Safari 4 - "The world's fastest browser" ... NOT!

Safari 4 - "The world's fastest browser" ... NOT!

Summary: At yesterday's WWDC 09 keynote speech Apple announced the release of Safari 4 web browser for Windows and Mac. Apple claims this browser is "the world’s fastest browser" ... let's find out.


At yesterday's WWDC 09 keynote speech Apple announced the release of Safari 4 web browser for Windows and Mac. Apple claims this browser is "the world’s fastest browser" ... let's find out.

I'm always suspicious of all claims that are along the lines of "world's blankiest blank," especially when those claims come from Apple.

So, is Safari 4 the fastest browser? Let's test out Apple's claim by pitting it against Google Chrome.

I'm using my standard test bed - QX9770 Core 2 Extreme running at 3.2GHz, with 2GB or RAM on a fully up-to-date Windows Vista Ultimate 32-bit platform.

I'm putting Safari 4 against Google Chrome because that's currently the fastest browser on the block. If Safari 4 can beat that, it earns the title of "world's fastest browser." If it can't beat Chrome, then the claim is puff.

As usual, two tests - SunSpider JavaScript, and V8 benchmarks.

The results -->



While it's clear that Safari 4 is fast, and it's almost caught up with Chrome in the V8 benchmark, Safari 4 is clearly NOT the world's fastest browser.

I replicated the same tests on a number of other systems and came up with similar results. Google Chrome still outperforms Safari 4. The gap is closing, sure, so it's close but no cigar.

Detailed benchmark results -->

Google Chrome

V8 Benchmark

Score: 3623 Richards: 3297 DeltaBlue: 3747 Crypto: 3043 RayTrace: 4006 EarleyBoyer: 5777 RegExp: 1265 Splay: 7441

Sunspider JavaScript Benchmark

============================================ RESULTS (means and 95% confidence intervals) -------------------------------------------- Total:                  802.8ms +/- 2.5% --------------------------------------------

  3d:                   128.4ms +/- 9.2%     cube:                36.8ms +/- 5.6%     morph:               54.6ms +/- 25.7%     raytrace:            37.0ms +/- 2.4%

  access:                71.8ms +/- 11.9%     binary-trees:         4.6ms +/- 14.8%     fannkuch:            27.6ms +/- 19.0%     nbody:               28.4ms +/- 12.2%     nsieve:              11.2ms +/- 9.3%

  bitops:                49.2ms +/- 8.6%     3bit-bits-in-byte:    4.2ms +/- 24.8%     bits-in-byte:         9.8ms +/- 5.7%     bitwise-and:         13.6ms +/- 5.0%     nsieve-bits:         21.6ms +/- 19.3%

  controlflow:            3.4ms +/- 20.0%     recursive:            3.4ms +/- 20.0%

  crypto:                57.4ms +/- 6.2%     aes:                 18.4ms +/- 3.7%     md5:                 21.6ms +/- 8.7%     sha1:                17.4ms +/- 10.8%

  date:                 109.6ms +/- 11.9%     format-tofte:        57.0ms +/- 4.4%     format-xparb:        52.6ms +/- 22.4%

  math:                  72.4ms +/- 4.0%     cordic:              25.8ms +/- 4.0%     partial-sums:        35.8ms +/- 5.2%     spectral-norm:       10.8ms +/- 5.1%

  regexp:                29.6ms +/- 7.0%     dna:                 29.6ms +/- 7.0%

  string:               281.0ms +/- 3.3%     base64:              37.4ms +/- 4.5%     fasta:               42.8ms +/- 5.2%     tagcloud:            52.6ms +/- 3.6%     unpack-code:         91.2ms +/- 6.4%     validate-input:      57.0ms +/- 3.1%

Safari 4.0 (530.17)

V8 Benchmark

Score: 2494 Richards: 3844 DeltaBlue: 2965 Crypto: 3325 RayTrace: 2829 EarleyBoyer: 3710 RegExp: 1252 Splay: 1206

V8 Benchmark

============================================ RESULTS (means and 95% confidence intervals) -------------------------------------------- Total:                  846.2ms +/- 3.5% --------------------------------------------

  3d:                   187.4ms +/- 14.1%     cube:                74.0ms +/- 23.8%     morph:               62.6ms +/- 16.7%     raytrace:            50.8ms +/- 18.2%

  access:                90.6ms +/- 9.3%     binary-trees:         8.4ms +/- 8.1%     fannkuch:            20.0ms +/- 9.8%     nbody:               54.0ms +/- 20.8%     nsieve:               8.2ms +/- 24.9%

  bitops:                41.0ms +/- 3.7%     3bit-bits-in-byte:    4.6ms +/- 14.8%     bits-in-byte:         8.6ms +/- 7.9%     bitwise-and:          3.8ms +/- 14.6%     nsieve-bits:         24.0ms +/- 0.0%

  controlflow:            4.8ms +/- 11.6%     recursive:            4.8ms +/- 11.6%

  crypto:                64.4ms +/- 10.4%     aes:                 14.4ms +/- 40.8%     md5:                 25.0ms +/- 6.1%     sha1:                25.0ms +/- 6.1%

  date:                  78.0ms +/- 28.0%     format-tofte:        27.6ms +/- 26.8%     format-xparb:        50.4ms +/- 47.7%

  math:                 135.4ms +/- 5.6%     cordic:              54.6ms +/- 7.3%     partial-sums:        55.6ms +/- 15.8%     spectral-norm:       25.2ms +/- 2.2%

  regexp:                33.0ms +/- 8.8%     dna:                 33.0ms +/- 8.8%

  string:               211.6ms +/- 6.9%     base64:              28.4ms +/- 2.4%     fasta:               43.8ms +/- 21.9%     tagcloud:            42.8ms +/- 18.2%     unpack-code:         55.0ms +/- 26.8%     validate-input:      41.6ms +/- 5.8%

<< Home >>

Topics: Browser, Apple, Google, Operating Systems

Kick off your day with ZDNet's daily email newsletter. It's the freshest tech news and opinion, served hot. Get it.


Log in or register to join the discussion
  • Your text and pictures don't match.

    As to what is better or worse. What is what?
    • To clarify, are shorter bars better?

      If so, Safari 4 wins according to your posting.
      • It does say next to the graphs...

        ...but to clarify:

        For SunSpider (the first)shorter is better so Chrome wins.

        For V8 (the second) longer is better so Chrome wins.
        • AKH updated the figures. -nt-

          The original graphs were reversed.
    • Graphs and numbers? What's it all mean?!

      Its amazing the support Apple fans bring to the table.

      They can blog all night and day and be the first to comment.

      But can they read graphs, charts, numbers, and do math?

      • And the Prophet quoth

        "Judge not all Apple fans as one, lest they should judge all your kind the

        And later, after a beer or two:

        "graphs, charts, numbers, and math doth confuse many a pilgrim"
        Graham Ellison
        • Apple users drink beer?

          Well then, maybe they aren't so bad afterall!
        • Oh the poetry of it

          Thy chart doth smelleth fishy.

          WWDC Keynote
          00:29:55 "Fastest browser on any platform"
          00:22:51 Apple's Chart on SunSpider results
          - IE 8 = 1.0x
          - Firefox 3 = 1.7x
          - Chrome 2 = 5.4x
          - Safari 4 = 7.8x

          You want us to believe that they've made a mistake or lied, while we believe that you haven't made a mistake or lied...

          Apple has worked hard to earn their good reputation. Those who challenge it better make sure they can back up their claims or they're going to make themselves look pretty silly.

          Some suggestions for protecting your reputation from becoming that of a mac basher (a person who bashes macs regardless of the truth) you might want to use a legitimate test environment, not just "on a number of other systems". And don't say "similar results", instead, actually do the math and tell us what the margin of error is.

          I'm sure you don't want your articles to be treated like they're just an unprofessional opinion.
          • RTFA

            Sleeper Service
          • Apple's reputation...

            [b]Apple has worked hard to earn their good reputation. Those who challenge it better make sure they can back up their claims or they're going to make themselves look pretty silly. [/b]

            Ah yes... Apple's 'stellar' reputation... Let's take a good hard look at that, shall we?

            Their Mac Vs PC ads - are 99.999999999% LIES.

            Example: The most fun you can have with a PC is playing with spreadsheets. While the most fun you can have on a Mac is making slide show DVDs for the family.

            If you ignore that pesky multi-billion dollar PC gaming industry, they might have a point. But it's still going to be that big elephant standing squarely in the middle of the room. And it hasn't quite gone away.

            And please... Slide shows? WTF? I can do that on my XP box, my Vista box and my Win 7 box. Who's kidding whom? This is considered 'fun'??

            Example: More up to date ad: PC goes to the future...

            PC, in an attempt to see if they ever figured out how to stop Windows from freezing in the future, finds that his future incarnation is just as prone to freezing as ever.

            The simple truth is - PCs don't freeze that often. They don't even BSOD all that much any more - unless the driver for a device is poorly written or the device itself is failing. I haven't seen a 'freeze' since I retired Windows 98. XP just doesn't freeze. It might BSOD, it might crash an app, it may even go into an unresponsive mode where it's busy for a long time, but it just doesn't lock up.

            And let's be real honest here - Apple computers aren't perfect by any stretch of the imagination. They too have issues, kernal panics, BSODs, etc... There's NO such thing as a perfect computer. Not Windows, nor Linux, nor Mac.

            I'd go on with more examples, but I got better things to do with my life than detail Apple's FUD factory output.

            So yeah.. Apple and their ad agency worked really hard on their 'reputation' - for spreading FUD and other bovine droppings. They're quite good at it - and good at getting people to believe it as gospel fact.

            Given this, what makes you think Apple can't ever possibly stretch the truth when it comes to their statistics? Hmmmm...
          • Why Would Apple Lie???

            Hmmmm... Why would they lie?? The Mac is the bestest fastest coolest whitest prettiest computer in the whole world!

            Yeah, sure...

            I remember working for a computer dealer back in the 1984 when the Mac's came out... They had us set them up next to a PC to show how much better they were for graphics...

            They put a MAC with 512x342 monochrome graphics (white letters) next to a PC with CGA Monochrome 320x200 - Oh, and the MAC had a smaller (9") screen than the PC (13").

            Now, as anyone can tell you, especially back then, looking at a larger screen with the SAME resolution (like a giant screen TV of the day compared with a tiny portable) the SMALLER screen always looked MUCH sharper when in fact the actual resolution was the same.

            Mac took it TWO steps further:

            1) They compared 320x200 to 512x342. Gee which one would win?
            2) Ignoring the fact that EGA graphics was the latest for PC's and PURPOSELY not allowing computer dealers to setup any EGA system within 30' of a MAC! EGA being 640x350 after all and in COLOR (yeah so was CGA) would not fair well for the MAC.

            But hey, it wasn't exactly a LIE, was it??

            No, not at all... Even when mac went to 600x400 (eventually) as VGA (640x480) had come out...

            But they sold MACS as having "The Best Graphics".

            They still try to do so today...
          • And Apple did this?

            Hrm...I don't see a margin of error given for
            Apple's tests. Heck, I don't even see any
            mention of the system they're being run
            in...suspicious. Our author here at least gave
            basic system specs and gave us actual results,
            not dumbed down multipliers.

            You want me to just flat out believe the people
            who stand to gain or lose millions based on a
            claim instead of some guy who just writes in a
            blog and honestly won't make a buck or two more
            for it? Who's got more to gain from lying?

            So, please, get Apple to provide us with system
            specs - and they better be specific and varied,
            I'll want to see Power of at least .99 in their
            tests with Type 1 Error Rates (alpha) fixed at
            .001 just to be absolutely certain. Then they
            need to provide us with the software they used
            to do the tests (none of that MiniTab point-
            and-click crap, I want SAS or R being used).
            And what do you think, Mr. I-Know-All-About-
            Statistics, should they use a completely
            randomized design or should they block based on
            OS? And since you can't get all the same
            hardware under both operating systems, should
            we do an incomplete block design or just switch
            back the completely randomized?

            I'm sick of everybody on here talking about
            margin of error when you have no freaking clue
            where it comes from or what it would be based
          • I concur!

            Here here, let sanity reign!
          • Sanity? You're on the wrong board! (NT)

      • If they could read numbers

        They wouldn't be buying Apple. Apple has such a small variety, they could just as well say there is no choice.

        1st of all I am an AMD guy, buy from the little guy, it keeps the prices down. That leaves me out of the loop with Apple, thank God. I don't use Apple apps because the are restrictive.

        My 1st PC had proprietary devices, when you needed to add or fix something, you only had one supplier. They charge what they want, regardless of how much profit they are getting. I didn't really dislike the computer, but when my PSU failed, I was informed I needed a complete new monitor, the power supply was embedded in it.

        The other reason for not going with Apple, they are not really very fast, because they don't offer the state of the arts CPUs or components. Why can't you get a Blue Ray player in an Apple?
        • BluRay

          "Why can't you get a Blue Ray player in an Apple?"

          It's not a built to order option, but there's been BluRay players for
          Mac since 2006.

          As for why Apple doesn't BTO with BluRay, it's probably got to do with
          DRM issues.

          Jobs really doesn't like DRM.

          Rumor has it that BluRay will be supported soon, a matter of months.
    • Spurious reporting by a Mac hack!

      I concur! This is just another spurious nonsensical comparison of
      platforms and OS's. The fact is neither Goggle Chrome nor Microsoft
      Explorer IE 7 or 8 are available for the Mac in any final version. Safari
      is however available on Mac & PC therefore your test is baseless
      except as an expression of poor journalism, if that is indeed what you
      are, clearly your objective of sensationalism is transparent.
      Had you compared the speed of Safari and Chrome on a comparable
      Mac and Windows machine the results would have been more
      meaningful to us as users.

      Additionally, as WINDOWS XP and VISTA cannot run Mac OS X and
      Mac OS X can run any iteration of Windows OS using Boot Camp,
      your test would have been more credible had you run the test using
      Safari in on Mac and Widows on the Mac as a native install using Boot
      Camp on the same machine, thus removing OS Platform and any
      hardware variables.

      Sir, all you have demonstrated here is your inability to peruse facts or
      factual reporting to you readership. Therefore, while we Mac
      users wuill hence review your words and sentiment as you stated:

      Quote "I?m always suspicious of all claims that are along the lines of
      ?world?s blankiest blank, especially when those claims come from
      Apple. "End Quote!

      Sir, we too will now equally in turn be forever suspicious of any
      claims made by Adrian Kingsley-Hughes relating to Apple and the
      veracity of your so called unbiased testing. Your journalistic bias
      is showing Mr. ingsley-Hughes. The additional fact that you author a
      blog "The PC Doctor" indicates there is not as much need for a "Mac
      Doctor" blog relating to fixes and support not already offered by Apple
      Inc., excellent award winning support, which again only indicates that
      you are extolling another hand to feeds you, this time instead of
      Microsoft you have chose Goggle.

      I believe you are transparently biased; as am! However, my bias is for
      the best user experience regardless of manufacturer and regard for
      what each user can afford for themselves and their particular
      computing need, even when it is free as in the case of the Goggle
      browser still in Beta. Free is never free as indicated by the add content
      displayed as part of the Goggle browser. Had you pointed that out I
      would have considered you a better journalist instead of just another
      hack with a suspicion of every thing Apple says and does. Please, if
      you are going to report something as fact then you must level the
      playing field and be transparent.

      Dr. A. de Mandeville PhD
      • Get a Life!!!!

        People... do yourselves a favor and GET A LIFE!!! Nothing you all say here puts money in your pockets, brings about world peace, or saves the environment... Please find something better to do with your time... and maybe write the perfect OS or application... then you all will have something of value to say.... until then.... STFU.

        Thank You!
        • re: Get a Life!!!!

          Why don't you STFU ?

          You have such grandiose ideas about life in general, why are you reading these opinions like the rest of us 'LOSERS' ????

      • you CAN run OSX on a PC but WHY??

        There are plenty of methods for getting OSX to run on a PC, but why would you want to??

        I have done it on several boxes just for kicks, but I sure wouldn't want to pay Apple for the OS (sure, like they pay for XP/Vista when they put it on a Mac!!)

        Don't think so? Just google something like "OSX on a PC"...

        Of course you will be breaking Apple's software rules which state that it must ONLY be run on a MAC, but it can be done and pretty easiely actually...