Announcement

Collapse
No announcement yet.

Linux Now Faster Than Windows 11 For Intel Core i9 12900K With Latest Kernel

Collapse
X
 
  • Filter
  • Time
  • Show
Clear All
new posts

  • vb_linux
    replied
    Originally posted by Danny3 View Post
    Wow, Ubuntu 22.04 definitely sucks for using such an old kernel by default for the time when it will be released!
    And thinking that is also full of snap crap it will probably be very hard to find a distro slower than that.
    IBM employees attacking Ubuntu. So, 5.15 is a really old kernel, Wow!! If they release 5.16 and encounter bugs then these same accounts will attack them again.

    Leave a comment:


  • edwaleni
    replied
    Originally posted by cl333r View Post

    What a loser, I was writing the Linux kernel in 1962 before Linus stole it from me.
    What are you, a 92 year old engineer from Bell Labs? Stand in line, there are few other Ma Bell retirees making the same claim.

    Leave a comment:


  • andyprough
    replied
    Originally posted by mitcoes View Post
    I do not understand, why there is no arch based distribution in this kind of benchmarks, when the Steam survey shows that Manjaro + Arch are about 1/4 as much or more than deb based distributions. And in enterprise liGnux RPM distributions as SUSE and RedHat are the kings.

    My suggestion is to add SUSE or RedHat, and Manjaro or Arch instead of other versions of Ubuntu when comparing MS WOS with GNU/Linux (liGnux), to be able to know more how liGnux perform vs MS WOS.
    Arch-based and rpm-based distros routinely lose multi-distro tests on Phoronix. What Michael should start doing again is throwing the latest and most prior versions of Debian into the tests, like adding Buster and Bullseye. Sometimes those showed fairly large wins in testing. It's harder to do though, since a lot of new hardware requires one of the latest kernels and a recent toolchain.

    Using Clear in tests of desktop distros and OS's is kind of a meme, since nearly no one uses Clear as a desktop distro. Intel no longer even tries to support it as a desktop distro, and the last survey of Intel engineers showed that almost none of them were using it as a desktop distro. Clear is out of place in these competitions.

    Leave a comment:


  • Volta
    replied
    Originally posted by HEL88 View Post
    Thaks. I didn't know that Ubuntu has Apparmor enabled by default. Maybe that's why it's slower than Clear Linux.
    It enabled by default, but AppArmor has rather tiny impact on performance. Ubuntu is using slower CPU governor: powersave, while ClearLinux is set to performance. Clear Linux has also additional optimizations. You can achieve this with Ubuntu, but you'll have to compile packages on your own.

    Leave a comment:


  • mitcoes
    replied
    I do not understand, why there is no arch based distribution in this kind of benchmarks, when the Steam survey shows that Manjaro + Arch are about 1/4 as much or more than deb based distributions. And in enterprise liGnux RPM distributions as SUSE and RedHat are the kings.

    My suggestion is to add SUSE or RedHat, and Manjaro or Arch instead of other versions of Ubuntu when comparing MS WOS with GNU/Linux (liGnux), to be able to know more how liGnux perform vs MS WOS.

    Leave a comment:


  • Azrael5
    replied
    By optimization, Linux Oses could be far better than windows oses also beyond ten times. The main matter is the way to implement the optimization in the easiest modality so to reduce the complexity. Clear linux continues to be the best Os among the linux Oses. This condition should arise from the ability of intel developers to make the Os CPu conformant to the CPu's instructions in order to get benefit from optimization. Obviously, hardware producers have such resources and above all knowledge to maximize the performance of their own software to their technological solutions.
    Last edited by Azrael5; 11 February 2022, 09:04 AM.

    Leave a comment:


  • Draget
    replied
    Interesting benchmarks. Of course, tuning those kernel things always is a balance. One might be faster 5% in one benchmark, but draw 20% more power in general. Having a power-draw comparison every now and then between the different loads would be interesting, too.

    Leave a comment:


  • HEL88
    replied
    Originally posted by Volta View Post

    It was done many times. Fedora has SELinux enabled. Ubuntu has Apparmor. It will be fair to make benchmarks with having it disabled.
    Thaks. I didn't know that Ubuntu has Apparmor enabled by default. Maybe that's why it's slower than Clear Linux.

    However, it probably wouldn't make much difference. The question is if CPU mitigations are the same on both system. They have higher impact.
    Probably similar.

    Leave a comment:


  • Volta
    replied
    Originally posted by schmidtbag View Post
    Good ol' Phoronix comments section: full of people who can't WAIT to prove how much smarter they are than you by whilst being a total prick.
    It's about not being lazy. Everyone can verify most of the things in few minutes.

    Leave a comment:


  • Volta
    replied
    Originally posted by HEL88 View Post

    Windows has too. But SELinux has much more complex .

    So maybe should bechmarked Windows vs linux with selinux implemented - it would be fair
    It was done many times. Fedora has SELinux enabled. Ubuntu has Apparmor. It will be fair to make benchmarks with having it disabled. However, it probably wouldn't make much difference. The question is if CPU mitigations are the same on both system. They have higher impact.

    Leave a comment:

Working...
X