We seem to be having a browser week here at OMG! towers – posts on Chrome taking a massive chunk of Firefox’s Linux share AND shaking off the beta tag, noting Midori 0.2.6 being released yesterday, the blind-sided news that soon-to-be Kubuntu default browser Rekonq is adding support for Chrome extensions and, of course, even this weeks poll is browser based!
With all that in mind I couldn’t help but stoke the ever-tender embers of one question that continues to be a red-hot topic: “What is the fastest browser on Linux?”
Sunny Wunny Spider Crawled up the Java spout…
I ran the test 3 times in each browser. Each browser had a clean cache and no extensions loaded. The average results for each browser were as follows: -
It’s no surprise that Opera is giving Chrome a run for it’s cookies. The new Carakan rendering engine is immensely powerful - as we noted when we tested the very first alpha.
Browser Name: Number of tests failed out of 5246 (less fails = better)
- Rekonq: Couldn’t complete the test
- Firefox 264
- Midori: 195
- Chrome: 135
- Opera: 81
Both Opera, Midori & Rekonq have problems with many popular sites: Midori with GMail, Opera with Facebook and Reknoq pretty much anything you throw its way.
Snake in a browser (bring it on)
In the video below I measure the time it takes for 5 browsers (Epiphany excluded this time) to load the same website. Each browser had a clean cache, history etc and no extensions loaded. The results were actually surprising…
Midori did atrociously! Which is surprising as it generally feels snappy and swift. Perhaps Cobrastarship.com is badly designed or has a vendetta against Midori.
Below is a (bad quality) video showing that Midori is as capable as the hype suggests (which is why Midori rocks even though it can’t load Gmail!).
The results were as follows, in ascending order.
- Opera: 100/100
- Rekonq: 100/100
- Epiphany: 100/100
- Chrome: 100/100
- Midori: 98/100
- Firefox: 94/100
Midori’s result seems a bit of a fluke as it is *supposed* to pass fully with 100/100. I ran the test 4 times and got the same result each time.
As with any test conducted the environment and platform of said test will have a bearing. All of my tests were carried out on the same computer in the same operating system with the same applications running during each (i.e. Docky).
My results are not definitive by any means but hopefully give a slight insight into the performance of the current crop of Linux browsers.