Follow Slashdot stories on Twitter

 



Forgot your password?
typodupeerror
×
Windows Red Hat Software Ubuntu Linux

Linux Distros Beat Windows 11 in Phoronix Performance Testing (phoronix.com) 58

Phoronix ran some fun performance tests this week. "Now that Windows 11 has been out as stable and the initial round of updates coming out, I've been running fresh Windows 11 vs. Linux benchmarks for seeing how Microsoft's latest operating system release compares to the fresh batch of Linux distributions." First up is the fresh look at the Windows 11 vs. Linux performance on an Intel Core i9 11900K Rocket Lake system... The Windows 11 performance was being compared to all of the latest prominent Linux distributions, including:

- Ubuntu 20.04.3 LTS
- Ubuntu 21.10
- Arch Linux (latest rolling)
- Fedora Workstation 35
- Clear Linux 35150

[...] Each operating system was cleanly installed and then run at its OS default settings for seeing how the out-of-the-box OS performance compares for these five Linux distributions to Microsoft Windows 11 Pro...

The geometric mean for all 44 tests showed Linux clearly in front of Windows 11 for this current-generation Intel platform. Ubuntu / Arch / Fedora were about 11% faster overall than Windows 11 Pro on this system. Meanwhile, Clear Linux was about 18% faster than Windows 11 and enjoyed about 5% better performance overall than the other Linux distributions.

Out of 44 tests, here's a breakdown of how many first-place wins were scored by each OS:
  • Clear Linux: 33 (75%)
  • Fedora Workstation 35: 4 (9.1%)
  • Windows 11 Pro: 3 (6.8%)
  • Ubuntu 20.04.3 LTS: 2 (4.5%)
  • Arch Linux: 1 (2.3%)
  • Ubuntu 21.10: 1 (2.3%)

This discussion has been archived. No new comments can be posted.

Linux Distros Beat Windows 11 in Phoronix Performance Testing

Comments Filter:
  • Dupe (Score:4, Insightful)

    by williamyf ( 227051 ) on Saturday October 30, 2021 @12:37PM (#61942365)

    this was published a few days ago. Slow news saturday?

  • Vista and/or ME. Something hastily thrown out to maintain brand recognition and keep people on an upgrade treadmill.
    • ME was never used by sane people as 2000 was out first. I refused to take up "kids edition" windows xp until DX stopped getting 2000 updates.

      Vista was fine from SP1 and muchlyu required becuase XP 64 was unsupported and a joke (and yes I had rthe ram to make it useful and no, pagiong on the server XP was not supported well by games.

      11 is great. It is nothing much differnet it may as well be called 10 but they got scared by apple. It is not a new os, just an update. Yes explorer crashes all the time on right

      • ME was never used by sane people as 2000 was out first. I refused to take up "kids edition" windows xp until DX stopped getting 2000 updates.

        Vista was fine from SP1 and muchlyu required becuase XP 64 was unsupported and a joke (and yes I had rthe ram to make it useful and no, pagiong on the server XP was not supported well by games.

        11 is great. It is nothing much differnet it may as well be called 10 but they got scared by apple. It is not a new os, just an update. Yes explorer crashes all the time on right clicks but hey ho not much trouble caused. Far more positives to talk about. Like working HDR for one.

        Vista's biggest issue was it didn't work on machines that they said it would work on. Many bought Vista ready Vista basic machines, and they simply weren't powerful enough. My sister bought my father a new Vista basic Desktop to replace the nicely functioning XP box I set up for him. The thing stunk. A group I worked with forced me to implement a Vista Basic system. Likewise, didn't work for crap. One of the guys on the decision team was a friend of Microsoft, so he convinced them it was great stuff. Are

        • Vista's biggest issue was it didn't work on machines that they said it would work on. Many bought Vista ready Vista basic machines, and they simply weren't powerful enough.

          Microsoft allowed it under pressure from Intel, who had a glut of outdated Intel Graphics chipsets to sell.

          • Vista's biggest issue was it didn't work on machines that they said it would work on. Many bought Vista ready Vista basic machines, and they simply weren't powerful enough.

            Microsoft allowed it under pressure from Intel, who had a glut of outdated Intel Graphics chipsets to sell.

            That didn't work out too well

        • Vista's biggest issue was it didn't work on machines that they said it would work on.

          Vista's biggest issue was that it was a pile of shit. Windows 7 is faster and uses less resources. If Vista were not a pile of shit then it would have worked on those machine. Windows 7 is practically nothing more than a bugfixed Vista (there are minor additions but in general that's what it is) and Windows 7 has dramatically lower requirements than Vista does. Vista won't run properly even with 2GB RAM, for example, while Windows 7 does okay with only 512MB, and that's probably the biggest problem with Vis

          • The biggest issue most machines had with vista was the new driver model. Even though windows 10 is more efficient most of the problems people had it were resolved because more drivers were updated to be compatible with 64-bit vista by the time windows 10 was out.
            • The biggest issue most machines had with vista was the new driver model.

              That was an issue for printers and scanners which speak dumb protocols requiring the PC to do all the work, because their drivers often weren't ported. But it wasn't really an issue for PCs themselves because statistically nobody upgrades to a new Windows version any way other than buying a new PC.

              The biggest problem with Vista was you'd buy a PC with a sticker on it saying it was good enough for Vista and then Vista would run like dogshit anyway because for some reason it was ludicrously memory hungry. Win

    • Re: (Score:2, Interesting)

      by thegarbz ( 1787294 )

      I'm genuinely curious how you can see a comparison between Windows 11 and Linux, and with it conclude something about Windows 11 vs a completely different system.

      Oh wait. rsilvergun post... No never mind, I'm not curious anymore.

      But in case you want to engage your last remaining braincell Windows 11 and Windows 10 are within a rounding error / run-to-run variance of each other performance wise in basically every benchmark with the exception of and AMD bug. The fact that Linux is faster in many benchmarks ha

      • Or did you just want to do a personal attack on me for the hell of it? Windows 11 is rife with performance issues. It's like Vista and Millennium were. This article is about performance. It shouldn't take the amazing Keskin to put two and two together, so I didn't think I needed to spell out the connection. I stand corrected.

        Also just out of curiosity, any chance you're running one of the alts it's been impersonating me? I don't mind I'm flattered and all...
      • by fferreres ( 525414 ) on Saturday October 30, 2021 @08:58PM (#61943527)

        What percentage of Top 500 supercomputers does Linux have? 100%
        What percentage of mobile phones run Linux? 81%
        Which one is the wealthiest company in the world? Microsoft.
        What desktop OS is faster? For that we have this benchmark.

        What is your complaint?

        • Maybe re-read my post before angry replying something irrelevant. Why are you comparing Windows 11 to Linux? The OP clearly was comparing Windows 11 to Vista/ME and other releases of Windows.

          With that context, maybe you'll understand my complaint.

      • by lsllll ( 830002 )
        Man, right now I'm just hoping that lsllll is not as memorable as rsilvergun. I don't want the wrath of thegarbz on me!
        • Wrath? Dude it's an internet forum. All I do is throw around words and call people stupid. If this "wrath" worries you I suggest eating a cup of cement and hardening up.

          Or ... just post something that makes sense unlike rsilvergun. Either way. Though I recommend the cement in any case.

    • Sure, if you completely ignore what Vista represented. Namely a major overhaul of key Windows subsystems and a near total rewrite of the GUI from a 2D software rendered system to a 3D accelerated app.

      Now granted MS made a few unforced errors like the whole "Vista Ready" thing which was just flat misleading if not outright dishonest. Still, without Vista, Windows 7-11 wouldn't have been possible. And in the case of 11, it just feels like they've been rewriting big chunks of the non-user facing parts of the O

  • Windows 11 is for me such a let down, far worse than W8 after W7. I'm not sure why people still don't look for the alternatives to such an arse rape. All the time, since Windows 7 actually, but it can be "tamed" at least. I use extremely "lite" version of W7 just to be able to produce music and it works well. I would like some kind of a RTOS for music, but that's not gonna happen, obviously. Otherwise Debian Linux with MATE all the way for 8 years now.

    A simple font rendering, that is readability of fonts co

    • Windows 11 is for me such a let down, far worse than W8 after W7. I'm not sure why people still don't look for the alternatives to such an arse rape.

      Maybe because there's no actual arses being raped. If you want to compare windows releases why not actually look at windows benchmarks rather than Linux benchmarks. There you may find that other than for a since fixed AMD bug Windows 11 has the same performance as Windows 10 within run-to-run variance and when it differs it do so by a rounding error in a fraction of a percent.

      I use extremely "lite" version of W7 just to be able to produce music and it works well.

      I hope you anti-vaxxers aren't on the internet with your unpatched buggy operating systems.

      A simple font rendering

      You're not talking about Linux are you? If

  • This appears to be more like an advertisement for Clear Linux. I wonder if this test was conducted with Intel mitigations enable.
    • by andydread ( 758754 ) on Saturday October 30, 2021 @03:01PM (#61942723)
      Well it's not news that Clear Linux with Intel's own optimizations vastly outperforms all other Linux on Intel hardware. And even on AMD hardware it's much faster than the rest of the lot.
      • by heson ( 915298 )
        Since when is 5% vast? Most common wins were by less than a percent and a few by about 10% .
        • by t0rkm3 ( 666910 )

          When you combine them as a whole, it makes quite a lot of difference. Consider a car analogy? Depending on the combination of attributes:

          The passenger's compartment is 1% more comfortable
          The trunk is .75% more spacious
          The ride is .05% less noisy
          The all-wheel independent braking is 7% more safe
          The engine is 10% more powerful
          The gas mileage is 0.4% greater

          The price is significantly lower and the car itself is easier to maintain as the components and engine compartment are designed to be worked on.

          Which would

    • The tests were utterly rigged. windows is on power save other lines on normal clear lines on performanceâ¦. The author is known for doing these gimmicks
      • All the other Linux distros are also on powersave, this is however not something that Michael have done deliberately, he is just running with the defaults and Clear Linux runs with the Performance governor by default. What Windows is on I don't know since it's not listed and I have zero clues on what it uses as default.
      • Yeah, without some transparency on the cpu clock behavior or even some mention of the cooling/tdp limits it's hard to say if this was rigged accidentally or not. Often benchmarks will score best on "balanced" on windows due to idle dropping the clocks which allows the benchmarks to run for longer in the short performance limits.

        Of course, there's also the possibility that windows 11 is causing bad performance due to the big/little support for alder lake apparently causing problems for other processors, mos

  • I'm not interested in the difference in results from some bollocks artificial benchmark or some random open source render or proprietary one I've never heard of that I'll never use, I want to know what the difference is in something like the FPS of my favourite games. And the answer to that currently is 100% in the favour of Windows because despite the advancements in Linux gaming and getting Windows games to run on Linux none of my top 5 currently do.
  • It's really impressive how much better Clear Linux's results are compared to the other distros. I wish those performance improvements would be integrated into other linux-distributions. Clear Linux is open source. Why don't they just copy over the best improvements? (Note that Clear Linux is still using gcc and clang, not Intel's icc compiler)
  • Really, and now?

    This is good? or is this bad? Hasn't this been this way since ever?

    I'll still go with windows 11 just because I have a use case there. And yes, I will still go with debian or ubuntu, because I have a use case there as well.

    Have a great day, and thanks for testing and thanks for breaking it as such big news..
    Have a great weekend.
  • by jdeisenberg ( 37914 ) on Saturday October 30, 2021 @06:44PM (#61943309) Homepage
    I am curious to know why they used the geometric mean instead of the arithmetic mean. Is this common in performance testing?
    • by piojo ( 995934 )

      I wonder that as well. It seems like it would let individual results have outsized influence over the final outcome.

      I gather arithmetic mean is used when the results logically add together. Harmonic mean is used with many questions involving ratios. Geometric mean is used when the numbers are logically multiplied (they're used as factors), such as comparing growth rates. What's the rationale for benchmarks?

      • by piojo ( 995934 )

        Side note: if there's no reason to use a geometric mean, it may indicate the author massaged the results by picking the average that showed the desired results. I hope that isn't the case.

  • Most of the tools are cross platform open source with primary development on Linux. So, of course. A bit surprised that OpenJDK is showing that much favor to Linux these days, but given how common that use case is, again, it makes perfect sense.

    It would have been nice to know what the build tools and settings were. I've seen open source tools that still use very old Visual Studio build tooling (like 2012 old) and that is a handicap.

  • Out of 44 tests, here's a breakdown of how many first-place wins were scored by each OS:

    Out of 44 (random) tests, the OS that "won" the most tests is considered best? The results show that every OS "won" at least one (or more) tests.

  • Windows 11 as has always been the case in desktop operating systems from MS prioritizes tasks in the foreground. This means rapid task swapping, UI rendering and such.

    All the tasks I see being tested are batch tasks. These are items we would all be happy if they were faster, but the tests run are overwhelmingly things which would be processed by farm based systems.

    Yeh… Linux is faster than Windows for raw processing. But, Iâ€(TM)d be interested in seeing numb
  • I don't run an O/S to run benchmarks. I run an O/S to host PROGRAMS that I want to use.

    All the programs I want to use (mainly music applications) run on Windows. So I use Windows. As long as it loads the programs I want to use and gets out of the way I don't care what O/S I'm using. Everyone knows Linux is a better O/S than WIndows. But very few people use it because the programs they use don't run on it.

    The Linux desktop crowd spend all their time reinventing the desktop metaphor, moving wiodgets abou

  • The results actually show Windows 11 beating Linux distributions by a significant margin. The only exception is an experimental Linux made by Intel purely to demonstrate CPU optimisations. How this article can be interpreting the results in such a dumb way astonishes me. I am no Windows fan, switched to Manjaro a year ago and love it. Never going back to Windows (apart from having to use it for work).
  • This was all that needed to be proved!

In the long run, every program becomes rococco, and then rubble. -- Alan Perlis

Working...