Rendering speed has long been a debatable matter and we all know there’s no completely fair and accurate method to discover the fastest browser. However, there are some tests which can be seen as the most relevant tests to see which engine is the most efficient at rendering pages, thus which one can be considered as the most efficient for your daily browsing.

The test I used are in the count of two: one for HTML rendering and one for javascript. For CSS render I used this test (running it from a local file), despite the warning that on Safari and Chrome it can’t measure the time accurately. The warning is not accurate at least for Chrome since due to the efficiency of its javascript engine, the relative running difference can be used for comparison. I chose these two because they test the browser under usual daily stress: rendering HTML and javascript (but not actual CSS). All types of code are embedded to eliminate the lag due to loading of external files, which doesn’t depend on browser but on internet connection. These tests all relate exclusively to browser engine and how the rendering engine deals with a loaded page.

To ensure fairness, the results were taken on the first visit, so we are not mislead by cached pages. Here we go …

Safari 4 Beta: 1) 1.62; 2) 60 ms; 3) 6ms

Internet Explorer 8.0.6: 1) 3.66 s; 2) 599 ms; 3) 14 ms

Mozilla Firefox 3.0.10: 1) 3.26 s; 2) 170 ms; 3) 35 ms

Mozilla Firefox 3.5: 1) 3.80; 2) 47 ms;

Google Chrome 2.0.172: 1) 2.11 s, 2) 22 ms; 3) 11 ms

Google Chrome 3: 1) 3.52  ;2) 16 ms

Opera 9.64: 1) 2.25 s; 2) 224 ms;  3) 6 ms

Opera 10 Beta 1: 1) 2.27 s; 2) 226 ms; 3) 7 ms

What can we say after seeing all the tests? First of all, please note that this is no standard compliace testing. In fact the code within the tests is old by some standards, exactly to ensure that all browsers can run it. It is a speed test, not standard compliace test. No browser reported any errors during testing, except IE 8 who kept complaining that the javascript test made it run slow (for good reason apparently). IE’s low tolerance for high load javascript applications is a really bad mark and in fact it may have tripled the real rendering time for JS. I can’t stress how bad and poor IE 8 behaved.

Other than that, I am really surprised by Chrome’s speed with the JS. I remade that test on several computers, all in the first run. Apparently Chrome is not only using few resources but it also uses them extremely efficient so even poorly written JS will still perform well on Chrome.

Between Firefox and Opera things look like a tie. Opera was better with HTML rendering while Firefox did well with JS rendering. But overal both were nothing else than average. Not good, not bad, just average.

Safari 4 was a very nice surprise behaving the best with the pure HTML rendering, while doing a great job with the javascript as well. I can only commend the new generation of browsers spearheaded by Safari and Chrome for speed and safety against which the old guard should watch out.

To me, Internet Explorer has been dead since version 7 and while version 8 has quite some improvements, it’s far from enough, far from good. Firefox still has to deal with a horrifyingly poor memory management algorithm which slows everything down and after hours of browsing ends up eating hundreds of megabytes of RAM. Opera is very efficient in this respect, but a website with a fair load of javascript can easily kill Opera. Also, the new Opera 10 Beta doesn’t seem to bring any improvements whatsoever in this respect.

UPDATE: Firefox 3.5 with very minor improvements, while Chrome 3 came out with some caveats unfortunately.

I will not comment on the rendering of CSS test due to the doubt surrounding it.

Considering the standard ACID 3 test results (found here), I can say that Safari 4 and Chrome 2 can easily take the browser crown (and share it).