The latest versions of the top five major web browsers were tested under six performance indicators to see how much they stack up against each other. Browsers were tested three times with unprimed caches except during caching performance benchmarks. The mean values are reported below.
CPU usage reveals how much system resources a browser needs: resource hogs show higher CPU utilization. Windows Resource Monitor was used to obtain an average CPU occupation (%) while SunSpider was running to simulate activity.
The faster the browser can select elements in a web page, the more responsive it is on asynchronous page updates (which most Web 2.0 apps heavily rely on). SlickSpeed was used to see how fast jQuery selects elements.
Browsers with fast CSS rendering speeds have faster page response times. The nontroppo.org CSS Rendering Benchmark was used to measure the onLoad duration for complete table-to-div conversion.
The total time it takes to load Yahoo.com’s front page was measured using Numion Stopwatch. Note that due to latency differences that occur with variable site traffic and server load, caution should be used when interpreting the results.
Total page load times for Yahoo.com with primed caches were measured to see how well browsers perform when you have visited a website already. The same variable latency difference may be relevant here with calculating page load times.
Based on the results, the relative rating of each web browser is displayed above.
In the above graphic you can find additional information, as well as the details for the testing machine.