Performance Benchmarks: HTML5 Hardware Acceleration And WebGL
HTML Hardware Acceleration
This is probably the area of most interest to those involved in the IE/Firefox rivalry. HTML5 hardware acceleration is currently unique to these two browser war veterans. And fittingly, we've included one test from each outfit. Although the other three browsers cannot take advantage of graphics hardware, they do understand HTML5 and can run these tests. This will let us see how well they perform without the help of the GPU.
Mozilla Hardware Acceleration Stress Test
As expected, only Firefox and Internet Explorer reached 60 frames per second (the test's ceiling), creating a tie for first place. Without hardware assistance, Opera manages to run this test at 22 frames per second, taking second place. Chrome finishes in third with 20 FPS, while Apple Safari only allows four frames per second.
Psychedelic Browsing
Once again, it's IE9 and Firefox 4 in the top spots, though this time IE9 pulls ahead by 25 points. Far behind in third place is Chrome, with a score of 123. Opera takes fourth with 65 and Safari only achieves 31 points, finishing last.
The first test was a tie, and IE9 wins the second test. But keep in mind that Firefox 4 is only 25 points behind in a benchmark where the winner scores nearly 1800 points. These two Web browsers essentially perform equivalently.
WebGL
With the release of Firefox 4, we now have enough WebGL-capable browsers to make a comparison. IE9, Opera 11, and Safari 5 do not support WebGL.
Khronos Particles
The Particles benchmark from the Khronos Group, overseers of WebGL, places Google Chrome in the lead with just over 60 frames per second, though Mozilla still manages to push out 55 FPS
WebGL Aquarium
Chrome owns this test, running at 55 frames per second compared to Firefox 4's 22 FPS.
Overall, the results in these two tests are mixed, with Firefox close behind Chrome in Particles, but trailing pretty far in WebGL Aquarium. Either way, Chrome has a measurable lead over Firefox 4 in WebGL performance.