Debian vs. Ubuntu vs. CentOS vs. openSUSE vs. Clear Linux Post-Meltdown Performance

[phoronix]

Phoronix: Debian vs. Ubuntu vs. CentOS vs. openSUSE vs. Clear Linux Post-Meltdown Performance

With Linux distributions being patched since last week's Meltdown and Spectre disclosure, here are benchmarks on some of the prominent distributions looking at their performance impact since being patched. Tested from an Intel Core i7 8700K system was CentOS, Clear Linux, Debian, openSUSE, and Ubuntu.

http://www.phoronix.com/vr.php?view=25839

[_Alex_]

I think one of the parameters that might affect system performance with the Meltdown patches, due to context switching, is 100hz/250hz/1000hz and pre-emption vs server (no preemption). The more context switches there are, the bigger the drop (especially in older hardware).

Perphaps a test of an identical kernel with preemption on / off, or different timer frequencies could reveal more about it, which would then lead distros to fine tune their distros according to the new data. Of course all these tests require time :P

[dungeon]

Debian has patched their Stretch, Jessie, and Wheezy kernels with KPTI for Spectre mitigation but Meltdown is still a work-in-progress.

S/M to M/S - permutation typo

I think one of the parameters that might affect system performance with the Meltdown patches, due to context switching, is 100hz/250hz/1000hz and pre-emption vs server (no preemption). The more context switches there are, the bigger the drop (especially in older hardware).

Perphaps a test of an identical kernel with preemption on / off, or different timer frequencies could reveal more about it, which would then lead distros to fine tune their distros according to the new data. Of course all these tests require time :P

Welcome at forum, everything require time... is that your first post ever 33 months after joining

[_Alex_]

Thanks... Yes... I think I tried to post when I registered back in the day, but something happened with the credentials or email - I don't even remember after all that time, heh...

[carewolf]

I think one of the parameters that might affect system performance with the Meltdown patches, due to context switching, is 100hz/250hz/1000hz and pre-emption vs server (no preemption). The more context switches there are, the bigger the drop (especially in older hardware).

Well the main context switch in those cases is between the two user processes, and that is unchanged. Though the overhead might indeed have increased as the logic looking up which process to switch in now has to work without cached virtual pages entries, but the overhead should be pretty minimal these days, I doubt it will make as big a difference as a process that does a lot of kernel calls and can get more changes between user and kernel during its timeslice than just the one at the end of its timeslice.

[geearf]

I'm still amazed at Clear Linux's performance, and wish the other distributions would follow suit (especially Arch that I use).

[Zoll]

I'm still amazed at Clear Linux's performance, and wish the other distributions would follow suit (especially Arch that I use).

It seems no one is taking steps to optimize the flags in the popular distros. Maybe it's too much work for them?

[norsetto]

Updated intel microcode was actually released by Debian through unstable recently (on 4 and 10 Jan).

[Djhg2000]

It seems no one is taking steps to optimize the flags in the popular distros. Maybe it's too much work for them?

I think a more likely reason is portability, you don't want to cause performance regressions on a binary distro. On a source distro like Gentoo people are putting a lot of effort into optimizing compiler flags because they're just local anyway.

[Steffen Bergner]

phoronix ... I would like to see a benchmarking of *lowlatency* & generic kernel now in the era of post-meltdown/spectre. OutOfOrder-Pipelining... is now limited(?), does lowlatency-kernel may have an positive impact? Just wondering...