Comments on: Optimize only when you have a reason to. http://marvintam.com/2010/02/optimize-only-when-you-have-a-reason-to/ Sat, 04 Feb 2012 04:25:26 +0000 hourly 1 http://wordpress.org/?v=3.3.1 By: amd nvidia benchmark http://marvintam.com/2010/02/optimize-only-when-you-have-a-reason-to/comment-page-1/#comment-438 amd nvidia benchmark Sat, 04 Feb 2012 04:25:26 +0000 http://marvintam.com/?p=413#comment-438 <strong>amd nvidia benchmark...</strong> [...]Marvin Tam | Optimize only when you have a reason to.[...]... amd nvidia benchmark…

[...]Marvin Tam | Optimize only when you have a reason to.[...]…

]]> By: Marvin http://marvintam.com/2010/02/optimize-only-when-you-have-a-reason-to/comment-page-1/#comment-101 Marvin Wed, 02 Jun 2010 04:20:04 +0000 http://marvintam.com/?p=413#comment-101 Carlos, thanks for taking your time to correct me! I certainly won't dispute any of your findings and assertions, but I do still maintain that script parsing and loading probably isn't the slowest part of most websites; there are often bigger gains to be found by cutting down requests, optimizing caches etc. My intention here was really just to stress the importance (to not much of an audience, really :) ) of profiling before optimizing blindly. Carlos, thanks for taking your time to correct me!

I certainly won’t dispute any of your findings and assertions, but I do still maintain that script parsing and loading probably isn’t the slowest part of most websites; there are often bigger gains to be found by cutting down requests, optimizing caches etc.

My intention here was really just to stress the importance (to not much of an audience, really :) ) of profiling before optimizing blindly.

]]> By: Carlos Bueno http://marvintam.com/2010/02/optimize-only-when-you-have-a-reason-to/comment-page-1/#comment-99 Carlos Bueno Sun, 30 May 2010 15:19:55 +0000 http://marvintam.com/?p=413#comment-99 Hey -- I feel I should clear something up. The tests contain 1,000 iterations, but each data point in the chart is the *average*, not the sum. It really does take tens to hundreds of milliseconds to parse and load a typical Javascript library on a typical browser as of 2010. Parse time is on the same order of magnitude as network time, and the purpose of this test was to show that it is something you need to pay attention to. You can download the benchmark here and try it out: http://github.com/aristus/parse-n-load Hey — I feel I should clear something up. The tests contain 1,000 iterations, but each data point in the chart is the *average*, not the sum.

It really does take tens to hundreds of milliseconds to parse and load a typical Javascript library on a typical browser as of 2010. Parse time is on the same order of magnitude as network time, and the purpose of this test was to show that it is something you need to pay attention to. You can download the benchmark here and try it out:
http://github.com/aristus/parse-n-load

]]>