Follow Slashdot blog updates by subscribing to our blog RSS feed


Forgot your password?
Chrome Google Ubuntu Hardware Linux

$250 Chromebook With Ubuntu Linux Is Very Fast 117

An anonymous reader writes "The Google Samsung Chromebook was already interesting for its competitive $250 price-tag and that it can be loaded with Linux distributions beyond Chrome OS, but it turns out that its performance is particularly good, too. When loaded with Ubuntu Linux, the Samsung Exynos 5 Dual ARM SoC on the Chrome notebook had outperformed a 1.8GHz Intel Atom, a quad-core Calxeda ARM server, and a TI OMAP4 PandaBoard."
This discussion has been archived. No new comments can be posted.

$250 Chromebook With Ubuntu Linux Is Very Fast

Comments Filter:
  • by luckymae ( 2691983 ) on Tuesday November 20, 2012 @12:26PM (#42042639)
    But will it blend? first post!
  • by timeOday ( 582209 ) on Tuesday November 20, 2012 @12:33PM (#42042749)
    This sounds like a potentially fun, cheap device. Does Ubuntu for ARM have all the same packages as x86? (From a check of the Ubuntu ARM web page it appears a lot of the focus for ARM is on the Server distro?)
    • I run debian/ARM on a qnap TS 409 (text only, no X). Some programs/kernel modules are x86 specific by their nature (example: virtual box). Everything else should be available.
    • Per the discussion here [](which includes a Google dev who wrote up some early instructions on running standard linuxes on this ARM Chromebook, and at least one Linaro project person, among other clueful types, Apparently mainline kernel support for the Exynos 5 SoC is expected in the near future but not 100% just yet.

      As for ARM packages, you are very likely out of luck for 3rd-party binaries(eg. Flash, Oracle JVM), and may be a more or less second-class citizen in some areas(the javascript JIT compilers in su

      • by snadrus ( 930168 )
        (to help GP)
        The open-source JVM is available though, which is very feature-complete. The open-source flash lags behind or (usually) isn't worth it, but YouTube has HTML5 videos now.
        Anything Windows-oriented is unavailable (FOSS or not) like ndiswrapper (which you shouldn't need), wine (not running windows apps directly may affect you). There's KVM (Kernel Virtualization Module), but it only virtualizes ARM systems. If you must, there's bochs to run x86 on ARM, but it's slow and limited.
        Other "common for
    • I haven't kept a close eye on ubuntu but debian armel and armhf are at about 98% while debian i386 and amd64 are at about 99.5%. Since ubuntu is based on debian i'd expect their figures to be similar.

    • I have two ARM systems at home. One is QNAP TS-110 1-bay Home NAS Server powered by Debian ARM Linux and another one is Raspberry Pi powered by Raspbian(again Debian -> Linux) which serve as XBMC media center application. They are contrast to each other, But mostly I am not disappointed in finding the package due to its ARM nature. I had to compile the source of eGalax touch screen for Raspberry system (it is DIY). I was not able to find dropbox support for ARM system (it is closed source). Except a very

  • by CajunArson ( 465943 ) on Tuesday November 20, 2012 @12:36PM (#42042791) Journal

    Now that the latest ARM chips from late 2012 are actually faster than a similarly clocked Atoms using the exact same architecture that was introduced in 2008 (well at least in some of those benchmarks, the Atom won some too), will we finally see the ARM fanboys talk-up Atom as Intel's best chip of all time?

    Remember, when you say that Atom is a complete PoS and simultaneously crow that you finally beat it in performance 4 years after it hit the market, you kind of sound like someone who bragged about cheating to win the Special Olympics...

    • by Anonymous Coward

      Architecture fanboys don't use benchmarks to form opinions, they use benchmarks to form propoganda.

      The only numbers that matter are whatever ones [insert favorite architecture here] is better at.

    • What the hell? Maybe if the Atom was operating in the same power margins. The whole point is the arm is using way less power, hence why you can see it in phones.
      • You haven't bothered to look at Anandtech's review of this system then. Considering the Exynos SoC is sucking down 8 watts of power running a single-threaded non-GPU Mozilla Kraken benchmark, you better believe that Samsung is going to have to cut down this chip's performance to run in a smartphone power envelope.

        • I was thinking of EXACTLY that review, where they mention all it needs is a lower clock speed. As opposed to the much slower atom used there, which has a higher consumption, or the apparently equal performing atom used in the test above which has a 35W draw AT IDLE with chipset. The Exynos 5 chromebook as a whole system including display has a draw of just over 11W when running a benchmark. So no, the Atom isn't even close on power draw, and clocking it down will not make it work in a phone.
          • are comparing a state-of-the-art 28nm SoC on the ARM side with a several years old 45 nm Intel netbook that includes a separate chipset.

            I find it hilarious that you only looked at that one part of the Anandtech review and declared victory for ARM when even you know that 32nm Medfield SoCs were on sale before the Exynos 5 even launched and have substantially better power/performance ratios than were exhibited in the Anandtech numbers.

            I find it even more hilarious that you summarily ignored the Has

            • I read the whole review where the Atom N570 fared extremely poorly in comparison for more power draw. Nothing in that review supports anything you are saying. And I don't see any links to any Haswell demos in any of your quotes. If you can find me anything showing equivalent performance per watt on the Intel side, please link it. I would be interested in seeing it.
              • by 0123456 ( 636235 )

                Yes, today's ARM uses less power than a 45nm Intel CPU released nearly two years ago. Shocking news!

              • This.

                Is it really so hard to grasp that what's most relevant in ARM vs x86 comparisons (probably other arch comparisons as well) is FLOPS/W?

            • by hattig ( 47930 )

              This Exynos 5250 is a 32nm SoC.

              The 28nm Exynos 5450 is coming out next year, with two more cores, twice the GPU and a faster clock speed.

              Haswell is simply not going to compete in this area of the market, where price, power consumption and performance come together. It will compete at higher price points, maybe even at low power in single-core ULV (ultra-low-clock too) variants, but not all three.

              Intel's 22nm next-generation Atoms are coming out in 2014.

    • by Hatta ( 162192 )

      You forget the importance of efficiency. Performance per cycle is irrelevant, performance per watt is what matters.

    • by rbmyers ( 587296 ) on Tuesday November 20, 2012 @01:58PM (#42044039)
      Ok, so I noticed that one system is apparently using a solid state disk and the other a conventional disk.

      Given that the limiting bottleneck of a notebook with a decent processor is almost always the disk subystem, I stopped reading. Did I miss something?
    • by hattig ( 47930 )

      The next generation of 22nm Atoms with the new core has been delayed until 2014.

      A best of breed current generation Atom (dual-core, quad-thread) is thoroughly beaten in many benchmarks by this slightly slower clock-speed, first generation, 32nm dual-core Cortex A15 product. Next year brings a lot of 28nm quad-core A15s at 2GHz including the Exynos 5450.

      And because it is cheap - no Intel tax - you can get a very decent computing device using it for $250, which is a darn sight better than the pricing of Atom

      • by Guspaz ( 556486 )

        Dual core haswell ULV parts are coming in at 8-10W TDP, which is far more interesting than Atom at those power levels. That's too much for a smartphone, but almost in tablet range, and definitely in net book range.

        • by hattig ( 47930 )

          Let me know when an 8W Haswell is available for under $40, as it would have to cost for it to be viable for a $250 netbook.

          The 8W Haswells will be used for Ultra-ultrabooks, or long-life-ultrabooks, costing at least $799.

    • Atom 5watts
      Intel chip set 20 watts
      Screw your atom

  • Those are completely unreadable, the text is about 3 pixels tall. Tried zooming, didn't work.

  • by Anonymous Coward

    This $250 device is listed in the UK for £250 which doesn't make it that competitive compared with the cheap netbooks that can be purchased for £170 and £200 with WIndows 7 starter shovelled onto the hard disk. AND you can install Ubuntu or whatever too!

    I've had a little Compaq netbook for just on 2 years now, which cost me £190 then. Ok the Atom CPU isn't much to write home about, but it works fine dual booting between Win7 Home Premium and Ubuntu. Its fast enough for all the stuf

  • The screen resolution (1366x768) is ridiculously bad on these things. I have that on my 2 year old tablet. I don't consider it a useable resolution. Give me the same dot pitch as my phone.
    • Cmon, be kind, It's a $250 laptop for Christ sake. I think we can allow an exception for the 1366x768 res in this case. HOWEVER, I still find it ridiculous to see the same resolution on laptops 3 times the price, especially since my 10+ year old thinkpad rocks 1600x1200. Oh god, don't get me started on how useless screen resolution is these days.
    • by admdrew ( 782761 )
      I'd also certainly like to see a better resolution, but what exactly is unusable about it on this laptop, given its overall size, speed, and intent?
  • couldn't they find less exotic tests?
    anyhow, the benches are from phoronix. the intel tested is the cheapest crappiest foxconn all-in-one you can find. you'd think they'd have plenty of test results from some other relevant machines too. I mean, who the fuck gives a fuck about how pandaboard does on the bench? nobody, that's who. and maybe bambino.
    and anyhow, intels lowest atom pricing is.. well, it is what it is due to competition having been what it is (that's right. they're selling atom as shit cpu as th

  • When can we run Windows XP on it?
  • I've been trying to find out if the internal storage can be upgraded. So far I've drawn a blank which is making me think that it can't (so far). For certain the RAM can't but I could live with that. Not with so little internal storage though.

    • by admdrew ( 782761 )
      The dude who posted instructions on how to through Ubuntu on there does mention SSD upgrading in passing ( - "Note: If you've installed a larger SSD in your Chrome device..."), but I'm not sure how easy/possible that is from a hardware perspective. I haven't opened mine up, but others have alluded to everything pretty much being soldered onto the mainboard.
    • The internal storage device is an eMMC module soldered to the motherboard. Unless you have a BGA rework setup and nerves of ice, no go.

      this gallery [] has motherboard shots.

      It does support SDHC cards and USB mass storage devices.

  • As the title suggests, I'd like to know if there is any way to run Android apps on ChromeOS. If yes, that would make Chrome OS the desktop Linux OS with the largest application base... even if most apps aren't designed for desktop use.

    • There is not. Chrome OS is literally that: an OS that runs a browser and only the apps that work inside of it.
    • by admdrew ( 782761 )

      As the title suggests

      What's being suggested? "Ubuntu on Chromebook" isn't "Ubuntu on ChromeOS", it's a straight up dual boot with ChromeOS and Ubuntu.

  • I'm pretty intrigued by these new chromebooks, and I am seriously considering getting one for my wife. She mostly just does web surfing, facebooking, and email checking, so I think it would be fine for her needs.

    • by admdrew ( 782761 )
      I'd definitely agree. I received my ARM-based Chromebook last week, and have played around with both ChromeOS and manually throwing Ubuntu on there - for the casual user looking to do some regular ole web surfing, ChromeOS is great. The machine itself is great too; form factor is awesome, doesn't overheat, keyboard is amazing (even if it's a ripoff of Macbooks, it's a great ripoff).
  • Of course a device built to run, linux derivative, chromeOS runs linux faster then a device built to run windows!
    • by admdrew ( 782761 )
      What? Atom-powered machines aren't "built to run windows", they're just simply x86 processors, which common Linux distros and Windows both happily work with.
  • Why are we comparing against antique Atom chips? Aren't AMD's Brazos 2.0 chips significantly faster at half the power draw?
    • by admdrew ( 782761 )
      Agreed, but - I suspect it's due to the ubiquity of Atom-powered machines; Brazos-based machines are still faaaaaar less common.
  • There's something not quite right about these benchmarks. A huge margin in FFTE is completely reversed on Apache. Often you can normalize this a bit by knowing which chip has how many cores and whether the floating point unit sucks or doesn't suck.

    This discrepancy is more extreme than normal. Usually you find out that one chip or the other was hobbled by software indigestion, then the discrepancy dissipates in subsequent rounds.

  • at a staggering 12-30fps on minimum settings


Real Programmers don't write in PL/I. PL/I is for programmers who can't decide whether to write in COBOL or FORTRAN.