Web Browser Grand Prix: The Top Five, Tested And Ranked

[husker]

There is no single answer for any of this. Complex systems with this many variables can only be compared when you are looking at very specific issues which can then NOT be generalized. A trivial example: I can prove that walking is faster than taking a commercial flight. How? We start at my house and you take the plane and I'll walk. The first one to my neighbor's house wins.

 

[Guest]

>>If you haven't yet downloaded Google Chrome, you just don't know what you're missing.

I know exactly what I'm missing, the Find-as-you-type feature of FireFox which will prevent me from ever switching to any other browser. Hitting Ctrl-F slows down my browsing experience more than all those tests combined.

 

[Guest]

The results should had been presented as the mean values of rank, and not the most first places. The browser should be good in average.

From order of ranks:

In[5]:= (7*1 + 2*9 + 2*3 + 4*4 + 5*5)/(7 + 9 + 2 + 4 + 3.)

Out[5]= 2.88

In[4]:= (7*1 + 2*5 + 6*3 + 4*4 + 5*5)/(7 + 9 + 2 + 4 + 3.)

Out[4]= 3.04

In[3]:= (1*5 + 2*4 + 12*3 + 2*4 + 2*5)/(7 + 9 + 2 + 4 + 3.)

Out[3]= 2.68

In[6]:= (4*1 + 2*5 + 4*3 + 10*4 + 2*5)/(7 + 9 + 2 + 4 + 3.)

Out[6]= 3.04

In[7]:= (2*1 + 2*3 + 3*3 + 3*4 + 14*5)/(7 + 9 + 2 + 4 + 3.)

Out[7]= 3.96

 

[Guest]

@mitch074

Indeed compiler options and "supported" hardware make a difference.

Chrome is aggressively compiled for P4, for instance, so trying to display a on anything less will cause the page to crash. Whereas FF works perfectly fine.

 

[Guest]

Your memory test using multiple tabs is bunk. You made the assumption that lower was better in the memory test and on a single tab you're right. However Chrome has a liner growth in memory because it does everything it can to keep the web pages separate in memory and process so a crash in one doesn't bring down the entire browser. If the browser doesn't have a liner growth in memory as you open more tabs then it isn't doing this right.

 

[Guest]

I've been saying Opera was my favorite browser for many years. I use them all, but I'm happiest with Opera.

 

[JohnnyLucky]

WOW! Eight pages of comments. Did Tom's Hardware push the hot buttom or what? Is this a new record for posts?

 

[JohnnyLucky]

WOW! Eight pages full of comments. Did Tom's Hardware push the hot button or what? Is this a new record?

 

[ominous prime]

In Windows 7 x64 I experience major slowdowns when using Chrome. In some instances I can load a page in Chrome, then open firefox and load the same page, and it will finish in Firefox before Chrome.

The benchmarks were interesting, I would like to use Chrome but it's unusable for me. Another pet peeve would be the ad hiders on Chrome, no blocking yet.

 

[endo13]

" I've been running Chrome as my primary browser for several months now. Before that, I was a Firefox user since it's very first release. I'm reluctant to admit that I switched to Internet Explorer for version 3.0, and I suppose it must have been Netscape Navigator before that. I personally haven't used Internet Explorer in ages, and truth be told, I have never actually used Safari before this. I check out Opera just about every time they come out with a new release, but two weeks is probably the longest uninterrupted duration that I have ever seriously used that browser."

I find that interesting in how eerily similar it is to my own browser usage history.

 

[Guest]

I use only Linux, and I've found that Firefox is much faster for me than all of my Windows-using friends.

 

[Guest]

You do realize the nontroppo tests are fundamentally broken, right? They use onload to decide when the page is done loading, and Webkit lies about when onload fires. See http://weblogs.mozillazine.org/bz/archives/020434.html

 

[Guest]

One more comment. The CelticKane tests are very buggy (in the sense that they don't measure what they claim to measure) and very easy to game. They're not particularly indicative of actual JS performance, or DOM performance, or really anything other than how one does on them. So as a benchmark, they're pretty useless.

 

[Guest]

What about on the Mac? Besides that Chrome is buggy on Mac OS X, Safari runs much much faster on it's native OS - Mac OS X. I'd love to see a benchmark suite on a different os

 

[crom]

For me customization, plugins, and security are key. That's why I use Firefox as my primary browser. It isn't the fastest one out there, but its by far the most extendable. Adblock, NoScript, etc. It really allows me to surf the web in a much more secure way and view it as I'd like to see it, not what marketers want me to see. If that was the case I'd use IE or Safari.

 

[robertbradbury]

You are testing/benchmarking "toy" sessions, not real world sessions that people care about. Real world sessions may involve dozens of windows and hundreds of tabs and things that people care about are:
1) When do I run out of CPU? (what Mozilla based browsers are prone to do);
2) When do I run out of memory? (what Chrome is prone to do);
3) How long does a large session restore take? (if it isn't maxing out the network/CPU for 10-20 minutes you aren't testing a "real world" scenario);
4) How "Green" is the browser? (when left alone for extended periods does it recognize this and attempt to minimize CPU/memory/network use to be extremely "Green"?) -- When people have to upgrade PCs in order to support large sessions in a browser (when one is generally only looking at a single tab in a single window) then there are problems with the OS/Browser implementation.
5) How user friendly is the browser? I.e. how effectively can one control things like Flash, Ads and Javascript? -- because its my computer not a proxy tool for commercial entities.

 

[chechak]

Google Chrome is the real speed king but when talking about options opera combine both speed and a lot of feathers

 

[Guest]

The memory test is useless. Most browsers now allow you to fine tune what, where, and how much is cache. So, in reality you are only testing what the default setups are for each browser. For example, Opera by default, for at least the 10.x series releases, is set to heavily use memory and cache a lot of stuff. This design decision has gotten a lot of play on the Opera forums and mailing lists.

 

[the_man]

The article is slightly biased towards Chrome.
Chrome suppsedly wins the ACID3 category only by looking at the screen refresh, which may be down-prioritized (and rightly so) to execute JS, etc. The Acid3 test accounts for this, and gives the test timings by clicking the A in Acid3. On my 3-year old laptop, the timings are:

Opera 10.50:
Failed 0 tests.
Test 26 passed, but took 353ms (less than 30fps)
Test 65 passed, but took 138ms (less than 30fps)
Test 69 passed, but took 16 attempts (less than perfect).
Test 79 passed, but took 41ms (less than 30fps)
Total elapsed time: 1.45s

Chrome 4.0.249.89:
Failed 0 tests.
Test 14 passed, but took 44ms (less than 30fps)
Test 15 passed, but took 42ms (less than 30fps)
Test 26 passed, but took 116ms (less than 30fps)
Test 69 passed, but took 91 attempts less than perfect)
Test 70 passed, but took 44ms (less than 30fps)
Test 71 passed, but took 49ms (less than 30fps)
Test 74 passed, but took 45ms (less than 30fps)
Total elapsed time: 2.91s

Safari 4.0.4
Failed 0 tests.
Test 00 passed, but took 91ms (less than 30fps)
Test 26 passed, but took 98ms (less than 30fps)
Test 43 passed, but took 51ms (less than 30fps)
Test 65 passed, but took 41ms (less than 30fps)
Test 69 passed, but took 46 attempts (less than perfect).
Total elapsed time: 7.84s

On my computer Chrome uses twice as long for Acid3 as Opera, uses too long time on almost twice as many tests, and still Chrome wins. And no data to back it up. Unless this has been checked carefully on the test system and measured differently BY THE ACID TEST ITSELF, there is no way Chrome is the sole winner of Acid3.

If the dubious ACID3 test results are dropped, the results are:

Test wins:
Opera 7
Chrome 6
Safari 5
Firefox 4
Internet Explorer 2

Category wins:
Opera: 4 wins
Safari: 3 wins
Firefox: 3 wins
Chrome: 2 wins

To me, it does not look like Chrome is the winner. Opera is.

 

[obarthelemy]

Hey, I got an idea ! let's compare perfs between

1- a fully usable Opera (with its integrated mouse gestures, adblock, noscript, synch...)

2-an unusable barebones Firebox with 0 addons.

 

[m2pc]

@tomtompiper: "Chrome records everything you type in the search bar and sends it back to Google. If you want big brother get Chrome."

All browsers with a "suggestion" feature do this. It's impossible to get a list of suggestions without sending what you've typed to a server to retrieve a list of possible choices!

Try running a packet sniffer on any browser with this feature, and I'm sure you will see similar activity. Not all "phoning home" is bad.

 

[Ironeinstein]

I wonder which one sponsored the article....

 

[techguy911]

Well try logging onto gmail with chrome CRASH!.

http://www.maximumpc.com/article/news/chrome_glitch_nixes_offline_gmail

 

[Guest]

The conclusions don't support the numbers. I ran these numbers again using z-scores. (In each test, the top browser scored 5, and the bottom browser scored 0, the rest scored somewhere in between.) This way, if a browser was narrowly beat, they still got almost as many points as the browser who scored first. When the results were analyzed correctly, Opera comes out on top. Here are the results:

Opera: 74.7
Chrome: 74.1
Firefox: 72.2
Safari: 69.9
Explorer: 36.3

Here you can see just how close the race is between Chrome and Opera. And, you see how terribly Explorer performed. While it is close, Chrome is second to Opera, not first.

In case you're curious, here's how to find these scores:

1. Find the difference between the top and bottom result in each test.
2. Divide that difference by 5 to get the "standard deviation"
3. Find the difference between each result and the top result.
4. Divide that difference by the standard deviation.
5. This result is always 0-5 (first place gets 0, and last place gets 5.)
6. Subtract this number from 5 to "invert it" now the high score gets 5, and the low score gets 0.
7. Find the sum of all the scores for each browser. The winner is the browser with the most points.

 

[Guest]

The reason IE is the most used, is because at a lot of companies, like mine, management like to buy crapplications that a bunch of VB monkeys created, and it required IE...