iPad Test Notes: Speed (Versus Tegra 3)

iPad Test Notes: Speed (Versus Tegra 3)


Apple made some bold claims about its new A5X chip, claiming it would give a 4x graphical boost over NVIDIA’s Tegra 3. At the same time, the Tegra 3 is a powerhouse of computational processing. So lets put ’em to the test.

First, the usual caveat: benchmarks only mean so much. There are a lot of anomalies that are hard to account for, and benchmarking across different operating systems is problematic at best. User experience is more important.

Unfortunately, there are only three tests that work well on both iOS and Android: the SunSpider JavaScript benchmark, the BrowserMark benchmark and GLBenchmark 2.1. SunSpider and BrowserMark are both in-browser tests that indicate how well your device (or, more specifically, your device’s browser) handles JavaScript, HTML and other browsery tasks. We used default browsers (Safari and stock Android) for both of these. GLBenchmark 2.1 tests graphics performance, with an eye toward gaming. (Note: We planned to use the age-old Linpack standard test to evaluate CPU performance, but the Linpack apps for iOS and Android are very different, and their results could not be accurately compared.) For our test we’ll be using the new iPad and the ASUS Transformer Prime.

SunSpider: Tie

Results in the SunSpider test were extremely inconsistent, but the two tablets scored within the same range. In a race-type test in which lower times win, both had peaks in the 1680ms range and both had troughs down to the mid-1800s. We’re calling this one a draw, but we go on record as being suspicious of the results because of the variance.

BrowserMark: Transformer Prime

The BrowserMark test gave much more consistent results, with less deviation, which made it seem more reliable. Higher scores win this test, which is designed to test Java and HTML rendering. After several runs, the Transformer Prime averaged 107,632 and the iPad averaged 99,908, making the Android device 7.2 per cent faster. But again, different browsers will yield different results.

GLBenchmark 2.1: iPad

The iPad definitely won the graphical performance tests. In the Egypt standard test, the iPad processed 6740 frames at 60fps; in the Pro standard test it processed 2974 frames at 60fps. Compare that to the Prime, which clocked 5080 frames at 52fps and 2775 frames at 56fps, and the winner is clear. Watching the tests go, the iPad looked very smooth, whereas the Transformer Prime struggled with artifacting and other graphical anomalies.

Did the iPad win? Yes, solidly. But not even close to the 4x performance Apple promised, so it’s unclear where the company got that number they were boasting about.

Subjective Speed Tests: Tie

To test user experience, we downloaded some apps that are available on both platforms (iOS and Android) and opened them both simultaneously. Facebook, Twitter, Fruit Ninja and others all opened and worked at nearly the exact same moment. How close was it? Close enough that it doesn’t matter at all. That’s the good news. Whichever tablet you have, it’s super fast and great. Stop trying to brag to your friends about milliseconds and just enjoy your awesome toy already.

Overall Winner: ???

Sorry to disappoint you, but there was no clear winner here. Results seem to indicate that for blasting through graphics, the GPU (graphics processing unit) in the iPad’s A5X is faster. But for your average day-to-day usage, the CPU on Transformer Prime takes it. I’d love to see a good, cross-platform CPU test emerge, but until then, we’ll just have to listen to nerds screaming at each other. Get your earplugs ready.

Image credit: Shutterstock/Ssuaphotos.


The Cheapest NBN 50 Plans

It’s the most popular NBN speed in Australia for a reason. Here are the cheapest plans available.

At Gizmodo, we independently select and write about stuff we love and think you'll like too. We have affiliate and advertising partnerships, which means we may collect a share of sales or other compensation from the links on this page. BTW – prices are accurate and items in stock at the time of posting.