Nvidia to Apple: Benchmarks or It Didn’t Happen

by: Bogdan PetrovanMarch 8, 2012

Tegra 3 vs AX5 TheVerge

Everyone knows the importance of benchmarks in the high-tech industry. With things like user experience or manufacturing quality, manufacturers, or, better said, marketers, can make unsupported claims, without ever risking to be called out. But in the world of cores and RAM modules, benchmarks are, almost always, the final argument in any dispute. So, when the largest tech company in the world claims that its new device is four times faster than a competing product, it’s only natural to ask for a benchmark.

During yesterday’s iPad press event, Apple’s SVP of Marketing, Phil Schiller made a claim that raised quite a few eyebrows. Schiller told the audience that the iPad’s new dual-core Apple A5X processor is, in terms of graphics performance, four times faster than the Tegra 3. Right on cue, a minimalist graph appeared behind Schiller, showing the A5X column towering over Tegra 3. And that’s it. No benchmarks cited, no fine print, no footnotes.

Now, I like Apple’s marketing tricks as much as the next guy (kidding, I don’t). And I know that Apple’s philosophy is to go beyond specs and focus on the user experience (how much RAM in the iPad 3, anyone?) But when making such strong claims, even if your name is Apple, you should at least be prepared to show some proof, right?

Nvidia: We are Flattered by Apple’s Attention

ZDNet reached out to Nvidia for commentary on Apple’s affirmation, and the Santa Clara-based giant responded by expressing their “gratitude” for being mentioned during Apple’s event. Irony aside, it’s interesting to note that Apple found it necessary to differentiate the new iPad by comparing its A5X processor with the Tegra 3 SoC. It’s another recognition that Nvidia gets for its Tegra 3 4-cores-plus-1 architecture, which it pushed hard over the last months and especially during the recently concluded MWC.

Tegra 3 vs iPad A5X

Nvidia’s representatives refrained from denying or confirming Apple’s claim, which is natural when you consider that Apple carefully avoided revealing the conditions of the tests or the benchmarks that it ran to compare A5X with Tegra 3. Also, Apple seemed to engage in some cherry-picking (apple-picking?), avoiding to mention how its dual-core processor does in terms of performance when compared with Nvidia’s SoC, or even about the amount of RAM in the iPad.

Did Apple make up their statistic about A5X being four times faster than Tegra 3? Probably not. I am sure that there are applications where Apple’s chip is beating Tegra by a long shot. Moreover, we might not ever know what exactly Apple was referring to when they made their claim.

Even if Apple refuses to show how it came up with the claims it flaunted yesterday, real-world benchmarks are just a few days away. Both Nvidia and a number of experts have announced their intention to compare the new iPad with Tegra’s finest, so we are bound to get some hard figures, sooner or later.

  • Graham Laight

    My opinion is that you have to focus carefully on the wording: they are only claiming 4x speed for graphics performance – it won’t necessarily do any better recalculating a spreadsheet, for example. The higher resolution display on the iPad 3 probably needs a lot of graphics processing.

    • fps_beaTt

      Due to algorithm efficiency, an increase in resolution of that magnitude will @ most increase the graphical requirement by 25-75%. So, it’ll still be noticeably quicker than iPad 2, which is twice as fast as the Tegra 3.

  • iRule

    To be honest .. i have an ipad 2 , so far comparing to all android based pads , including Galaxy Note 10.1 , its much faster , i mean way faster , i don’t know how apple does this , but with lower ram and processor , everything i run on my iphone 3gs and my ipad2 works smoother than my sgs II … i’ve had the chance to test a galaxy Note 10.1 , although it was new and empty , it was slower then my 16gb loaded ipad2 already running a lot of apps on the baground … therefor if apple says their new A5X is 4 times faster then Nvidia’s tegra 3 , then it is probably true ..

    • Rus

      My Asus Prime (overclocked to 1.7 GHz) is far away from your apple crap. So I’m sure that ipad3 having more resolution will fail too as the marketing and the real apps/benchmarking are different things.

      • fps_beaTt

        Again, check the Anandtech benchmark. The A5 @ just under 1GHz doubles the performance of the 1.4GHz Tegra 3. Accept it; the pathetic connection between your ego and your hatred for Apple products is no match for reality.

        • watchman626

          GLBenchmark is a joke. I have installed it on my Prime and can safely say that it is a very unstable app and that it is not optimized for Android. So why don’t you take your ego and hatred for any product that is not Apple and get the hell off this Android blog.

        • Douglas Anderson

          Anandtech has been discredited for a very long time. Instead check real benchmarks and you will find Apple is 12-24 months behind.

    • You have to define “faster” here, because if this is more of the “the app drawer doesn’t scroll at 60 fps” argument, that’s just silly. Portions of the Android UI can, at times, lag for an instant. This is because the Android OS is not only more complex than iOS, but also has to be able to run on lots of different kind of hardware. This is one thing Apple has going for it, but also why you don’t see big improvements in the specs from one device to the next.

      Some people prefer how iOS works. My girlfriend *loves* her iPhone. Personally, I like being able to tinker with stuff on my Android. I hardly use any of the stock applications. I’ve replaced the launcher, the keyboard, dialer, contacts, sms app, just to name a few.

      But you, like so many people over the past few days, are mis-reading the entire situation that’s causing the complaints:

      “if apple says their new A5X is 4 times faster then Nvidia’s tegra 3 , then it is probably true ..”

      Apple *didn’t* say that the A5X was 4 times faster. They said the *graphics* were 4x faster, and everyone is just asking for clarification on what that *means*. If you like iOS, good for you.

      All that said, a couple of things to note, if you’re getting lag in Android that’s more than an occasional stutter in the App drawer:

      * Tons of widgets will do it
      * Some apps are written badly and eat up CPU in the background. Check out System Panel in the market and see if you’re having this problem.
      * Check your WiFi sleep policy. Some manufacturers set there’s funny. Don’t know about Sammy.
      * DON’T USE TASK KILLERS. They actually fight against the OS and cause it to continuously have to re-load applications into memory. Android manages its own memory just fine, and app killers are completely unnecessary unless you’re still running Android 1.6.
      * Moving apps to the SD-Card will cause a couple of extra seconds of load time. It’s a great way to save space on the internal storage, but the SD-Card is slower to access. If you’ve got an app that takes too long to load, make sure it’s not pushed to the SD-Card.

    • Douglas Anderson

      You are in a dream world. The iPad 2 is not faster than any modern Android tablet. Heck, the iPad Air is not faster than any modern Android tablet.

      Apple is the King of fake benchmarks.

  • yodb

    I have a Prime and it’s pretty fast and neat. I have some hang ups, but probably it’s rather the OS than the CPU. Hopefully some updates will come to ICS and everything will be smoother. (ICS already changed a lot and became lot faster than it was.)
    Anyway, I don’t think they’re lying, they’re just using the good benchmarks. I remember having custom roms on my desire with 900-ish score on Quadrant, and everything was really really fast and usable. Other roms above a score of 1200-1300 were horrible. So, benchmarks doesn’t really matter I guess.

  • Bigdaddyrey10

    Sorry my galaxy tab 10.1 is way faster than ipad 2

    • jaaaaaaa

      I have galaxy tab, its by far faster than iPad 2, Apple fakes people by thinking your machine is fast, by ginving almost all priority to user interfaces, but in the background it still takes time @ least as long as other tablets. Other tablet do real multitasking, apple is still trying. When using unitask, its very simpel to fake performance.

  • Common Man
  • jasonponce5062

    You should change the articles name too

    Nvidia to Apple: Benchmarks It Didn’t Happen

  • fps_beaTt

    Ahhh, all the Apple haters conveniently ignoring the Anandtech benchmarks comparing the A5 dual core GPU to Tegra 3; it clearly shows the A5 doubling the performance of the Tegra 3. Google it.

    • They probably compared A5x vs 4-PLUS-1 Powersaver core of Tegra 3 =)) All 4 cores were resting =))

    • Fifthblindmouse

      Anandtech is one company with skewed results for the fanboys to rally behind, that being said there are many more benchmarks out there if you *ahem* “Google it” – If you ask me Anandtech is full of crap.


    • jaaaaaaaaaaaaa

      Yes but if you werent that stupid you’d also see Tegra 3 crushes A5x on all other fronts, including gaming.

  • Jj-snoop
  • Cooktim8

    PowerVR SGX 543MP2 Smokes Tegra 3’s GPU :)))

  • Bubbafats

    ok, the a5x does not process things faster itself, such as media converting, word processing. but the ipad has a high resolution screen so it needs a fast basic GPU to run it. so they put one in there and it is a fast running processor, BUT………it does not handle smoke, fog, lighting, physics, ragdoll effect, ambience as well as the tegra processor does. so will it run slightly smoother 60 fps vs 52 fps yes, but it will not make it look better, more life like, more realistic. that’s is the big difference, YES it runs faster, NO it will not handle extra effects that add that little realism.