There are only a few browsers even worthy of consideration at the moment – due to lack of support, infrequent updates, security holes, performance issues etc, I’ve decided to ignore all other browsers except for the main five. Legacy browsers (e.g. Firefox 2.0, IE7, Safari 3, Netscape were excluded), and the latest versions and betas available at time of testing were used.The versions I tested were:
- Google Chrome 2.0.172.33
- Google Chrome Beta 3.0.193.0
- Opera 9.64
- Opera 10 Beta Build 1631
- Safari 4.0.2
- Firefox 3.0.11
- Firefox 3.5
- Internet Explorer 8
Aside: FYI, I’ve long since lost all traces of browser loyalty – they’ve already wasted enough of my time and sanity for me to even bother blowing the X browser horn.
I don’t really believe in browser benchmarks – the only real way to test browsers out is to use them for a few days and monitor their performance and system resource usage. The SunSpider and V8 “industry standard” benchmarks are utter crap – every tech blog, journal and media portal seem to toss their figures around willy nilly, yet fail to realise that both benchmarks are BIASED and UNFAIR when used to compare competing browsers. They were designed to test performance of different builds, updates, improvements and so on, not the performance of a completely different competitors product. SunSpider was developed by WebKit (the basis of Safari) as an internal benchmark, as was V8 for Google Chrome. The only “un-biased” benchmark currently available is ACID 3, but even that doesn’t really simulate real world usage.
So, I gathered 301 different websites – each laced with images, Javascript, Java applets, Flash, Shockwave, Silverlight, pop-ups and AJAX and loaded them in each browser to check for memory leaks, abnormal CPU usage and disk I/O. I then tried using each browser with 30 tabs open for a day or two each, to see how they performed over time. As a test of each browsers tagging, history and bookmark searching capabilities, I also imported my 8MB bookmarks.html (where possible).
Each browser was setup to a “usable” state – i.e. pop-up and adblocking enabled if available, dangerous javascript options disabled. All other options were left at default, as settings shouldn’t need tweaking out of the box for a browser to be usable.
Test system specs were:
- Overclocked Intel Core 2 Quad 2.4ghz @ 2.95Ghz
- 4GB 4-4-4-12 DDR2
- 3x 250GB WD2500KS SATA Hard Disks in a RAID0 Array (used as system disk)
- Nvidia GeForce 7600GT
- Windows XP Professional SP3 32-Bit with latest patches as of 13/7/09 applied
- Norton Internet Security 2008 16.5
So by no means a slouch. Keep in mind the average PC will NOT be anywhere near as powerful.
More to come in following posts…
Recent Comments