Samsung ATIV Smart PC: Performance

Anand covered the CPU and GPU performance of the Atom Z2760 really thoroughly in his Clover Trail platform review with the Acer W510. At this point, Z2760 is pretty familiar to us—two Hyper-Threaded Saltwell cores at 1.8GHz, along with PowerVR’s SGX 545 GPU clocked at 533MHz. Saltwell is the 32nm shrink of Bonnell, which you probably better remember as the heart of the much beloved Atom N270 netbook processors. (I’m kidding about the much beloved part.) The new 22nm microarchitecture cannot come soon enough. Here's a rehash of Clover Trail performance, which is generally identical to the Acer W510.

SunSpider Javascript Benchmark 0.9.1

Mozilla Kraken Benchmark

For the first time, we’re seeing Intel lose its performance edge to Cortex-A15 based SoCs, but for now Clover Trail is still competitive from both compute and power efficiency standpoints. That’s more than can be said for SGX 545, even at such a high clock. Clover Trail+ changes that though, with the inclusion of two SGX 544 cores, and should offer graphics performance that is in the same range as Adreno 320, as well as being much more competitive with Apple’s recent SoCs. But CT+ hasn’t arrived yet, so for now we’re left with good old Clover Trail. I'm leaving Surface Pro's numbers in the following graphs, just so that you can see how much faster the ultra-low voltage IVB parts are when compared to Clover Trail. It's a pretty huge difference, even when looking at just the Core i5.

WebXPRT—Overall Score

TouchXPRT 2013—Photo Enhance

TouchXPRT 2013—Photo Sharing

TouchXPRT 2013—Video Sharing

TouchXPRT 2013—Podcast MP3 Export

TouchXPRT 2013—Photo Slideshow

Samsung ATIV Smart PC: Display Samsung ATIV Smart PC: Battery Life
POST A COMMENT

60 Comments

View All Comments

  • beginner99 - Tuesday, March 19, 2013 - link

    Suspected that but makes the whole benchmark even more flawed it is in the first place. It doesn't even mention the browser used, at least not easily visible. I assume it was IE then considering the crappy performance. Reply
  • Kidster3001 - Tuesday, March 19, 2013 - link

    Kraken ( and Sunspider ) are not CPU or even system benchmarks. They are Browser benchmarks only. Clovertrail and Clovertrial+ scale perfectly with frequency as compared to Medfield when they use the same browser. Reply
  • kyuu - Monday, March 18, 2013 - link

    Hope a review of the Vizio Tablet PC with the AMD C-60 is incoming soon! Reply
  • kyuu - Monday, March 18, 2013 - link

    Er, Z-60. Reply
  • nikon133 - Monday, March 18, 2013 - link

    What about Lenovo ThinkPad Tablet 2? I think proper docking station option (with LAN, external screen and a few USB ports) makes it really interesting among Clover Trail tablets, but it is not being mentioned here at all. Is it available in US? Reply
  • hughtwg - Tuesday, March 19, 2013 - link

    Yes, it's available in the US. Mine is on Ebay at the moment.

    I think docking the tablet directly to a keyboard like the Ativ, or Envy X2 is much more practical than a dock like that provided by the TPT2 since the keboards have essentially the same ports as the non mobile dock for the TPT2. I bought the TPT2 for the stylus support but found it rather disappointing due to a combination Windows 8 and Lenovo issues.
    Reply
  • sprockkets - Monday, March 18, 2013 - link

    Still no Nexus 10 review... Reply
  • A.J. - Tuesday, March 19, 2013 - link

    Pls try using the same Google Chrome/MS IE10 in Octane or Kraken javascript benchmark test!
    It's totally ridiculous to have the javascript benchmark score with different browsers. Am I looking at a browser test?

    with Chrome25:
    I got 2580 on Octane bench,
    about ~900ms in Sunspider,
    and ~12000ms in Mozilla Kraken.

    Intel Clover Trail Atom is way better than old cortexA9! Even close to cortexA15 in some cases.
    And with the remarkable 1.7W TDP, it totally smashes the A15 in efficiency!

    Curious to see 22nm BayTrail quad-core Atom vs Tegra4.
    Reply
  • paulbram - Tuesday, March 19, 2013 - link

    Although in general I agree it is not fair to compare different browsers, I think you also need to remember that Chrome on a Windows 8 CT chip is really not NEARLY as good as IE10 is. In fact, I've been incredibly impressed with IE10 perf on mine. So, sure you could run Chrome if you want to see poor benchmark results, but why not give the machine the best chance at success? Reply
  • Kidster3001 - Tuesday, March 19, 2013 - link

    Javascript benchmarks are "Browser" benchmarks, not CPU benchmarks. They are completely dependent on the javascript engine used in the browser. Reply

Log in

Don't have an account? Sign up now