Tom’s Hardware recently completed their 2013 browser benchmark. The thorough test looks at all aspects of browser performance including start-up times, rendering, memory use, reliability and standards support.
This year’s winner is … dum-de-de-dah … Firefox.
Yes, really. The browser’s reputation for speed and efficiency had been damaged in recent years, but Mozilla has turned the situation around. At least on Windows.
The test analyzed the five main Windows 8 browsers:
- Google Chrome 27 (nine previous wins)
- Microsoft Internet Explorer 10 (two wins)
- Mozilla Firefox 22 (three wins)
- Opera 12 (two wins)
- the beta version of Opera 15 which uses Blink and V8 (the final edition was released this week)
Start-up
The start-up time for a single tab was won by IE10. It was followed by Firefox, Opera 12, Opera 15 and Chrome which was two seconds slower than IE.
However, when the test is performed with eight tabs, Firefox is twice as fast as Opera and Chrome with IE10 coming last. However, that’s a cold-start; when you restart the browsers again, IE10 is slightly ahead of Firefox.
The aggregated results put Firefox at the top, followed by IE10, Opera 12, Opera 15 and Chrome.
Page Load
Chrome takes the lead for loading a single page. As expected, Opera 15 is almost identical followed by Opera 12, IE and Firefox.
Loading 40 cached tabs yields a slightly unusual result: Opera 15 is the fastest and Chrome the slowest. This could be resolved in Chrome 28 when it moves to Blink, but I suspect a combination of Chrome’s multi-process architecture and calling home to Google may hamper it. Opera 12 was in second place. IE was in third although the testers experienced stability issues. Firefox was fourth.
Overall, Opera 15 comes out on top but version 12, Firefox and IE are only marginally less. Despite loading a single page fastest, Chrome falls to fifth.
Performance
The test uses aggregate scores from Webkit SunSpider, JSBench, RIABench, Futuremark, Rightware and Mozilla Dromaeo benchmark suites to assess JavaScript execution and DOM manipulation. Chrome wins, followed by Opera 15, Firefox, Opera 12 and IE.
However, benchmarking is rarely a test of real-world performance so the testers used a number of HTML5 and CSS3 suites which assess tasks such as applying a series of photo effects, face detection and games. The overall result still has Chrome ahead, but Firefox and IE10 are very close.
Finally, the testers looked at WebGL and hardware acceleration. It was a photo finish for Chrome and Firefox with Opera 15 a little behind. IE and Opera 12 don’t support WebGL.
Memory Use
The overall efficiency result placed Opera 15 and Chrome at the top followed by Firefox, IE and Opera 12. However, it depends on what you’re doing. For a single tab, IE leads with Firefox in last place. With 40 tabs, Firefox uses half the memory required by Chrome, but is a little less efficient at releasing it.
Reliability
During the 40-tab reload, the team assessed the reliability of each browser to ensure all pages had fully rendered. The results were surprising:
- Firefox loaded all pages correctly
- Opera 12 required four tabs to be refreshed
- IE10 required eight (and the browser crashed first time)
- Opera 15 required nine refreshes
- Chrome required twelve — 30% of pages were incomplete or incorrectly loaded
Standards Support
When you take HTML5, CSS3 and JavaScript standards support into consideration, Chrome comes out top with Opera 15 a little behind. Firefox, Opera 12 and IE10 follow but the differences have narrowed considerably.
The Championship Results
Tallying the scores for all tests puts Firefox in a commanding lead:
Image may be NSFW.
Clik here to view.
Firefox comes top in relatively few tests but it’s often in a strong second or third place and offers good performance in all areas. Chrome leads in many speed tests but is let down by start-up times, multi-tab memory use and a slightly shocking reliability result.
A quick glance at the chart may lead you to conclude there’s a gaping chasm between the best and worst browsers. Don’t believe it. While tests such as this are interesting, the applications are closer than they’ve ever been — no single browser is an outright winner in all areas.
We’re at a fortunate point in web history where old IEs are (almost) irrelevant and it no longer matters which browser a user decides to use. Long may it continue.