
We have a whole lot of exciting geekery planned today. But first, let's get everyone up to speed. There have been a handful of developments in the world of Web browsing since Web Browser Grand Prix 5: Opera 11.50, Firefox 5, Chrome 12:
Recent Events
07/20/11: Apple Releases version 10.7 of Mac OS X, dubbed Lion, along with Safari 5.1
08/02/11: Google Releases Chrome 13
08/16/11: Mozilla Releases Firefox 6
08/25/11: Steve Jobs resigns as Apple's Chief Executive Officer. Contrary to what many speculated, Apple's stock remains strong.
Recent Drama
08/13/11: A Mozilla developer incites fear and chaos by suggesting the removal of version numbers altogether.
08/15/11: Microsoft Internet Explorer 9 is declared the Web browser most safe from phishing.
08/16/11: Google refutes the validity of that report.
Mac OS X
Back in July of last year, we threw Web Browser Grand Prix (WBGP) contenders Chrome, Firefox, and Opera onto an Ubuntu Linux track in Web Browser Grand Prix 2: Running The Linux Circuit. The results were compared to those of their Windows 7 counterparts from Web Browser Grand Prix 2: Top 5 Tested And Ranked. As the “Linux guy,” doing that article seemed perfectly natural to me. After all, how Web browsers perform in Windows really didn't matter to me, personally.
What we didn't see coming was the number of requests for a Mac-based Web Browser Grand Prix, and the calls for us to run Safari on OS X haven't stopped since.

In case you didn't notice, this is Web Browser Grand Prix VI, not Web Browser Grand Prix 6. That's because the twist this time is Mac OS X. We're running Chrome, Firefox, Opera, and Safari on Mac OS X, along with the usual suspects on Windows 7.
Tom's Hardware is going Mac?
That's not it at all. We said we're running Chrome, Firefox, Opera, and Safari, on Mac OS X. We did not say we're running them on a Mac. We're using the same Core i5-based desktop test system used for many of our other software-based stories. It just so happens we had OS X Snow Leopard testing in mind when it came to designing a reference cross-platform test system.
WBGP Test Suite v6.0
Once again, we tweaked our test suite for this story. The new composite page-load time tests are improved, expanded, and updated. There are all-new WebGL benchmarks, and Sputnik is replaced by Ecma test262. The methodology of the Maze Solver benchmark and memory management tests is improved to better reflect real-world usage scenarios. Enhanced placing tables and improved analysis round out the changes.
Which Web browser is best? Which operating system holds Web browsing supremacy? We'll answer that. But before we fire the starting pistol, let's take a deeper look at the contenders and course.
![]() | Chrome | |
|---|---|---|
| Vendor: | Google | |
| Debut: | 9/02/2008 | |
| Current Version: | 13 | |
| Layout Engine: | WebKit | |
| JavaScript Engine: | V8 | |
| Supported Platforms: | Windows, Mac, Linux, ChromeOS | |
| HTML5 Hardware Acceleration: | No | |
| WebGL: | Yes | |
| WBGP Wins: | 3 (WBGP1, WBGP2:Linux, and WBGP5) | |
| Download Chrome! | ||
Chrome 13 on Windows 7
Chrome 13 on Mac OS X Lion
![]() | Firefox | |
|---|---|---|
| Vendor: | Mozilla | |
| Debut: | 11/09/2004 | |
| Current Version: | 6 | |
| Layout Engine: | Gecko 2.0 | |
| JavaScript Engine: | JagerMonkey | |
| Supported Platforms: | Windows, Mac, Linux | |
| HTML Hardware Acceleration: | Yes | |
| WebGL: | Yes | |
| WBGP Wins: | None | |
| Download Firefox! | ||
Firefox 6 on Windows 7
Firefox 6 on Mac OS X Lion
![]() | Internet Explorer | |
|---|---|---|
| Vendor: | Microsoft | |
| Debut: | 8/16/1995 | |
| Current Version: | 9 | |
| Layout Engine: | Trident | |
| JavaScript Engine: | Chakra | |
| Supported Platforms: | Windows (Vista and 7) | |
| HTML5 Hardware Acceleration: | Yes | |
| WebGL: | No | |
| WBGP Wins: | 2 (WBGP3 and WBGP4) | |
| Download Internet Explorer! | ||
Internet Explorer 9 on Windows 7
![]() | Opera | |
|---|---|---|
| Vendor: | Opera Software | |
| Debut: | 12/09/1996 | |
| Current Version: | 11.50 | |
| Layout Engine: | Presto | |
| JavaScript Engine: | Carakan | |
| Supported Platforms: | Windows, Mac, Linux | |
| HTML5 Hardware Acceleration: | No | |
| WebGL: | No | |
| WBGP Wins: | 1 (WBGP2) | |
| Download Opera! | ||
Opera 11.50 on Windows 7
Opera 11.50 on Mac OS X Lion
![]() | Safari | |
|---|---|---|
| Vendor: | Apple | |
| Debut: | 1/07/2003 | |
| Current Version: | 5.1 | |
| Layout Engine: | WebKit 2 | |
| JavaScript Engine: | Nitro | |
| Supported Platforms: | Windows, Mac | |
| HTML5 Hardware Acceleration: | Yes (Mac only) | |
| WebGL: | No | |
| WBGP Wins: | None | |
| Download Safari! | ||
Safari 5.1 on Mac OS X Lion
Safari 5.1 on Windows 7
All Apple products are known as much for their form as their function, and Safari is no exception. Apple is a total experience company. They control the horizontal. They control the vertical. Pairing Apple products with other Apple products is how they're intended to be used. So, let's take a look at what Safari does on Apple's own operating system with a spotlight on Safari for Mac.
Safari 5.1 for Mac OS X Lion has some pretty cool features that aren't in the Windows version. While none of the capabilities in this spotlight enhance performance, efficiency, reliability, or conformance in any way, they could very well mean more to a user than any of the above. Every one of these features is completely unique to Safari for Mac. And in requisite Apple fashion, they all display a certain level of polish that only Cupertino can pull off.
| Full-Screen |
While full-screen browsing is certainly nothing new (most Web browsers have had that option in their view menus for ages), the way in which Safari handles it is more functional than the traditional completely full-screen method. Oh, and the animation is slick, too.
While in full-screen view, the navigation and tab bars remain on-screen. There's no more hovering to the uppermost edge to reveal these important controls. However, hovering to the top of the screen does open the Mac global menu bar for full control. And the most important thing about the new full-screen feature is that it's the only way to maximize Safari. Since Mac OS X doesn't have a maximize button (it uses an "intelligent resize" button instead), we sure do welcome full-screen.
| Multitouch Gestures |
Like Lion itself, Safari 5.1 supports several multitouch gestures, provided you have a suitable input device.
Scroll
Swiping upward with two fingers causes the page to scroll down. Likewise, swiping downward with two fingers scrolls the page up. The scroll gesture also responds to momentum. The faster you swipe, the faster and farther it scrolls.
Back/Forward
Using the same two-finger swipe as the scroll gesture, performed left and right, controls navigation. Swiping two fingers to the right navigates to the previous page in your history, and swiping left moves forward.
The back and forward browser buttons are probably some of the most clicked-on controls in any UI. But never before has it been both faster and easier to use them as it is with multitouch gestures.
Somewhat like the animations on tablets and certain e-readers, Safari mimics the page turning of books and magazines, which the Web never really tried to emulate (despite the fact we call them webpages).
Pinch-to-zoom
Like the iPad, iPhone, and iPod Touch, pinch-to-zoom now appears in OS X. Placing two fingers on the multitouch surface and moving them apart zooms the page in. Moving two fingers together causes the page to zoom out.
Tap-to-zoom
There is yet another way to zoom Safari with multitouch gestures in Mac OS X Lion. A double-tap with two fingers quickly zooms in on the portion of the screen near the cursor. A second two-fingered double-tap returns you to a neutral zoom level.
| Reader |
Apple has a unique feature in Safari called Reader. Reader essentially converts webpages into a clean "page," like you would expect to find in a PDF. This "page" is displayed above the original webpage, which is darkened out.
The Deus Ex 3 Article Before Safari Reader
After Reader
One of the best parts about Reader is that it ditches sidebars and ads. Overall, we're pretty surprised at just how effective this feature is in focusing on and clarifying the actual content of the webpage. Here's Reader and the next feature, Reading List, in action:
While not a new feature, the updated implementation is slicker, and the whole "page" illusion works much better with multitouch.
| Reading List |
Another portion of Reader is the Reading List. The Reading List is a sidebar for saving webpages for later reading. Hence, Reading List.
Readin List Safari 5.1 on Mac OS X Lion
When combined, full-screen, multitouch gestures, Reader, and Reading List succeed at making the Web look and feel like a traditional print publication. It really is fantastic. Not magical. Just fantastic.
Hardware Setup
| Test System Specs | |
|---|---|
| Operating System 1 | Microsoft Windows 7 Ultimate (64-bit) |
| Operating System 2 | Apple Mac OS X Lion (64-bit) |
| Processor | Intel Core i5-750 (Lynnfield) @ 2.8 GHz, Quad-Core |
| Motherboard | Gigabyte GA-P55A-UD7, LGA 1156, P55 Express, F7 BIOS |
| Memory | 8 GB Crucial DDR3 @ 1333 MT/s (2 x 4 GB) |
| Graphics | AMD Radeon HD 4870 Reference Boards 512 GB GDDR5 (PCI-e 2.0) |
| Storage | Seagate Barracuda 7200.12 500 GB SATA 3Gb/s, 7200 RPM, 16 MB Cache |
| Optical | Asus DRW-24B1ST/BLK/B/AS |
| Power Supply | Corsair TX750W (750 Watt Max) |
| Chassis | Zalman MS1000-HS2 |
| CPU Cooler | Scythe Mugen 2 Revision B |
The Windows 7 Test Installation
The OS X Lion Test Installation
| Local Web Server Specs | |
|---|---|
| Operating System | Ubuntu 10.04 LTS Server Edition "Lucid Lynx" (32-bit) |
| Processor | AMD Athlon @ 1150 MHz |
| Motherboard | Soyo Dragon Platinum |
| Memory | 512 MB DDR |
| Graphics | AMD Radeon 9550, 256 MB GDDR |
| Storage | 40 GB Western Digital HDD WD400BB |
| Optical | Samsung DVD-ROM SD-616T |
| Extra Packages | Apache2, MySQL Client, MySQL Server, PHP5, PHP-GD, PHP5-MySQL, PHPMyAdmin, SSH |
As requested by readers in our previous stories, the table below hosts additional information on the test network.
| Network Specs | |
|---|---|
| ISP Service | Cox Premium (28 Mb/s down, 5 Mb/s up) |
| Modem | Motorola SURFboard SBS101U |
| Router | Linksys WRT54G2 V1 |
Hackintosh Setup
The table below lists what we used to get the Lion-based Hackintosh up and running.
| Installer | xMove |
|---|---|
| Bootloader | iBoot |
| DSDT | GA-P55A-UD7 |
While these tests are conducted on Mac OS X, this test system is not an Apple-branded Mac. We had to use a different bootloader than what you'd find on a genuine Mac, and real Mac systems use EFI instead of BIOS. Therefore, performance may vary from the same tests conducted on an Apple-branded system.
Also keep in mind that this use of OS X is completely unauthorized, and anyone doing this at home should have no expectations of receiving any help or support from Apple.
Software Setup
Both our Windows 7 Ultimate and Mac OS X Lion installations were freshly installed and fully updated as of midnight on August 15th. Power management and automatic updates were disabled before testing. The Web browsers and additional software, along with the exact version numbers tested, are listed in the table below.
| Software | Windows Version | Mac Version |
|---|---|---|
| Chrome | 13.0.782.215 | 13.0.782.215 |
| Firefox | 6.0 | 6.0 |
| Internet Explorer | 9.0.8112.16421 | N/A |
| Opera | 11.50 (1074) | 11.50 (1074) |
| Safari | 5.1 (7534.50) | 5.1 (7534.48.3) |
| AMD Driver | 8.850.0.0 | ATI Radeon HD 4800 series |
| Adobe Flash | 10.3.183.5 | 10.3.183.5 |
| Microsoft Silverlight | 4.0.60531.0 | 4.0.60531.0 |
| Oracle Java | 6.0.260 | 14.0.3 |
Test Setup
We restart the computer and allow it to idle for a while before benchmarking the next browser. Other than the conformance benchmarks, all of our final scores are an average of several iterations. More iterations are run on tests that have short durations, lower scales, and/or higher variations.
This time around, we're rating the benchmarks themselves. The tests are placed into one of four groups: core, observation, dated, and quarantine.
Core tests are considered current. These tests are usually trusted industry standards or our own creations, and they make up the core of the WBGP suite. Examples include our own page-load time tests, FutureMark Peacekeeper, and Mozilla Dromaeo DOM.
Tests that are either generally unknown, mostly untested, or just too bleeding-edge are placed under observation. The WebGL tests and Ecma test262 are examples.
Tests classified as dated are either outdated, losing relevance, or otherwise need replacing. We are actively seeking community feedback and contributions regarding alternatives to these benchmarks. Examples of this group include Acid3 and GUIMark Java.
The final group is for quarantined benchmarks. Benchmarks find their way into quarantine by delivering dubious results or by being gamed. Examples include previous suite dropouts Google V8, Dromaeo JavaScript, and now SunSpider as well.
Whenever the results of benchmarks that test the same thing conflict, more weight is given to tests with a better rating when creating the analysis tables.
The table below lists all 41 of the tests currently in our test suite (along with a version number, where applicable), current rating in the Web Browser Grand Prix, and number of iterations performed:
| Web Browser Grand Prix Test Suite 6.0 | ||
|---|---|---|
| Test Name | Iterations | Rating |
| Performance Tests (32) | ||
| Startup Time: Single Tab | 5 | Core |
| Startup Time: Eight Tabs | 5 | Core |
| Page Load Time: Google | 5 | Core |
| Page Load Time: YouTube | 5 | Core |
| Page Load Time: Yahoo! | 5 | Core |
| Page Load Time: Amazon | 5 | Core |
| Page Load Time: Wikipedia | 5 | Core |
| Page Load Time: eBay | 5 | Core |
| Page Load Time: craigslist | 5 | Core |
| Page Load Time: The Huffington Post | 5 | Core |
| Page Load Time: Tom's Hardware | 5 | Core |
| Peacekeeper | 3 | Core |
| Kraken v1.1 | 3 | Core |
| SunSpider v0.9.1 | 3 | Quarantine |
| Dromaeo DOM | 3 | Core |
| Maze Solver | 5 | Core |
| JSGameBench v4.1 | 3 | Core |
| Asteroids HTML5 Canvas 2D And JavaScript | 3 | Observation |
| GUIMark 2 HTML5 Vector Charting (1 pixel variant) | 5 | Core |
| GUIMark 2 HTML5 Bitmap Gaming | 5 | Core |
| GUIMark 2 HTML5 Text Columns | 5 | Core |
| Psychedelic Browsing | 3 | Core |
| Hardware Acceleration Stress Test | 3 | Dated |
| WebGL FishIE | 5 | Observation |
| WebGL Solar System | 5 | Observation |
| ThoughtsInComputation Particles | 5 | Observation |
| GUIMark2 Flash Vector Charting | 5 | Core |
| GUIMark2 Flash Bitmap Gaming | 5 | Core |
| GUIMark2 Flash Text Columns | 5 | Core |
| Flash Benchmark 2008 | 3 | Core |
| GUIMark Java | 5 | Dated |
| Encog Silverlight | 5 | Dated |
| Efficiency Benchmarks (4) | ||
| Memory Usage: Single Tab | 3 | Core |
| Memory Usage: 40 Tabs | 3 | Core |
| Memory Management: -39 Tabs | 3 | Core |
| Memory Management: -39 Tabs (extra 5 minutes) | 3 | Core |
| Reliability Benchmarks (1) | ||
| Proper Page Loads | 3 | Core |
| Conformance Benchmarks (4) | ||
| HTML5Test.com | 1 | Core |
| CSS3 Selectors Test | 1 | Core |
| Ecma test262 | 1 | Observation |
| Acid3 | 1 | Dated |
You've seen the lineup and toured the track. Now it's off to the races.
Single Tab
The Google home page serves as our test in the single-tab startup time test.
Starting with a single tab, Google Chrome takes the lead, followed closely by IE9. Firefox 6 is the first browser to take more than one second to start up, landing Mozilla a third-place finish. Safari 5.1 ups the game, leap-frogging Opera to take fourth.
In Mac OS X, Apple's own Safari reaches the finish line first, just a fraction of a second behind Chrome's winning Windows 7 time. Google still manages to grab second place on OS X in under one second. Firefox 6 again comes in third at over one second, very close to the Windows 7 run. Opera still places last.
Eight Tabs
We used the Top Eight websites (according to Quantcast) for our eight-tab startup time test. These sites include: Google, Facebook, YouTube, Yahoo!, Twitter, MSN, Amazon, and Wikipedia.
When starting with eight tabs, it's Opera that really shines, finishing in just under 1.5 seconds. IE9 comes in second place. Chrome 13 places third at just under four seconds, while Firefox 6 takes fourth at 4.25 seconds. Safari falls far behind the pack at just under eight seconds.
Using Lion doesn't change the finishing order one bit. Opera still turns in the best time. Chrome takes second place at 4.25 seconds, Firefox is in third with 4.5, and Safari again places last with 5.5 seconds.
All of the third-party Web browsers start up in approximately the same time, or slightly slower, in OS X. Safari is the exception. We start to see that Safari has significant advantages on its native platform, starting up about 25% faster.
The newer versions of Safari and Firebox seem to have helped Apple and Mozilla's startup times. However, with Chrome 13, the times are higher than they were in version 12.
Our page load time testing was recently overhauled. We modified the timer script to render test pages at 1080p instead of the previous, netbook-friendly resolution.
The line-up is expanded from five webpages to nine. Facebook and MSN are gone. Google, YouTube, and Yahoo! all remain, but are now up to date with the most recent home page layouts. Added to the testing is Amazon, Wikipedia, eBay, Craigslist, The Huffington Post, and good old Tom's Hardware.
In order to better reflect real-world browsing, we're not using the home pages for Amazon, Wikipedia, eBay, or Craigslist. Instead, we're using the page for Computer Parts & Components at Amazon, the Wikipedia page for Tom's Hardware, an eBay Motors search for Cadillac DeVille, and the New York City page on Craigslist.
Like the conformance benchmarks and GUIMark2 tests, page load times in WBGP6 will be averaged into a composite score, although a detail view is still provided.
Page Load Time Detail: Windows 7
The chart below shows how each of the five Windows 7 Web browsers perform on each of the nine test webpages.

Page Load Time Detail: Mac OS X Lion
This chart contains the complete detail view of the four Web browsers in Mac OS X Lion.
Page Load Time Composite
The average time each Web browser takes to load all of the test pages on each platform is displayed in the chart below.

As you can see, Chrome 13 takes the top spot in both Windows 7 and OS X Lion. In fact, Chrome 13 on OS X 10.7 has the fastest average page load time overall. Whenever this occurs, we'll change the regularly green bar to red in order to highlight the existence of a performance advantage on OS X. Safari 5.1 falls a very close second, performing relatively similarly on both Microsoft's and Apple's operating systems. Internet Explorer 9 is a third-place finisher. Opera earns fourth place on Windows, but fifth on OS X. Likewise, Firefox loses in Windows 7, but takes fourth in Lion.
Peacekeeper

The placing in Windows 7 for Peacekeeper is: Chrome, Opera, Safari, IE, and Firefox. Safari 5.10 for Windows surges ahead of its 5.05 score by nearly 1700 points. Other than Safari, the Peacekeeper scores from Windows are virtually the same as we saw in WBGP5, with minor improvements favoring the other four browsers.
The placing in OS X is slightly different, with the number two spot going to Safari, and Opera taking third.
| JavaScript |
Kraken

The Kraken results are nearly the same as what we saw in WBGP5, only Chrome shows significant improvement, shaving 450 milliseconds off its already-winning time. Firefox again places second, with Opera in third, IE9 in fourth, and Safari picking up the rear.
The story changes in OS X. Safari earns third place on its native platform, beating Opera and shaving 5000 milliseconds off its Windows 7 score.
SunSpider

The finishing order is again the same as it was in WBGP5, with IE9 taking the gold, followed by Firefox, Opera, Chrome, and then Safari. The only notable change is that Safari 5.10 takes an additional 40 milliseconds over version 5.05.
The order completely changes in OS X. Firefox takes first place, followed by Safari and Chrome in a near-tie for third, with Opera coming in last.
| DOM |
Dromeao DOM

As usual, Opera beats everyone with ease in DOM testing. Firefox 6 earns second place with a score 100 runs per second higher than Firefox 5. Safari 5.10 adds about 100 runs over Safari 5.05, jumping from fourth up to third. Chrome 13 finishes fourth, with a score 200 runs less than Chrome 12. Internet Explorer 9 again winds up in last place, scoring only half of what Opera did.
All is not well for Opera on Mac OS X. The Norwegian Web browser that had been tearing up the competition in DOM benchmarks falls to nearly half of its Windows 7 score, barely taking third place.
| CSS |
Maze Solver
We tweaked the methodology for testing Maze Solver. The maze is now cleared and re-created for each iteration of this test, and more iterations are used for generating results.

Chrome still holds the lead, but Safari jumps from fourth place to second. Opera drops to third, and IE9 drops to fourth due to Safari 5.1's higher score. Firefox is again far behind the competition in this test, with miserable scores exceeding one minute. All of the Web browsers perform noticeably worse in Mac OS X than in Windows 7.
| Flash |
GUIMark 2 Flash
The chart below lists the results of all three GUIMark2 Flash tests for each of the five Windows 7 Web browsers.

The next chart has the same data for the four OS X browsers.

This final chart hosts the cross-platform composite scores, which average the results of the three GUIMark2 Flash tests.

Microsoft's Internet Explorer 9 achieves the highest score in the GUIMark2 Flash Windows 7 testing, closely followed by Opera and Safari. Firefox 6 only earns a fourth-place victory, with Chrome 13 in last place.
The highest score from all of our GUIMark2 Flash testing goes to Opera on OS X. Chrome 13 manages to rocket into second place on the Hackintosh, while Safari 5.1 and Firefox 6 practically tie for last.
Flash Benchmark 2008

Safari 5.1 nabs the highest score in Flash Benchmark 2008 in Windows 7, followed by Microsoft's IE9. Opera takes the third-place spot, with Chrome 13 and Firefox 6 far in tow.
Opera for Mac achieves the highest score in all of our testing, more than 10 000 points ahead of Safari's Window 7 victory. Safari takes second in Lion, closely followed by Firefox 6, then Chrome 13.
It appears that Opera on OS X is the Flash performance king.
| Java |
GUIMark Java

Firefox 6 leads in GUIMark Java under Windows 7. Mozilla is followed by Chrome 13 with a 5 FPS disadvantage. Opera takes third, followed by Safari 5.1 barely ahead of IE9.
In OSX, Opera takes the lead, followed by Chrome, Safari, and Firefox.
This benchmark shows Windows 7 to have a significant advantage over Mac OS X in Java performance.
| Silverlight |
Encog Silverlight

The Encog Silverlight scores are pretty close together. In Windows 7, Opera 11.50 holds onto first place. Version 13 allows Chrome to move from third place to second. Firefox 6 catapults Mozilla up one spot, from fourth place to third. Microsoft Internet Explorer 9 drops from second place to fourth, and Safari remains in last place.
The OS X scores are all very far behind the Windows 7 results (by about double). Firefox 6 takes the lead in Lion, followed by Safari 5.1. Opera settles into third place on the Apple platform, and Chrome finishes last.
Silverlight is the second plug-in where Windows has a serious edge on OS X, but this wasn't entirely unexpected. Afterall, Silverlight is Microsoft's own attempt at a Flash-killer.
JSGameBench

The placing in Windows 7 remains the same as what we saw in WBGP5. Mozilla Firefox again demolishes the competition in the Facebook JSGameBench test, retaining first place. IE9 also holds onto its solid second-place finish. Chrome lands in a distant third, with Safari and Opera even further behind in fourth and fifth (respectively).
In OS X, there is no Internet Explorer. Also, Chrome and Firefox for Mac lack HTML5 hardware acceleration. But Safari includes it, making it the winner, as well as the only viable HTML5 hardware-accelerated option for Mac OS X. Chrome 13 finishes a distant second, followed by Firefox 6. Opera, the only Web browser with absolutely no HTML5 hardware acceleration support, predictably ends up in last place.
GUIMark 2 HTML5
The chart below shows how each of the five Web browsers performs in the three GUIMark2 HTML tests running in Window 7.

This chart has the same data, but for the four Mac OS X Lion browsers.

The final chart is our cross-platform composite score, achieved by averaging the results of the three GUIMark2 HTML5 tests.

The Windows 7 placing in the GUIMark2 composite score remains mostly the same as what we saw in WBGP5. Firefox 6 is again the Windows 7 winner, followed by IE9 at two frames per second. Chrome 13 is still in third place, 10 FPS behind the leader. This time, Safari 5.1 is barely ahead of Opera in fourth place, both at just over 30 FPS.
Safari on OS X earns the highest score in our competition, archiving just under 50 FPS. Firefox 6 coasts into second place at over 32 FPS, while Opera steals third with just under 30 FPS. Chrome 13 places last on OS X, barely over 26 FPS.
Asteroids HTML5 Canvas 2D

On Windows, Internet Explorer 9 remains top dog in the Asteroids HTML5 Canvas 2D benchmark. Safari 5.1 offers a marked improvement over 5.05 in this test, moving from fourth all the way up to second place. Google's browser falls close behind Apple's in third, while Opera trails in fourth, and Firefox 6 brings up the rear with a more pitiful sub-30 FPS score.
Once again, the highest score in this benchmark occurs on OS X. Safari 5.1 obliterates the other Web browsers, earning just shy of 80 FPS. Second place on OS X goes to Opera 11.50, with 36 FPS. Third place goes to Chrome with 33 FPS, while Firefox 6 again comes up short at 22 FPS.
Safari running on OS X really gives Firefox and IE9 on Windows 7 a run for their money with regard to HTML5 performance.
| HTML5 Hardware Acceleration |
Psychedelic Browsing

The scores and placing of this benchmark are similar to results generated for WBGP5: Firefox, IE9, Chrome, Opera, and Safari. The difference is that Chrome 13 earns a little over 1.5 points more than version 12. On Mac OS X, Safari's weak score of eight points rockets up to more than 1800, allowing Apple's browser to come in less than two points behind Microsoft's Internet Explorer in Windows.
Hardware Acceleration Stress Test

Only browsers supporting HTML5 hardware acceleration achieve good scores in this benchmark. In fact, they max it out. First place on Windows is a tie between Firefox 6 and Internet Explorer 9. Second place goes to Chrome, followed closely by Opera, with the Windows version of Safari 5.1 falling behind in last place.
In OS X Lion, Safari 5.1 maxes out the benchmark at 60+ FPS, creating a cross-platform tie for first with the Windows versions of Firefox 6 and IE9. Opera lands in second place on OS X, with Chrome and Firefox in the distance.
For those wanting to utilize HTML5 hardware acceleration today, Firefox 4+ and IE9 are still the only options for Windows users, while Safari is the one and only hope on a Mac.
| WebGL |
We've completely changed the lineup of WebGL benchmarks. We replaced the Kronos Particles test with a variation from ThoughtsInComputation. Its version has more going on, which lowers the frame rate so the test cannot be maxed out. It also has the option of adding more particles and other graphics, ensuring that this test is scalable for some time to come.
The WebGL Aquarium from the Chrome Experiments site is being replaced with WebGL Solar System for much the same reason (WebGL Solar System is more taxing and configurable than WebGL Aquarium).
Finally, we added a Mozilla-created WebGL variant of the famous FishIE HTML5 hardware acceleration test from the IE Test Drive site. All three tests yield lower frame rates, are configurable, have higher FPS limits, and provide steadier FPS counts.
ThoughtsInComputation Particles

Unlike the Khronos Particles benchmark, the ThoughtsInComputing variant puts Firefox ahead of Chrome by a significant margin: 62 FPS versus 39. The same result is seen in Mac OS X. However, Firefox only manages to produce 46 FPS on the Apple platform.
WebGL Solar System

On Windows 7, Chrome 13 beats Firefox 6 in the WebGL Solar System demo, 24 FPS to 16. The placing is reversed on OS X, with Mozilla besting Google by less than two FPS.
WebGL FishIE

Chrome 13 regains its lead in the WebGL version of the FishIE test, beating Firefox 6 by 10 FPS. The punishment gets worse for Mozilla on Mac OS X, where Chrome beats Firefox 53 FPS to 27.
The edge in WebGL performance goes to Chrome 13 in Windows 7 and Firefox 6 in OS X Lion.
Light Load

The placing for single-tab memory usage remains the same as what we saw in WBGP5, with IE9 in the lead, closely followed by Chrome, Safari, Firefox, and Opera.
On Mac OS X, all of the memory usage totals are higher than Windows 7, and the order is re-arranged. The top spot goes to Opera. Second place goes to Safari 5.1, with just under 100 MB. Chrome takes third, while Firefox uses nearly 130 MB to finish last.
Heavy Load
For the first time in the WBGP, Safari manages to fully load all 40 tabs at the same time without freezing. Previously, we had to open each of the 40 tabs individually in Apple's browser.
In WBGP5, we also reported some odd behavior from Internet Explorer 9 in these tests. Microsoft's browser would display certain white space-heavy pages in all black, and sometimes freeze entirely. This time around, we used the WHQL-certified AMD graphics driver provided by Windows Update instead of the company's website, and this phenomenon did not occur. After running through the full benchmark suite, we installed the latest Catalyst suite and retested. Sure enough, the wonky behavior returns with the Catalyst driver installed. The video below shows the white space issue on Wikipedia at 1:20:
The next video show the complete freeze-up at 1:50:
Now on to this story's 40-tab memory test:
The order again looks similar to what we saw last time. Safari uses the least amount of RAM to open 40 tabs. Firefox retains its second-place status, consuming three-quarters of a gigabyte. Opera is not far behind in third, still under 800 MB. Chrome 13 drops to fifth, allowing Microsoft's IE9 to jump into fourth.
In OS X, Safari 5.1 still uses the least memory of any Web browser, though nearly 100 MB more than it does in Windows. Opera moves to second place with 40 tabs open in OS X, hovering just over the 800 MB line. Firefox 6 uses nearly 1.2 GB to take third, while last-place finisher Chrome 13 uses a a whopping 1.8 GB.
We're tweaking the memory management testing methodology. Instead of recording the memory totals five minutes after closing 39 of the 40 tabs, we record the totals immediately. We also throw out the 10 minute totals. Ten minutes is just too long to wait for an application to return unused memory to the operating system. The 10 minute figure is replaced by five minutes.
-39 Tabs

While Chrome demonstrates heavy memory usage, it manages that memory superbly, giving back all but 75 MB to Windows immediately after closing 39 of the 40 tabs. IE9 is in second place, dropping its memory usage to just under 200 MB. Apple Safari takes third place with 325 MB. Firefox 6 holds onto nearly 400 MB, landing itself in fourth place, while Opera actually hits 400 MB, ensuring the Norwegian Web browser a last-place finish.
Chrome also wins in OS X, dropping from 1.8 GB to only 200 MB, which is still lower than any other browser's total in Windows 7. Opera comes in second with nearly 520 MB (not quite half of its 40-tab total). Surprisingly, Apple Safari only places third at 620 MB, just 110 MB less than the 40-tab count. Firefox 6 is the worst performer for immediate memory release on OS X, still using over 1 GB of RAM.
Five More Minutes

After an additional five minutes, Chrome remains firmly in the lead, now only using 50 MB. IE9 also retains second place, giving back another 70 MB to close at 130 MB. Firefox 6 jumps to third, dropping from nearly 400 MB to just over 250. Safari 5.1 loses an extra 60 MB, but not enough to keep Apple out of fourth place. Opera still uses the most memory, only releasing 40 MB more to Windows.
Chrome leads again after five minutes in OS X, dropping down another 70 MB to weigh in at 130 MB. Opera is still in second place at 510 MB, only dropping another 10 MB. Safari remains in third with 560 MB, just 55 MB less. Firefox 6 finally releases a significant amount (300 MB) of memory back to OS X. Unfortunately, that still leaves nearly 750 MB tied up.
Chrome and Internet Explorer are the clear winners in light-load memory usage and management. However, they also eat up the most memory under heavy load. Firefox and Safari are pretty much equals in Windows 7, and Opera still performs the worst when it comes to memory efficiency. However, on OS X it's Firefox that loses big, while Opera and Safari are close together in the middle. Overall, the Windows 7-based browsers have a substantial edge over their Mac OS X Lion counterparts in this discipline.
Proper Page Loads

Opera manages to come out on top for page loading reliability on both Windows and OS X. Strangely, Safari 5.1 for Windows comes in second, while it loses big time on its native platform. Firefox 6 grabs second for OS X and third for Windows, representing a major drop from Firefox 5. Google comes in second-to-last on the Microsoft and Apple ecosystems. Microsoft Internet Explorer 9 comes in dead last on its one and only supported platform, and manages to have the worst overall reliability.
Like the memory testing, reliability appears to be better on Windows 7 than Mac OS X Lion across the board.
| HTML5 |
HTML5Test.com

In the HTML5Test, Chrome again leads the pack with a score of 340. Firefox 6 surges ahead of Opera to take second place, scoring nearly 50 points more than Firefox 5. Opera drops to third place. While Safari still places fourth, version 5.10 adds 65 points and two bonus points to its version 5.05 score. IE9 remains in last place, though it now scores 11 points higher than it did in WBGP5.

In OS X, Chrome is still in the top spot, but with one more point than in Windows 7. Firefox remains in second, while Opera drops to fourth, both with the same score as in Windows. Safari jumps to third, gaining 41 regular and six bonus points over the Windows edition.
| CSS3 |
CSS3 Selectors Test

With Chrome updating from 12 to 13, all five Web browsers now earn a perfect score of 574 in the CSS3 Selectors Test. Since they all pass with 100%, a comparison is pretty meaningless. So, we're officially retiring this test from the WBGP. The results will not be factored in to the Conformance Composite.
| JavaScript |
test262
The Ecma test262 benchmark is replacing the Sputnik JavaScript benchmark. Sputnik is one of the many casualties from the scuttling of Google Labs. Fortunately, the Sputnik developers gave their code to Ecma, another group interested in JavaScript conformance, so Sputnik is now a part of test262. 
The results of test262 in Windows largely differ from the Sputnik scores in WBGP5. In test262, Firefox earns the top spot, followed by IE9, and then Chrome. Safari takes fourth and Opera completely tanks the test. The only difference between the test262 scores in Windows and OS X is that Safari earns one extra point on its native platform.
| DOM |
Acid 3

The Acid 3 scores remain the same from WBGP3. Firefox 6 does not change Mozilla's score of 97. And Microsoft still holds a 95. The other browsers pass completely.
| Conformance Composite |
The table below displays the grade each Web browser earns for overall standards compliance. These figures are achieved by converting the test scores to a percentage of 100, then averaging the aggregate.

Google Chrome is still the conformance king with version 13, earning a B+ score of 89.6. Firefox 6 moves up to second place, earning a B rating and score of 87.4. Safari also earns a B by placing third with a Windows 7 score of 82.6 and 85.6 in OS X Lion. Opera 11.50 falls from second place to fourth as a result of our switch from Sputnik to Ecma's test262, landing it a solid C. Predicatbly, the HTML5Test shoots IE9 out of the sky; Microsoft's browser only achieves a 74.5, earning it a C grade.
Windows 7 Placing Tables
| Performance Placing | |||||
|---|---|---|---|---|---|
| 1st | 2nd | 3rd | 4th | 5th | |
| Chrome | 5 | 2 | 5 | 4 | 1 |
| Firefox | 5 | 4 | 1 | 2 | 4 |
| Internet Explorer | 4 | 6 | 2 | 3 | 2 |
| Opera | 3 | 2 | 6 | 3 | 3 |
| Safari | 1 | 3 | 3 | 5 | 6 |
When we only take raw placing in the performance tests into account, Chrome 13 appears to have a slight lead over Firefox 6 and IE9.
| WebGL Placing | ||
|---|---|---|
| 1st | 2nd | |
| Chrome | 2 | 1 |
| Firefox | 1 | 2 |
The WebGL testing on Windows 7 gives the edge to Google Chrome 13 over Mozilla Firefox 6.
| Efficiency Placing | |||||
|---|---|---|---|---|---|
| 1st | 2nd | 3rd | 4th | 5th | |
| Chrome | 2 | 1 | 1 | ||
| Firefox | 1 | 1 | 2 | ||
| Internet Explorer | 1 | 2 | 1 | ||
| Opera | 1 | 3 | |||
| Safari | 1 | 2 | 1 | ||
Chrome appears to win here, followed closely by IE9, then Safari, Firefox, and Opera.
| Reliability Placing | ||||
|---|---|---|---|---|
| 1st | 2nd | 3rd | 4th | 5th |
| Opera | Safari | Firefox | Chrome | Internet Explorer |
With a single test, the reliability placing is pretty straightforward.
| Conformance Placing | |||||
|---|---|---|---|---|---|
| 1st | 2nd | 3rd | 4th | 5th | |
| Chrome | 3 | 1 | |||
| Firefox | 2 | 2 | |||
| Internet Explorer | 1 | 1 | 1 | 1 | |
| Opera | 2 | 1 | 1 | ||
| Safari | 2 | 2 | |||
The overall conformance placing skews the results of our composite score. Despite the placing table, Chrome 13 sweeps the conformance testing in Windows 7, followed closely by Firefox 6. Safari 5.1 earns third, Opera drops to fourth, and IE9 come in last.
| Total Placing | |||||
|---|---|---|---|---|---|
| 1st | 2nd | 3rd | 4th | 5th | |
| Chrome | 10 | 3 | 6 | 5 | 2 |
| Firefox | 7 | 7 | 3 | 4 | 4 |
| Internet Explorer | 6 | 9 | 3 | 4 | 4 |
| Opera | 6 | 2 | 8 | 3 | 7 |
| Safari | 4 | 4 | 5 | 8 | 6 |
The total placing clearly has Chrome in the lead, followed by Firefox. Internet Explorer grabs third, while Opera takes fourth, and Safari for Windows places last.
Mac OS X Lion Placing Tables
| Performance Placing | ||||
|---|---|---|---|---|
| 1st | 2nd | 3rd | 4th | |
| Chrome | 3 | 5 | 4 | 5 |
| Firefox | 3 | 2 | 6 | 6 |
| Opera | 5 | 2 | 5 | 5 |
| Safari | 6 | 8 | 2 | 1 |
The Mac OS X Lion-based performance results offer a stark contrast to the Windows 7 table. In OS X, Safari holds a clear advantage, followed by Opera. Windows 7-winner Chrome only places third on the Apple platform, while Firefox takes fourth.
| WebGL Placing | ||
|---|---|---|
| 1st | 2nd | |
| Chrome | 1 | 2 |
| Firefox | 2 | 1 |
The WebGL placing is also reversed in OS X, with Firefox beating Chrome.
| Reliability Placing | |||
|---|---|---|---|
| 1st | 2nd | 3rd | 4th |
| Opera | Firefox | Chrome | Safari |
Strangely, Safari falls short of its Windows 7 numbers in reliability placing for OS X.
| Conformance Placing | ||||
|---|---|---|---|---|
| 1st | 2nd | 3rd | 4th | |
| Chrome | 3 | 1 | ||
| Firefox | 2 | 2 | ||
| Opera | 2 | 2 | ||
| Safari | 2 | 2 | ||
Conformance placing in Lion is nearly the same as in Windows 7, just with one less browser.
| Total Placing | ||||
|---|---|---|---|---|
| 1st | 2nd | 3rd | 4th | |
| Chrome | 8 | 6 | 6 | 6 |
| Firefox | 5 | 5 | 7 | 9 |
| Opera | 9 | 5 | 5 | 7 |
| Safari | 9 | 9 | 6 | 2 |
The OS X total placing shows Safari to be the clear winner, with Opera and Chrome duking it out for second. Firefox is most definitely the last-place finisher.
Windows 7 Analysis Table
| Winner | Strong | Acceptable | Weak | |
|---|---|---|---|---|
| Performance Benchmarks | ||||
| Startup Time | Opera | Chrome, Internet Explorer | Firefox | Safari |
| Page Load Time | Chrome | Internet Explorer, Safari | Firefox, Opera | |
| JavaScript | Chrome | Firefox | Internet Explorer, Opera | Safari |
| DOM | Opera | Firefox | Safari | Chrome, Internet Explorer |
| CSS | Chrome | Opera, Safari | Internet Explorer | Firefox |
| Flash | Internet Explorer | Opera, Safari | Chrome, Firefox | |
| Java | Firefox | Chrome, Internet Explorer, Opera, Safari | ||
| Silverlight | Opera | Chrome, Firefox | Internet Explorer | Safari |
| HTML5 | Internet Explorer | Firefox | Chrome, Opera, Safari | |
| HTML5 Hardware Acceleration | Firefox | Internet Explorer | Chrome, Opera, Safari | |
| WebGL | Chrome | Firefox | Internet Explorer, Opera, Safari | |
| Efficiency Benchmarks | ||||
| Memory Usage: Light | Internet Explorer | Chrome, Safari | Firefox, Opera | |
| Memory Usage: Heavy | Safari | Firefox, Internet Explorer, Opera | Chrome | |
| Memory Management | Chrome | Internet Explorer | Firefox, Safari | Opera |
| Reliability Benchmarks | ||||
| Proper Page Loads | Opera | Safari | Firefox | Chrome, Internet Explorer |
| Conformance Benchmarks | ||||
| HTML5 | Chrome | Firefox | Opera, Safari | Internet Explorer |
| CSS3 | All 5 | |||
| JavaScript | Firefox | Internet Explorer, Chrome | Safari | Opera |
| DOM | Chrome, Opera, Safari | Firefox, Internet Explorer | ||
And for OS X...
Mac OS X Lion Analysis Table
| Winner | Strong | Acceptable | Weak | |
|---|---|---|---|---|
| Performance Benchmarks | ||||
| Startup Time | Opera | Chrome, Firefox, Safari | ||
| Page Load Time | Chrome | Safari | Firefox, Opera | |
| JavaScript | Chrome | Firefox | Safari | Opera |
| DOM | Firefox | Safari | Chrome, Opera | |
| CSS | Chrome | Safari | Opera | Firefox |
| Flash | Opera | Chrome, Firefox, Safari | ||
| Java | Opera | Chrome, Firefox, Safari | ||
| Silverlight | Firefox | Safari | Safari | Chrome, Opera |
| HTML5 | Safari | Chrome, Firefox, Opera | ||
| HTML Hardware Acceleration | Safari | Chrome, Firefox, Opera | ||
| WebGL | Firefox | Chrome | Opera, Safari | |
| Efficiency Benchmarks | ||||
| Memory Usage: Light | Opera | Chrome, Safari | Firefox | |
| Memory Usage: Heavy | Safari | Opera | Chrome, Firefox | |
| Memory Management | Chrome | Opera, Safari | Firefox | |
| Reliability Benchmarks | ||||
| Proper Page Loads | Opera | Firefox | Chrome, Safari | |
| Conformance Benchmarks | ||||
| HTML5 | Chrome | Firefox, Safari | Opera | |
| CSS3 | All 4 | |||
| JavaScript | Firefox | Chrome | Safari | Opera |
| DOM | Chrome, Opera, Safari | Firefox | ||
Let's crown the champion of Web Browser Grand Prix 6.
We're handing out two championship prizes: one for Windows and another for Mac OS X.
Your Windows Champion
Version 13 once again earns Google Chrome the Web Browser Grand Prix championship. Chrome's sheer number of wins nearly discounts its weaknesses.

With only one weakness and the highest number of non-winning strong scores, Mozilla Firefox is once again our runner-up. A bit of trouble in CSS performance, as well as a high number of merely acceptable scores, are all that hold back Firefox from taking the gold. Although the new rapid development cycle may be hurting Firefox in our reliability benchmarks, it also allows Mozilla to keep pace with Google Chrome on the performance front. Firefox 7 is supposed to bring improvements to both JavaScript and memory usage/management, which are two key areas that might allow Firefox to pull into a first-place finish.
Microsoft Internet Explorer 9 claims third place, though it is falling further behind the more quickly updated Chrome and Firefox builds. Its scores in Flash, HTML5, and memory management are keeping IE9 on the medal stand.
Opera takes fourth place, but is still very close to unseating IE9 for third. Poor memory management, plus a lack of HTML5 hardware acceleration and WebGL support put a hurt on the Norwegien Web browser. A large number of non-standout scores isn't helping, either.
Safari crawls into last place with the least number of wins and most losses. Like Opera, no HTML5 hardware acceleration and WebGL support make it hard for Safari to compete. But Safari for Windows is barely a shell of its OS X counterpart.
Mac OS X Champion
On its native platform, Safari is definitely no slouch. In fact, the performance of Safari 5.1 in OS X Lion matches that of Firefox 6 in Windows 7.

Chrome takes second place on Lion, though its performance isn't even close to what we saw in Windows 7. A lack of HTML5 hardware acceleration, hoggish memory usage, and poor reliability hold back Google's browser here.
Opera takes third place on OS X, and its cross-platform scores are really close, earning Opera the distinction of providing a steady experience across multiple platforms. This browser suffers the same issues on OS X as in Windows 7, with the added problems of lowered JavaScript and DOM performance, as well as slower page load times.
Firefox 6 is pretty close behind Opera. Like Chrome, though, its Windows 7 build far out-paces the OS X version. Firefox's problems in OS X mirror what it suffers in Windows, with the addition of memory problems found in OS X.
Operating System Comparison
Mac OS X Lion is a beauty to behold, and its benefits aren't just skin-deep. The score for Safari 5.1 on OS X is really close to Chrome 13 running on Windows 7, and it might even beat Firefox 6 for Windows. So, if you throw Safari 5.1 for OS X into the regular Windows 7 mix, Apple takes or shares second place! It appears that your Mac friends were right afterall, so stop hassling them.
Also, remember these tests were not conducted on an actual Apple-branded Mac system. Based on what we saw in the results of WBGP2: Linux, we really didn't expect the OS X scores to get so close to the best from Windows. With such a slim margin of victory favoring Chrome in Windows (versus Safari on OS X), it is entirely within the realm of possibility that running these tests on a genuine Apple rig could tip the scales in favor of Safari (or back the other way). A more in-depth Hackintosh versus Macintosh comparison would be needed to confirm one way or the other.
As it stands, a Mac-based browser matched or beat the best score from Windows-based browsers in 10 out of 29 scored tests. In fact, even on a Hackintosh, Mac OS X is capable of providing better results than Windows 7 in Flash, HTML5, WebGL, and the ever-important page load times. Better standards compliance in Safari and Chrome for Mac even tip conformance in favor of OS X.
On the flip side, that leaves 19 tests in which a Windows 7-based browser provides the best score. For now, Redmond can still claim Web browsing supremacy on the desktop. But that edge is eroding, and you can bet Cupertino won't quit trying to usurp its competition.
Follow Adam on Twitter.





