Comments on: Investigating JavaScript Array Iteration Performance http://benhollis.net/blog/2009/12/13/investigating-javascript-array-iteration-performance/ News about BenHollis.net and articles about Ben's interests Fri, 03 Sep 2010 02:00:35 +0000 hourly 1 http://wordpress.org/?v=3.0.1 By: Yet Another JavaScript Blog » Google Summer of Code 2010, Final Update http://benhollis.net/blog/2009/12/13/investigating-javascript-array-iteration-performance/comment-page-1/#comment-9749 Yet Another JavaScript Blog » Google Summer of Code 2010, Final Update Mon, 16 Aug 2010 20:23:30 +0000 http://brh.numbera.com/blog/?p=390#comment-9749 [...] and augmenting Array.prototype for compatibility with Internet Explorer. However, these methods are often slower than using a regular for loop, especially in IE, and so it would be good to swap them [...] [...] and augmenting Array.prototype for compatibility with Internet Explorer. However, these methods are often slower than using a regular for loop, especially in IE, and so it would be good to swap them [...]

]]> By: Diego Perini http://benhollis.net/blog/2009/12/13/investigating-javascript-array-iteration-performance/comment-page-1/#comment-9688 Diego Perini Wed, 23 Jun 2010 15:08:27 +0000 http://brh.numbera.com/blog/?p=390#comment-9688 Nice and extensive review of iterators speed and available alternatives. As you pointed out the slowdown is due to calling a function for each iteration. This seems to be one of the most expensive operation for a browser since it has to save the current stack/scope and retrieve it back on each function return. That's pretty easy to understand. And since loops are used for everything from animation to DOM manipulation using the best alternative will give us a generic overall boost in our applications. The gain is so big that cannot be left out. John David Dalton (of FuesJS) have setup an interesting speed demo related to this: http://dl.dropbox.com/u/513327/loop_speed/index.html have a look and see what we have been missing :) Nice and extensive review of iterators speed and available alternatives.

As you pointed out the slowdown is due to calling a function for each iteration. This seems to be one of the most expensive operation for a browser since it has to save the current stack/scope and retrieve it back on each function return.

That’s pretty easy to understand. And since loops are used for everything from animation to DOM manipulation using the best alternative will give us a generic overall boost in our applications. The gain is so big that cannot be left out.

John David Dalton (of FuesJS) have setup an interesting speed demo related to this:

http://dl.dropbox.com/u/513327/loop_speed/index.html

have a look and see what we have been missing :)

]]> By: Colin http://benhollis.net/blog/2009/12/13/investigating-javascript-array-iteration-performance/comment-page-1/#comment-9481 Colin Fri, 19 Mar 2010 23:16:39 +0000 http://brh.numbera.com/blog/?p=390#comment-9481 Good article, exactly what I was looking for after I ran into the problem with unhoisted for loops in IE. My problem was much more severe, though. It was taking 7 seconds(!) to iterate over an 8500 item array. I guess the difference was that the items were DOM elements instead of numbers... but I don't understand why that would change the time taken to get ary.length. In any case, when I put the array length into a variable, the execution time dropped to around 60ms. Keep in mind that all these differences will be amplified for slower computers, so make sure you're being efficient on those big loops. Good article, exactly what I was looking for after I ran into the problem with unhoisted for loops in IE. My problem was much more severe, though. It was taking 7 seconds(!) to iterate over an 8500 item array. I guess the difference was that the items were DOM elements instead of numbers… but I don’t understand why that would change the time taken to get ary.length. In any case, when I put the array length into a variable, the execution time dropped to around 60ms.

Keep in mind that all these differences will be amplified for slower computers, so make sure you’re being efficient on those big loops.

]]> By: jebaird http://benhollis.net/blog/2009/12/13/investigating-javascript-array-iteration-performance/comment-page-1/#comment-9446 jebaird Fri, 05 Mar 2010 15:49:56 +0000 http://brh.numbera.com/blog/?p=390#comment-9446 Thanks for the good info. I didnt know about Underscore.js or the _.each method. I'm checking it out now. Also, I did notice a small typo, differece should be difference thanks again Thanks for the good info. I didnt know about Underscore.js or the _.each method. I’m checking it out now.

Also, I did notice a small typo, differece should be difference

thanks again

]]> By: Iain http://benhollis.net/blog/2009/12/13/investigating-javascript-array-iteration-performance/comment-page-1/#comment-9372 Iain Wed, 10 Feb 2010 18:58:06 +0000 http://brh.numbera.com/blog/?p=390#comment-9372 You forgot foreach in, which loops over values, not keys. Of course, it's just as bad as for in, in that it will loop over unwanted values. You forgot foreach in, which loops over values, not keys. Of course, it’s just as bad as for in, in that it will loop over unwanted values.

]]>