Web Browser Benchmark Results Comparison

I compared web browser benchmark results back in 2008. Illustrious browsers such as Google Chrome 1, Firefox 3.1, Internet Explorer 8 beta 2 or Opera 10 alpha were tested and compared. The results from back then are history, with the majority of tested web browser versions no longer available. Google for instance has recently released version 12 of the Chrome browser, Mozilla is closing in on Firefox 4 (which will apparently be released on March 22), Microsoft just released Internet Explorer 9 and Opera the first beta of Opera 11.10.

Time for a now Internet browser benchmark comparison. All benchmarks are run on a 64-bit Windows 7 Professional PC with an Intel Core i7 860 Ghz cpu, 8 Gigabytes of RAM and a NVIDIA GeForce GTX 470 display adapter.

The following browsers have been included in the benchmark comparison:

• Firefox 3.6.15
• Firefox 4 RC
• Google Chrome 11.0.696.12 Dev
• Google Chrome 12.0.705.0 Canary
• Opera 11.01
• Opera 11.10 Beta
• Internet Explorer 9
• Safari 5.0.4

The following benchmarks have been utilized:

• HTML5 Test
• Acid 3 Test
• Sunspider JavaScript Benchmark
• Fish IE Tank

Results

Both Google Chrome versions are leading the HTML5 test. They are the only browser that reach a score higher than 300 on the test. Internet Explorer 9 and Firefox 3.6.15 bring up the rear with scores below 200 points.

The three browsers Safari, Google Chrome 12 and Opera 11.10 Beta manage to score 100 out of 100 points in the Acid 3 test. The remaining browsers all score more than 90 points, with Firefox 3.6.15 bringing up the rear with 94 points. It is a little surprising that the current Dev version of Google Chrome scores 96 points.

Microsoft’s Internet Explorer 9 scores the best result in the SunSpider JavaScript benchmark. It took the browser 196.9 ms to complete the benchmark. Firefox 4 RC, Google Chrome and the Opera browsers all completed the benchmark in less than 235 milliseconds. Only Apple’s Safari with more than 300ms and Firefox 3.6.15 with more than 650ms scored considerably worse in the benchmark.

The last benchmark, Microsoft’s IE Fishtank, benefits highly from hardware acceleration. All browsers with hardware acceleration ran the demonstration in 35 fps or more, with Internet Explorer 9 reaching the maximum of 60 fps.

Firefox 3, Opera 11 and Safari do not utilize hardware acceleration yet which explains the mediocre score of 2 respectively 3 fps in the benchmark. Opera Software did however release a special build of the browser not long ago that supports hardware acceleration. The test ran in 60fps in that version of the Opera web browser (see Opera Hardware Acceleration Demo Is Blazingly Fast).

And here are all results in one chart. Red values indicate the lowest test result, green the highest.

Click on the screenshot to see it in its original size.

Closing Words

Benchmarks are not everything, especially since all companies and developers have improved the speed and web standards support of their browsers significantly. Some of the benchmarks may increase in importance over time, the HTML5 benchmark for instance or the Fish Tank demo that benefits from hardware acceleration. They become more relevant when websites start implementing technologies on their websites that make use of those features.

Internet Explorer 9 and Google Chrome have the best scores in two of the four different benchmarks. The browser with the worst scoring is Firefox 3.6.15 which will soon be replaced on many systems with the upcoming Firefox 4 which scored significantly better in most benchmarks.

All browsers have improved significantly when looking back at the test results of the benchmarking test in 2008. Both Google Chrome and Firefox were the two only browsers that completed the SunSpider JavaScript benchmark in less than 1000ms. That’s almost five times as much as today’s browsers need.

Related Articles:

Author: Martin Brinkmann, Thursday March 17, 2011 -
Tags:acid3, benchmark, browser benchmark, html5 test, sunspider, web browser

Responses so far:

1. Hans says:

   March 17, 2011 at 3:33 pm

   Your Acid 3 score for Opera 11.01 is definitely wrong. With a clean profile you get 100/100.
   And in Fishtank demo I have with Opera (not HW-accelerated) at least half of the fps I get in FF4 (HW-accelerated) or Chrome (HW-accelerated).

   Reply
   • Martin Brinkmann says:

     March 17, 2011 at 4:03 pm

     Hans I’m not sure what you mean by a “clear profile”. Are you running the Fishtank demo with 1000 fish? I highly doubt it that Opera 11.01 without hardware acceleration scores near the number that browsers with hardware acceleration get.

     Reply
2. Icedrake says:

   March 17, 2011 at 3:36 pm

   Hiya there,

   Reading through the article, I may have found a slight mistake. In the paragraph above the Acid3 test chart, you say that Internet Explorer scored 100/100. However, if you look on the chart, Internet Explorer scored 95/100 there, while Safari got the 100/100 (and it wasn’t mentioned in the writing as getting 100/100). Is this a mistake in the chart or in the writing?

   Thanks,
   Icedrake

   Reply
   • Martin Brinkmann says:

     March 17, 2011 at 4:04 pm

     Icedrake you are right. The charts are correct, the mistake was in the writing. I have corrected the issue. IE9 scores 95 out of 100, Safari scores the full 100 points.

     Reply
3. ne1scott says:

   March 17, 2011 at 9:18 pm

   Don’t you think the fact that you are running a “64-bit Windows 7 Professional PC with an Intel Core i7 860 Ghz cpu, 8 Gigabytes of RAM and a NVIDIA GeForce GTX 470 display adapter” instead of your old 2008 test rig has something to do with the 5X browser speed improvements ?

   Reply
   • Martin Brinkmann says:

     March 17, 2011 at 9:27 pm

     Yes that has something to do with it as well, my PC back then was slower. Still, I believe the main factor are improvements implemented by the developers.

     Reply
4. ne1scott says:

   March 17, 2011 at 9:21 pm

   Just curious why Google Chrome 10.0.648.134 was omitted since Google only pushed it out to consumers last week.

   Also, do you have any benchmarks that are heavily flash based ? With the popularity of facebook flash games that absolutely bring IE8 to it’s knees and still cause some issues in chrome and firefox, I think it would be a very relevant benchmark.

   Reply
   • Martin Brinkmann says:

     March 17, 2011 at 9:29 pm

     I did not do this on purpose. Had only access to the dev and canary builds on the PC.

     Reply
5. Steve says:

   April 7, 2011 at 1:51 am

   I recently did some benchmarks myself and posted it on my new blog. I used the current versions as of 7th of April 2011

   http://cyberb1og.blogspot.com/2011/04/battle-of-browsers.html

   Reply
6. fakhry tatanaki says:

   August 12, 2011 at 4:45 pm

   i like google chrome

   Reply