Announcement

Collapse
No announcement yet.

Even With A $199 Laptop, Clear Linux Can Offer Superior Performance To Fedora Or Ubuntu

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

  • skeevy420
    replied
    Originally posted by Britoid View Post

    Did you not look at the recent benchmarks? The affect of SELinux is minimal.
    effect

    Leave a comment:


  • Britoid
    replied
    Originally posted by Volta View Post
    Does it have trash like SELinux and AppArmor disabled? They affect performance.
    Did you not look at the recent benchmarks? The a̶f̶f̶e̶c̶t effect of SELinux is minimal.
    Last edited by Britoid; 06 February 2020, 10:53 AM.

    Leave a comment:


  • Volta
    replied
    Does it have trash like SELinux and AppArmor disabled? They affect performance.

    Leave a comment:


  • wizard69
    replied
    Clear Linux might be fast but they are not friendly at all. Due to this article I was poking around the Clear Linux web site and tried to determine what version of the kernel and Mesa was currently shipping. Sadly the web site is one big suck! Maybe it is partly due to my knowledge of Fedora and the location of everything, but I can find all of that stuff relatively easily. Beyond all of that their approach to packaging seems to be a bit strained.

    In the end I'm not convinced there is huge advantages for me with Clear Linux. by the way the kernel and Mesa version could be having a significant impact on APU performance which is good reason to want to know that info.

    Leave a comment:


  • finalzone
    replied
    Originally posted by MastaG View Post
    It surprises me that Fedora does worse in all gaming benchmarks compared to Ubuntu.
    I always thought that Fedora would ship with a more up to date kernel/Mesa and even gnome/mutter?
    In majority of cases, Fedora release synchronizes with GNOME release. Both new kernel and Mesa needed to go through a quality check before getting into a stable repository.
    About gaming benchmarks, an average difference of 5 frames above 30 Hz is hardly a big deal considering the hardening features implemented in Fedora. Additionally, frame rate is only a portion of an entire context.

    Leave a comment:


  • MastaG
    replied
    It surprises me that Fedora does worse in all gaming benchmarks compared to Ubuntu.
    I always thought that Fedora would ship with a more up to date kernel/Mesa and even gnome/mutter?

    Leave a comment:


  • stormcrow
    replied
    Originally posted by set135 View Post

    I would say there two caveats here. One, Clear linux has shown better performance on all the x86-64 platforms it has been tested on by Phoronix, notably in this article, one of their main competitors low end apus. Two, they demonstrate that there are low hanging fruit that pretty much every distribution could be looking at that targets marginally recent x86-64 platforms... ie. I would not dismiss Clear as an impractical toy.
    I didn't dismiss Clear as a toy. Quite the opposite. I said it's a relatively small (in comparison to Intel's other research efforts in terms of funding) research project into what's possible if people are willing to make those changes. It is an advertisement for their hardware. That's part of the point. Yes those options show improvements to AMD hardware too, but that's not the primary emphasis with Clear Linux. They're upfront and honestly showcasing what their hardware can do even against their only real competition on the same software and optimization level. Macy used to use a similar advertising tactic against Gimbel Brothers back in the day. Macy upstaged Gimbel's by being upfront and honest with customers where to find products, even if it sent them to arch rival Gimbel Brothers if Macy didn't carry the product.

    Leave a comment:


  • set135
    replied
    Originally posted by stormcrow View Post
    [...] Clear Linux is one big advertisement for Intel hardware. It's a showcase and relatively small scale research project of what's possible should anyone care to tune for raw performance. That's all it's meant to be.

    The reality is generic distributions don't tune for raw performance. They conservatively tune for various ranges of compatibility and security settings instead, hopefully without sacrificing too much performance. That's not an unreasonable goal when you have no control over what hardware your software is going to be deployed on.
    I would say there two caveats here. One, Clear linux has shown better performance on all the x86-64 platforms it has been tested on by Phoronix, notably in this article, one of their main competitors low end apus. Two, they demonstrate that there are low hanging fruit that pretty much every distribution could be looking at that targets marginally recent x86-64 platforms... ie. I would not dismiss Clear as an impractical toy.

    Leave a comment:


  • skeevy420
    replied
    Originally posted by lowlands View Post
    The outcome is hardly surprising given the difference in focus between Clear Linux (pure speed) and Fedora. Here are the default compiler flags for Fedora 31:

    -O2 -g -pipe -Wall -Werror=format-security -Wp,-D_FORTIFY_SOURCE=2 -Wp,-D_GLIBCXX_ASSERTIONS -fexceptions -fstack-protector-strong -grecord-gcc-switches -specs=/usr/lib/rpm/redhat/redhat-hardened-cc1 -specs=/usr/lib/rpm/redhat/redhat-annobin-cc1 -m64 -mtune=generic -fasynchronous-unwind-tables -fstack-clash-protection -fcf-protection

    Not really Formula 1 settings I doubt that Clear Linux would still be faster if they recompiled their stuff with these flags.
    That.

    Defining an "-mtune=generic-aes" profile that's essentially Westmere in all but name would do wonders for modern desktops and laptops, especially on distributions where security is a major focus; Xeon Westmeres, Core i7-9xx, AMD FX Bulldozers & Opterons would be the minimum targeted CPUs. "-mtune=generic -march=x86_64", which is what is used if nothing is defined, doesn't enable AES and it's something that you'd want if you're using encryption and more hardened settings like Red Hat/Fedora.

    One of the "tricks" that Clear does is target Sandy Bridge and up (one gen newer than Westmere) because Sandy is the first Intel CPU generation with AVX support. I figure by going one generation earlier and targeting AES as a minimum, they (the distributions) won't have to worry about people who buy cheap and gimped Intel Celerons and Atoms and can say "CPUs released from 2010 and up are supported".

    Leave a comment:


  • stormcrow
    replied
    Originally posted by EarthMind View Post
    Clear Linux keeps kicking but. It's strange that they don't yet turn it into a commercial project yet by offering support services for it. It could be useful in specific cases where the best performance possible is required
    It won't be Intel that does it. Clear Linux is one big advertisement for Intel hardware. It's a showcase and relatively small scale research project of what's possible should anyone care to tune for raw performance. That's all it's meant to be.

    The reality is generic distributions don't tune for raw performance. They conservatively tune for various ranges of compatibility and security settings instead, hopefully without sacrificing too much performance. That's not an unreasonable goal when you have no control over what hardware your software is going to be deployed on.

    Leave a comment:

Working...
X