Submit News Alternative Tip Form

Benchmarks Show Modest Performance From 8-Core Samsung Chromebook

exynoseWith a mammoth eight-cores tucked up tightly inside, the new Samsung Chromebook 2 sounds like it has all the oomph to outrun its older 2012 self. 

Initial benchmarks testing specific computational functions of the Exynos Octa chip have revealed something surprising: the gains over the dual-core version used in the previous model are modest rather than magnificent.

The video below, courtesy of mobilegeeks.de, shows the 11.6-inch Samsung Chromebook 2 as it’s subjected to a series of standard browser-based tests.

)

So what gives?

Raw Data is Unprocessed Relevance

At face value, the results of the tests in the video above show that – in pure, raw performance terms – Chrome OS running on the Exynos 5420 SoC does not deliver a massive leap in power over that offered by the older, dual-core Exynos 5250.

Better? For sure. Worth upgrading for? Jury’s out.

With both chips coming from the same 5x series there’s an argument to be made that some of the lofty expectations had for it were misplaced to begin with. It’s also true that, while useful, stats stressing the number-crunching capabilities do not always accurately reflect the perceptible benefits that seemingly small bumps in results can have on user experience.

Low Octane Processor

Of the benchmarks run by mobile geeks and shown in the video above, the Octane test is useful in determining how well a computer copes with loading complex, JavaScript-heavy pages like Gmail, Google Music and Google+. The higher the score the faster and more responsive page rendering will be in real time.

The Exynos Octa 5420 shipping in the 11.6-inch Samsung Chromebook 2 (the 13-inch model sports a marginally faster chip) shows Octane results nearing 6,000 — more than 2,500 higher than those delivered by the dual core version.

This is actually a great score, and anyone upgrading from the older model will notice a more responsive web experience. On the flip side it remains well below the scores reputedly being achieved by the Intel Celeron 2955U, which canny Googling shows regularly passes 10,000.

With more cores at hand for processes to balance across, more RAM and a superior GPU, Chrome OS running on the Octa chip will feel faster, will multitask better and, as web technologies evolve to take advantage of multiple cores, should post better results as time goes on.

High Octane Caveats

Browser benchmarks are, like the oft-touted ‘horse power’ of a car, only part of the story. Other conditions, like usage load, temperature, battery life, etc. also play a role in determining how well a chip performs. This makes determining the true experiential performance of a chip a less qualifiable task; rigid benchmarks are rigid.

As Chrome OS renders each tab as a separate process, eight-cores – in theory – will deliver a better experience than two, regardless of how well those cores perform individually. Think of it like balancing plates: the more arms (‘cores’) you have, the easier it is to do.

Add in more RAM and Chrome should deliver a more responsive experience, with smoother switching between tabs and faster page loading times.

Benchmarks, for all their use, don’t test typical user behaviour.

Specific Stats

In the snippets below I’ve highlighted the Browsermark results (higher is better) in bold. This is because, more than any of the benchmarks posted thus far, this is the one that gives the best indication of overall performance, as it tests things like general rendering and graphics handling.

Samsung Chromebook (Exynos Dual)

  • SunSpider: 668ms
  • Browsermark 2.0: 2170
  • Octane: 3465

Samsung Chromebook 2 (Exynos Octa)

  • SunSpider: 619ms
  • Browsermark 2.0: 3320
  • Octane:  5982

The Takeaway

It’s easy to get caught up in a game of raw numbers, like kids playing a game of Top Trumps.

However, as I keep reiterating, real world experience matters more to most of us than which chip has the higher score. Noticeable improvements in multitasking, particularly with JavaScript-heavy websites open, is experientially better on the Samsung Chromebook 2.

It won’t beat out its Intel Haswell siblings any time soon, but it won’t crease up as quickly as current ARM offerings do, either.

  • t

    wat

  • http://lukelarris.com/ Luke Larris

    The reason it’s not performing much better is because the Javascript processor in Chrome does not support multiple cores, and I’m guessing the new CPU in this thing only gets a marginally faster clock speed.
    Chrome is a multi-process browser though, so while the CPU may not beat up its older brother in terms of performance on one site, it will outperform it when you’re loading a lot of tabs at once.

    • Sean Lumly

      But it *is* performing significantly better. Consider that the two credible tests Octane and Browsermark show significant improvements. Octane is 70% higher than the predecessor, and the Browsermark score is 50% higher. Something like 10-15% would be considered marginal, but anything as high as 50% is significantly improved.

      • Sean Lumly

        It seems that the Beta channel (and presumably the Dev channel) are seeing scores as high as 7000 in Octane! If true, expect Chromebook 2 scores (and performance) to go up considerably as the source for these channels is folded into the Stable builds. According to the article above, that is a hearty 2x the JavaScript performance of the original Chromebook!

  • starryhope

    Yeah, I’m more interested to see real world performance with multiple tabs. Hopefully, Facebook and other such sites won’t bring the whole machine to a stop. I’m also interested to see the screen. Hopefully, it’ll be better than the old one!

  • Random

    The real world value of the Samsung Chromebook Exynos Dual, performed poorly in any demanding avenue, especially videos, having both side by side with my C720, it is like night and day in terms of performance… Let’s hope these second edition Chromebooks can do better.

  • W.G.

    The Octane score for the Samsung Chromebook (Exynos Dual) is abnormally low, I usually score over 5000.

  • Wesley Files

    Those Browsermark and Octane scores for the Series 3 Samsung are really low. My Guest Mode scores were only 5% and 10% lower, which, maybe isn’t fair because I don’t think that original test video was performed in a fresh Guest Mode.

  • Kenny Strawn

    There certainly needs to be more massively multithreaded benchmarking utilities out there…

  • Na7noo7

    What about the 13 inch?

  • Joe Montfort

    Oh, here we go…

    I’m all for getting as good performance as possible, but I hope Chromebooks don’t go down the PC road of just hyping speeds and feeds in their sales pitches. I’d rather see someone make a mid-priced ‘book with more immediately recognizable features to distinguish it. How ’bout lighted keyboards? Metal chassis? Better screen?

    The “Haswell! Haswell! Haswell!” charge is already getting boring. Geeks (and I use that term in a positive sense) aren’t the only ones buying these things. There needs to be some mass appeal. Citing a bunch of specs without context won’t attract normal people…

    • Anonymoused

      I agree wholly. While real-world performance definitely matters (despite how beautiful my HP 11 is, it lags so often and it’s frustrating to use from time to time), I really just care about design and ease of use. If I wanted speed, I wouldn’t be using any kind of laptop at all ;) I’ve got a desktop for speed.

    • j4ckl3

      I agree with you as well. Even coming from a 48 year old geek, it has to be an eclectic blend. Would I pay the $2.00 more for the cost of installing a couple of 6 cent led’s on the bottom of the keyboard. As Apple has found once from removing the option from on of their models, it was highly sought after to bring it back.

  • Kenny Strawn

    355.5ms for my Acer C720 on SunSpider, and 9000+ regularly on Octane… Yup, those easily outperform single-threaded operations. That’s why the benchmarking utilities should do tests like, for example, a


    while(true) {
    var w = window.open();
    w.document.write(document.documentElement.outerHTML||document.documentElement.innerHTML
    }


    test to see how many tabs can be opened before the Chromebook slows down, then close all of them when it approaches a threshold. That should probably give a more acccurate reading of Chromebook performance especially

  • umbrarchist

    Maybe Android doing non-browser benchmarks would show a bigger difference. Could the problem be that Chromebooks don’t don’t make use of what the chips can do.

  • Craig Hansen

    I want the Chromebook 2 for its 1080p, but I prefer the processing power of the new Intel Bay Trail models, and touchscreen would be nice, too. Can’t any of these companies get their act together to come up with a single Chromebook that’s 1080p, 4GB, touchscreen and Intel Bay Trail, all in one device? Come up with that, under $400, and that’s the Chromebook I’ll buy.

    • http://geekfed.com/ Cody Engel

      …Chromebook Pixel.

      • Craig Hansen

        No, the Pixel is about 3X too spendy.

  • John Scott

    From the get go Samsung has had issues with the Exynos 8 core. I really don’t see the chip making great strides as Intel is with Hazwell. My only complaint with my Samsung Chromebook 3 is the bad performance with multi tasking or video playback. But at $200 which is what I paid, I did not expect big performance. Now that Samsung has raised prices on its Chromebooks. I think the new Exynos chips fall short of justifying that price bump. My next Chromebook won’t be a Samsung Exynos powered one.