backtop


Print 106 comment(s) - last by chick0n.. on Aug 13 at 11:31 AM


Futuremark, makers of 3DMark, have launched PeaceKeeper, the first comprehensive browser benchmarking suite.

The results are intriguing -- Apple's Safari and Google Chrome are the top browsers speed-wise. Windows 7 beats Windows Vista with similar hardware and the same browser version, but still lags behind OS X. And the hardware on the system, particularly the CPU dramatically influences results.  (Source: PeaceKeeper)
The makers of 3DMark take on the web

Mozilla, Apple, Google, Opera, and Microsoft all claim the same thing -- that their browser is the fastest.  So who is telling the truth? 

Until recently, that was a tough question to answer.  Sure there was a handful of free tests, such as SunSpider's JavaScript benchmark, Celtic Kane's JavaScript benchmark (a bit older), or the crude CSS benchmark posted on HowToCreate.  However, these tests all had one thing in common -- they did not fully test the browser's speed across a variety of rich-content standards.

FutureMark, makers of the popular 3DMark hardware benchmarking software, know all about how to provide a cohesive benchmarking suite.  So amid ongoing work on PCMark, 3DMark, and even a video game (Shattered Horizon), FutureMark decided to try to settle the browser bragging match once and for all by releasing a comprehensive test suite.

The result is the Peacekeeper benchmarking suite, which is now available as a free online applicationDailyTech had the pleasure of talking with FutureMark President Oliver Baltuch and learning more about this exciting new test, as well as some of its more interesting results.

Currently, the test covers page rendering using everything from HTML 5.0 to CSS.  It also includes a number of tests to gauge JavaScript performance with typical algorithms such as encrypts, filters, parses, sorts, and array manipulation.  DOM performance is also tested.  Currently, flash performance tests are not included.  Mr. Baltuch says that a second version is in the works "that would include more of the HTML 5.0 and something that works to measure flash performance as well as possibly power performance."

The core set of tests now simulate conditions that might be encountered on content-heavy pages such as YouTube, Facebook, or Meebo.  The results are intriguing.  For the same browser version tested across different OS implementations running on virtually the same hardware Mr. Baltuch states, "The Apple [computers] tend to be 5-10 percent faster [than Windows machines].  We believe that is based on the middleware of the OS.  We believe that it is more streamlined."

While OS X holds the speed lead, Mr. Baltuch noted that Windows 7 featured substantial gains over Windows Vista.  However, Internet Explorer 8.0, Microsoft's flagship browser remains much slower than the industry's speed-leading browsers -- Google's Chrome and Apple's Safari.  Mr. Baltuch comments, "Windows 7 is faster than Vista is.  But Internet Explorer 8.0 is 5 times slower than Chrome and Safari."

Despite Chrome and Safari being the fastest, Mr. Baltuch says he uses Opera 10.0 beta 2 for browsing because he likes the user interface and features the best.  He also praised Mozilla users for their very helpful feedback.  He says that some browser makers have embraced his company's efforts, while others remain standoffish.

He states, "Some seem to be happy.  The Norwegians (Opera) seem to hate us at the moment (FutureMark is Finnish).  The guys at Microsoft have been silent so far."

For those interested, Peacekeaper is funded by advertising revenue and is also funded by sales of consulting services to large businesses.  Also, the benchmark has helped to increased the adoption of FutureMark's paid products like PCMark, according to Mr. Baltuch.

A couple of final interesting notes.  Mr. Baltuch says that the test is compatible with the popular Apple iPhone.  He also notes that there is a large difference between web performance, going from a netbook processor (such as the Intel Atom) to a high-end desktop processor (such as the Intel Core i7).  This is due to the CPU's pivotal role in running increasingly heavy web applications.  He also says that for users looking to try something off the typical beaten path, the Epiphany (browser packaged with Gnome Linux), Shiira (Webkit-based), and Midori (also Webkit-based) browsers all provide relative good Linux browsing performance.

Mr. Baltuch says his company loves feedback, so feel free to head over to the Peacekeeper page, check it out, and leave your comments.



Comments     Threshold


This article is over a month old, voting and posting comments is disabled

Self-Contradiction
By MrDiSante on 8/10/2009 2:49:03 PM , Rating: 5
quote:
Why are there no Linux/Mac/Other lists of the top scores? Currently, our system scan technology only works on Windows PCs. Without knowing the hardware details of a system, the performance score alone is not enough to create a list of top scores.

and
quote:
The Apple [computers] tend to be 5-10 percent faster [than Windows machines]


Now let's see if I have this right: you don't know the hardware specs of the machines running the tests, but you go ahead and claim Macs are faster anyhow? Yeah, Windows tends to run on Atom processors among other things and yeah an Atom is gonna be a lot slower than C2D/Q. Alternatively, if you're testing on the same Mac then it's still hardly any wonder - the Windows drivers for Macs are notoriously shoddy.




RE: Self-Contradiction
By johnsonx on 8/10/09, Rating: -1
RE: Self-Contradiction
By StevoLincolnite on 8/10/2009 4:17:55 PM , Rating: 5
And then we get into the numbers game.

Because Dailytech failed to provide hard numbers yet again and instead gave us a percentage we have no real hard facts on the actual performance differences, it's merely speculation.

Plus Percentages can skew results to make one platform seem significantly superior, for all we know, there was only a single point difference, and it is merely a margin of error. (Where the tide can turn and Windows might end up 10% faster just by running the test again).

I find it poor and frustrating how there was zero hard numbers included in the article.


RE: Self-Contradiction
By Alexstarfire on 8/10/2009 5:09:45 PM , Rating: 2
Just look at the author of the article and you'll know why.


RE: Self-Contradiction
By Mojo the Monkey on 8/11/2009 1:42:25 PM , Rating: 2
Its a blog posting, not an article. Give Mick a break.


RE: Self-Contradiction
By sebmel on 8/10/2009 3:28:42 PM , Rating: 2
I believe that the Peacekeeper FAQ you are referring to is talking about their online test, which I can confirm doesn't work on a Mac, online.

I suspect one will find that they did not initially test their own software online.

I understand your point but I doubt that guys with the intelligence to write the browser test code are so unintelligent that they don't understand that processors of different speed will produce different results. They've used a hackintosh or a dual booting Mac.


RE: Self-Contradiction
By MrDiSante on 8/10/2009 6:25:53 PM , Rating: 2
I very much doubt that they would have used a hackintosh since that would mean entering a legal quagmire that no one wants to get into. As for a dual-booting Mac, I did address that in my previous post: "the Windows drivers for Macs are notoriously shoddy".


"If they're going to pirate somebody, we want it to be us rather than somebody else." -- Microsoft Business Group President Jeff Raikes














botimage
Copyright 2014 DailyTech LLC. - RSS Feed | Advertise | About Us | Ethics | FAQ | Terms, Conditions & Privacy Information | Kristopher Kubicki