8-Way Spring 2020 Linux Distribution Performance Comparison With 240+ Benchmarks

[Michael]

i'm sure fedora had selinux enabled

Yes each OS at defaults.

[pal666]

Tell you what, the lead Clear Linux is showing in the codec space is making it economically desirable for a video transcode & quality farm.

for video transcode farm you'd just recompile your transcode app targeting your cpu and get it fast on any distro

[V1tol]

I was suprised Manjaro is so slow in this benchmarks. I tried to use Ubuntu and Fedora on my notebook and Manjaro feels to be the most responsible. No benchmarks, just daily developer+gamer usage. And I always disable every security things in Linux distros (Intel is already as slow as Pentium 3 with those and just because I can, unlike on Windows or macOS), and currently in Manjaro I have "mitigations=off" and "apparmor=0".

[240Hz]

I wonder how much better Opensuse would be if you chose XFS for root and home at the installation that btrfs

[skeevy420]

i'm sure fedora had selinux enabled

It did, but somewhere I read, I think it was a post from here actually, that was talking about how apparmor is slower than selinux for certain contexts. I'm hoping someone who knows more about the two will chime in.

I also noticed that Manjaro has the least amount of options set for their EXT4 fstab compared to the others and wondered if that matters.

V1tol

I switched from Ubuntu 20.04 to Manjaro 20.0 KDE last night. Night and day difference in how much slower Manjaro is to start. Logon screen to desktop is an extra 20 seconds with Manjaro...seriously...it's stupid slow and I'd link to their forum posts discussing how 20.0 is slower to log on if I wasn't lazy and felt like searching for it again.

Once it's running, I agree that it actually does seem faster and more responsive, but the slowness in start up and logging on is very damn noticeable.

[torsionbar28]

i'm sure fedora had selinux enabled

As it should be. Selinux is a powerful security tool, and costs only a few percents in the benchmark numbers. Disabling it is foolish. I wonder how many folks will complain about Fedora being 2% slower in this thread, and then hop over to an intel chip review, and trash intel for pursuing performance over security.

[tildearrow]

Fedora did pop back up behind Clear Linux when the SVT-VP9 video encode performance.

Apparently you wanted to finish this sentence, but when you.

[skeevy420]

I must be in the wrong hemisphere

Just remember that the X-ray is her siren song if you're looking for the way out else you'll pass into Olympus.

[Mario Junior]

As it should be. Selinux is a powerful security tool, and costs only a few percents in the benchmark numbers. Disabling it is foolish. I wonder how many folks will complain about Fedora being 2% slower in this thread, and then hop over to an intel chip review, and trash intel for pursuing performance over security.

And another thing: disabling selinux will not change things so much, since it is not only selinux that offers an extra layer of security, but also the various security flags used in compiling packages on Fedora.

[Random_Jerk]

Note 2:
I understand why they are tested with the default governor that they ship with but it would be interesting a round of testing where they have a performance governor set for each.

On one of my old desktop machine with an Intel i5-23xx chip, Manjaro had the powersave governor on by default. Felt it was an interesting choice. Don't know what other distros do.