Catch up on stories from the past week (and beyond) at the Slashdot story archive

 


Forgot your password?
Close
typodupeerror
×
Intel AMD Ubuntu Windows Linux

Intel's 'Clear Linux' Distro Beats Ubuntu and Windows 10 -- on an AMD Laptop (msn.com) 51

Posted by EditorDavid from the benchmark-battles dept.
An anonymous reader quotes TechRadar: Intel's Clear Linux distribution looks like it could be the best operating system to run on cheap AMD hardware, with benchmarks showing it outperforms Windows 10 and Ubuntu on a $199 laptop with a budget AMD Ryzen 3200U processor. The Phoronix website ran a series of benchmarks on a super-cheap AMD laptop from Walmart, and found that Intel Clear Linux beat popular Linux distros Fedora and Ubuntu for 78% of the tests.

Not only is it remarkable that a relatively unknown Linux distro is so easily outperforming established operating systems, the fact that Intel is the company behind the distro is particularly ironic. As you can imagine, Clear Linux is optimized for Intel processors, but it seems like it works brilliantly on AMD hardware as well.
This discussion has been archived. No new comments can be posted.

Intel's 'Clear Linux' Distro Beats Ubuntu and Windows 10 -- on an AMD Laptop More

Intel's 'Clear Linux' Distro Beats Ubuntu and Windows 10 -- on an AMD Laptop

Comments Filter:

  • a fine line (Score:5, Insightful)

    by redback ( 15527 ) on Saturday February 08, 2020 @06:39PM (#59706026)

    its a fine line between optimised for intel cpus, and just plain optimised.

    • It's probably just compiled using Intel's compiler instead of GCC.

      • That's what I was thinking, but it is also still completely possible that Intel has expertise to do this with plain old gcc and a few tweaked flags. "Benchmarking" isn't a big target for most major Linux distros.

      • Re: (Score:3)

        by lgw ( 121541 )

        Are there AMD optimization flags for the Intel compiler? It wouldn't surprise me, but I haven't heard of such.

        Now, when are we getting the Microsoft Linux distro so we can have a fair comparison with the MS compiler?

        • Re: (Score:3)

          by gtall ( 79522 )

          You nuts, why would you want Microsoft to bastardize a Linux distro so it can phone home with all your vital "statistics" and be embedded with Microsoft malware?

        • There was a period when the Intel compiler produced pessimal code for AMD processors. The company came under antitrust scrutiny from that practice so they had to stop. There are also fewer differences between optimal code for Ryzen and recent generations of Intel Core than there were in earlier generations.

      • Re: a fine line (Score:5, Insightful)

        by m.dillon ( 147925 ) on Sunday February 09, 2020 @02:05AM (#59706674) Homepage

        I don't know the answer to that (which compiler they are using), but it almost doesn't matter. Modern CPUs do such a good job optimizing the instruction stream that is handed to them that vendor-specific micro-optimizations don't really do a whole lot these days. Over the years we have removed most of them because they just don't do anything any more.

        I can think of only one micro-optimization that is vendor-specific, and that's Intel's optimization of stosb and movsb which is designed to support generic memcpy() and memset() operations. But it won't matter on a machine with this little memory.

        All other optimizations are relatively agnostic and apply to both CPU vendors. Just common sense, really. Doing things like avoiding certain pushq/popq sequences (which has to manipulate %rsp) in favor of pre-adjusting %rsp and doing normal movq's, aligning code and certain branch targets, collapsing call/ret sequences, linker-based inlining, and so forth.

        -Matt

        • I don't know the answer to that (which compiler they are using), but it almost doesn't matter. Modern CPUs do such a good job optimizing the instruction stream that is handed to them that vendor-specific micro-optimizations don't really do a whole lot these days. Over the years we have removed most of them because they just don't do anything any more.

          I can think of only one micro-optimization that is vendor-specific, and that's Intel's optimization of stosb and movsb which is designed to support generic memcpy() and memset() operations. But it won't matter on a machine with this little memory.

          All other optimizations are relatively agnostic and apply to both CPU vendors. Just common sense, really. Doing things like avoiding certain pushq/popq sequences (which has to manipulate %rsp) in favor of pre-adjusting %rsp and doing normal movq's, aligning code and certain branch targets, collapsing call/ret sequences, linker-based inlining, and so forth.

          -Matt

          When I use Clang with Clear Linux, my compiled programs come out smaller and execute no slower than before. In fact, somewhat faster. The compilers are producing code that needs linking. The libraries that my code requires have been highly optimized by Intel. I would suspect that the Intel stuff uses new instructions that are common to Intel and to the latest AMD instructions (I have an AMD 2700x). If Clear Linux supported mp4 and mpeg codecs, I would switch to CL without hesitation.

    • Exactly. Why not show us how Intel's Linux compares to the other "established operating systems" on Intel Hardware?

      And as for that "so easily outperforming" claim, the differences between Clear Linux and other Linux OSes look more like rounding errors that any sort of dramatic or meaningful difference.

  • Slackware (Score:3)

    by packrat0x ( 798359 ) on Saturday February 08, 2020 @06:47PM (#59706046)

    Have they tried Slackware [slackware.org]?
    Any test of distros is incomplete without a baseline.

  • Intel has this stupid fixation on Gnome, which already bit them in the ass before when they got involved in crushing the formerly vibrant Maemo project in part by changing out its excellent KDE desktop for the monumentally brain damaged Gnome project, that old shuffling zombie duck taped together with horrid GTK and libglib crap in marginally maintainable C-trying-to-OOP. Did Intel learn? Nope.

    Well, not much. If you install Clear Linux "desktop" package you get Gnome. If you install "desktop-kde" you get KD

    • i agree, i dont like gnome-3.x and i checked distrowatch and i notice clear linux uses the RPM package manager, i would rather see it use something like Slackware's tgz/txz, and i dont like split packages so if i want to install all of libusb or libjpeg i dont have to also find libusb-dev and libjpeg-dev all that crap debian does splitting packages into separate components is not only making more work out of it they are also annoying to users that want the whole package intact

      • i notice clear linux uses the RPM package manager

        Yep, Intel didn't learn a thing. Install this PHB inspired junk for entertainment only. Don't expect it to last: conceived by PHBs, it will end when somebody changes the powerpoint colors.

      • i would rather see it use something like Slackware's tgz/txz, and i dont like split packages so if i want to install all of libusb or libjpeg i dont have to also find libusb-dev and libjpeg-dev all that crap debian does splitting packages into separate components is not only making more work out of it they are also annoying to users that want the whole package intact

        You can't be serious. You actually think that dependencies don't matter and your entire OS should a monolithic install?
        A critical flaw in a single package like openssl and what do you do? Reinstall the entire operating system?

        • i am serious about package management, but another poster said they dont use RPMs anymore, i guess distrowatch needs an update, i used to use slackware for years, but have since given up waiting for slackware-15 to be released, and switched to devuan, i wont install clear linux i was just commenting on it only
      • I've used it recently. It does not use RPM for package management. It uses swupd or flatpak.
    • It also comes with systemd.
  • People aren't buying computers because they don't need upgrades to do the things they want to do with computers. This also means benchmarks don't matter as much and marginal difference in benchmarks matter less than ever. Everyone's too busy with data mining mobile platforms to focus on real computing. Just keeping people clicking on stuff and data mining them is way more profitable than actually thinking up real ideas.

  • Clear is a light-weight Linux optimised for the x64, in particular running as a virtual machine on the cloud. It's not surprising that that it out performed a full Ubuntu desktop install or a Fedora Workstation by a bit. Especially at Python and graphic operations optimisation counts. There are probably many other light-weight Linuxes with similar performance though. They only bench-marked these ones because they are popular corporate Linuxes.

    Having watched a friend's Celeron netbook struggle with Windows 1

  • Too bad... (Score:4, Insightful)

    by dicobalt ( 1536225 ) on Saturday February 08, 2020 @08:48PM (#59706202)
    it uses the Windows 8 of Linux... GNOME.

  • That laptop only has 4 GB of RAM (Score:3, Insightful)

    by Solandri ( 704621 ) on Saturday February 08, 2020 @09:17PM (#59706272)
    As with the previous benchmark dick-waving contest [slashdot.org], the distinguishing factor is probably how well the OS runs in 4 GB of RAM, not how well the benchmark runs on the OS. Unless you actually plan to use the OS with just 4 GB, comparative benchmarks on a rig with 16 GB of RAM would be more informative.

    • Virtually guaranteed to be the case. You could probably make any linux distro match it just with a little cache-tuning for low-memory. Lots of bits and pieces of the Phoronix test suite are going to give skewed results based on available ram due to filesystem caching and other effects.

      -Matt

  • What's ironic (Score:3)

    by drinkypoo ( 153816 ) <drink@hyperlogos.org> on Saturday February 08, 2020 @11:43PM (#59706478) Homepage Journal

    "Not only is it remarkable that a relatively unknown Linux distro is so easily outperforming established operating systems, the fact that Intel is the company behind the distro is particularly ironic."

    What's particularly ironic is that Intel could have had a dominant Linux distribution a long time ago with Moblin, which instead begat Meego. But they explicitly made it only work on Intel processors, and nobody seemed to want to do the work to put the AMD parts back in, and it withered.

    Now that Intel is using AMD's instruction set, maybe Intel will be less uppity. AMD is now the performance leader again, as they were when Athlon came out. Intel is playing catch-up. And there's no indication that they will in fact do so any time soon despite their anticompetitive (oh look, that word isn't in gboard's dictionary, this is my surprised face) behavior.

  • Call me, when their thing has matured, hardened and stabilized to the point that those OSes have.

    Cause *of course* it's gonna be faster and leaner, when you haven't added all the special cases and quirk workarounds yet because you are too new.

    And to argue this on an *Intel* product, of ALL things in the world!

  • Do people who buy the cheapest laptop available really care about a few percent performance difference?

    • People that buy a $200 laptop for one of two reasons: a) idle fascination at what you can get for $200 these days, or b) they need a Windows laptop and can only afford $200.

      The former would invest hours to squeeze even the tiniest performance boost for bragging rights, the latter will never consider running anything other than Windows, and would never consider upgrading the RAM (likely the best way to improve the laptop).

      I recently bought a $100 laptop because it had a QHD display AND I could add a SSD to e

  • Really, Intel has interest in Linux for a long time, longer than I've got direct knowledge of.

    I've known a few engineers at Intel, even did a brief temp rack+stack project for them which was huge fun, and among the more interesting things I've becomer aware of:

    - Intel takes driver development seriously.
    - Intel was fussing over drivers for their RST/RSTe hardware a ways back, mostly to maintain market parity I expect.
    - They were committed to Linux support way back when, .
    - And Intel might be developing a bit

Slashdot Top Deals

A triangle which has an angle of 135 degrees is called an obscene triangle.

Close