Safari is fast

Lifehacker every now and than comes out with a browser test article, discussing which browser on the market is the fastest. They usually give windows centric focus, though to be fair at times--- at least I remember there are test cases on the Mac side.

Chrome, for the first time in a long time, leaped into first place, no doubt due to its new prerendering feature and always-awesome JavaScript tests

Browser Speed Tests: Chrome 17, Firefox 10, Internet Explorer 9, and Opera 11.61 -Whitson Gordan

I'm not trying to discredit any of Mr. Gordon’s point, just throwing alternative scenarios out there.

I wanted to write this brief article showing that Safari on the Mac is clearly the best browser on the market for the Mac.

My process is to use the same test-suite Lifehacker uses on Windows and make the closest approximations on my own system. All browsers are up to date and have been completely reset. Websites used in tab-loading test were the top 9 Alexa sites.

Cold Boot

!Cold Boot

I wanted to use Instruments to accurately measure the finish of loading an app on my Mac but Lifehacker stuck to a stopwatch method so, I shall too. Firefox wins here and I'm not surprised, they are doing some good word in their aurora builds. Do the typical person the other results were nearly instantaneous. Opera did not fair to well, I had a noticeable ball bounce in the dock. Interesting note: the ball-bouncing in the dock will soon be a thing of the past as app load times becomes instant.


Tab Loading

Now I rarely launch a browser with 9 tabs but there are those that do. This result surprised me. I expected Chrome to come out on top with its highly multiprocess architecture. But, Opera came out on top. All the others were noticeable same.



Test are run multiple times in the Dromaeo suite and an average given, so point in running the test multiple times.

Opera trailed the pack while Safari and Firefox came in a close 2nd. But Chrome ran away with its performance.



These are more intensive tests, in the JavaScript library tests of the suite each browser ran off with memory. I didn't like this; anything that causes Safari to run off with 1.96 GB of RAM in its Web process will slow down a system and I feel this influenced the results. All the browsers experienced this behavior so I concluded to leave this particular test out of my results. DOM and CSS scores are important because they effect a visual element that is clearly seen by the user.

Opera dominates in the CSS score but is far behind the park in DOM score. I found this confusing.

Memory Usage (with Nine Tabs Open)


Typically I'm a memory fanatic; I like my system having a good amount of free resources at any time. I run 4-5 apps simultaneous in my workflow and having a browser open is one of them. I loaded the same sites that Lifehackers used via an applescript and waited for all the progress indicators to finish in the tabs. Just like the tab-loading test. I left the browser running for 15 minutes than recorded the memory usage via Activity Monitor. Opera and Firefox have always been good at memory (for me at least) that I was not surprised they led the pack.

I wanted to point out one note, with Chrome and Safari, you never really should quit the app, Command + W has always been my way to "quit" the app. This leaves the parent process running in the background with a minimal footprint. Additional browsing sessions are near instantaneousness in this behavior. This process architecture is an Apple defined behavior so if I were to subtract the parent process from the actually window and logic process for Safari or Chrome I would get much lower scores.


After a week of use on each browser, I've come to the conclusion that each browser does one thing really well and average on other tests. There is no one browser wins all. Safari tends to launch faster but not noticeably different from Chrome. Chrome handles multiple tabs really well and its memory footprint is not so bad unless you keep it running (especially with extensions) Firefox is average in all areas but this being my 2nd browser-testing article, they are improving significantly. Opera is an odd app. It loads slowly but tabs are loading quickly, I suspect it’s their unique caching strategy. But actually speed performance, its last in JavaScript and this might effect is DOM and CSS scores. Which leads me to those, completely lopsided results I intend to investigate are they doing some test hacking? Or is just a standards issue.

Overall I will be sticking with Safari and here are my reasons. * On the Mac, it is the one browser that runs the smoothest. * It’s not fancy, its minimal appearance. I don't need favicons in my tabs; I like the extra screen real estate on the top. * It is the most memory efficient with extensions. Yes, they are the most restricted but I only use two or three at a time. It’s a quality over quantity matter. * ICloud syncing. I like my bookmarks synced across my devices without any additional programs. And I do use the Reading list feature extensively.

So, every now and than I will go over to Chrome and do some testing, see if anything changes and I'll find this enjoyable. But I always end up coming back to Safari.