http://magnetiq.com/2008/10/19/timing-code-accurately/ Ates Goral's Personal Playground and Project Repository Wed, 01 Apr 2009 23:28:36 -0500 http://wordpress.org/?v=2.8.4 hourly 1 http://magnetiq.com/2008/10/19/timing-code-accurately/comment-page-1/#comment-38789 Ates Goral Mon, 20 Oct 2008 18:34:02 +0000 http://magnetiq.com/?p=94#comment-38789 @cfq: Thanks for that insight! I used JavaScript just for illustration purposes; in order to avoid using pseudo-code that couldn't be properly highlighted by the automatic syntax highlighter that I'm currently using. I've successfully used this method to do benchmarking in C++. The benchmark program would first set the process priority to high so that my measurement wouldn't get affected by Windows doing stuff in the background. @cfq: Thanks for that insight! I used JavaScript just for illustration purposes; in order to avoid using pseudo-code that couldn’t be properly highlighted by the automatic syntax highlighter that I’m currently using. I’ve successfully used this method to do benchmarking in C++. The benchmark program would first set the process priority to high so that my measurement wouldn’t get affected by Windows doing stuff in the background.

]]> http://magnetiq.com/2008/10/19/timing-code-accurately/comment-page-1/#comment-38788 cfq Mon, 20 Oct 2008 12:47:24 +0000 http://magnetiq.com/?p=94#comment-38788 Javascript execution time in browsers isn't bounded only by the JS engine though. Render engine may interrupt the process if myFunction is doing any DOM modifications, which would stop the JS timer and the resultant timing values may not be accurate. If myFunction is not doing any DOM operations I'd recommend testing function optimizations under rhino, so you can be sure there are no interruptions. (I realize you may already be doing that, sorry if I'm repeating something you already know) Javascript execution time in browsers isn’t bounded only by the JS engine though. Render engine may interrupt the process if myFunction is doing any DOM modifications, which would stop the JS timer and the resultant timing values may not be accurate. If myFunction is not doing any DOM operations I’d recommend testing function optimizations under rhino, so you can be sure there are no interruptions. (I realize you may already be doing that, sorry if I’m repeating something you already know)

]]>