers) are clamped to fire no more often than once per second in inactive tabs.

You can find more information on

https://developer.mozilla.org/en-US/docs/Web/API/WindowOrWorkerGlobalScope/setInterval

https://developer.mozilla.org/en-US/docs/Web/API/Web_Workers_API/Using_web_workers#Timeouts_and_intervals


I updated my jQuery core to 1.9.1, and it solved the Interval discrepancy in inactive tabs. I would try that first, then look into other code override options.


here is my solution which gets the current millisecond, and compares it to the millisecond that the function was created. for interval, it will update the millisecond when it runs the function. you can also grab the interval/timeout by an id.

<script>

var nowMillisTimeout = [];
var timeout = [];
var nowMillisInterval = [];
var interval = [];

function getCurrentMillis(){
    var d = new Date();
    var now = d.getHours()+""+d.getMinutes()+""+d.getSeconds()+""+d.getMilliseconds();
    return now;
}

function setAccurateTimeout(callbackfunction, millis, id=0){
    nowMillisTimeout[id] = getCurrentMillis();
    timeout[id] = setInterval(function(){ var now = getCurrentMillis(); if(now >= (+nowMillisTimeout[id] + +millis)){callbackfunction.call(); clearInterval(timeout[id]);} }, 10);
}

function setAccurateInterval(callbackfunction, millis, id=0){
    nowMillisInterval[id] = getCurrentMillis();
    interval[id] = setInterval(function(){ var now = getCurrentMillis(); if(now >= (+nowMillisInterval[id] + +millis)){callbackfunction.call(); nowMillisInterval[id] = getCurrentMillis();} }, 10);
}

//usage
setAccurateTimeout(function(){ console.log('test timeout'); }, 1000, 1);

setAccurateInterval(function(){ console.log('test interval'); }, 1000, 1);

</script>