The Performance Of Six Linux Distributions On The HP Dev One

[hax0r]

Firefox performance on Ubuntu 22.04 is just terrible, its users just got snap'ped.

I was also expecting better Python performance out of OpenSuse, kind of disappointing seeing it lose a lot.

[gbschenkel]

Would love see the benchmark with Gentoo and Slackware on this machine too.

[DanL]

Firefox performance on Ubuntu 22.04 is just terrible, its users just got snap'ped.

Agreed

I was also expecting better Python performance out of OpenSuse, kind of disappointing seeing it lose a lot.

Meh, should be good enough for most desktop users and the casual python script. Hardcore Python devs probably want specialized setups anyway.

[Linuxxx]

Michael

Thanks, these results are quite interesting for a variety of reasons:

First of all, it clearly (as in Clear Linux) shows that Linux users have no reason whatsoever to fear the performance governor, even on battery-powered devices.
It also demonstrates that even on an APU with a shared power-budget between the CPU & GPU, the performance governor can boost frame-rates & -times quite a bit.

Which leads me to a benchmark request I'm surprised You haven't conducted yet:

How are the different CPU governors affecting the Van Gogh APU inside the Steam Deck?
And how about enabling the elusive AMD-PSTATE driver on this device, since Valve still prefers to stick with acpi-cpufreq by default?

I'm certain such a benchmark would be of utmost interest to a sizeable audience, with the Steam Deck becoming a more mainstream product by the day.

Anyhow, hope You can make that happen -- otherwise, I'm afraid no-one else will...
[Cue that Dark Knight quote about the hero we need, but don't deserve! ]

[Jahimself]

Good to know that if you buy a laptop preinstalled with PopOS it will perform reasonably well compared to other distributions, without requiring particular tuning. Meanwhile clear linux keeps flying I would love to be tech savy enough to be able to create a Clear Linux gaming distro. Would also love to see how it performs with gaming on a fully updated AMD & Intel platform, as well as againt windows 11 with like 2 or 3 cards from AMD and Nvidia.

[reba]

I'm curious how a setup of stock Debian unstable + xanmod-edge-amd64v2 kernel would fare.

[Evorster]

As this laptop uses an AMD CPU, the correct zen kernel would have been more appropriate for this test in Arch.

[Evorster]

As this laptop uses an AMD CPU, the correct zen kernel would have been more appropriate for this test in Arch.

Just did a bit of digging. Proper Arch kernel for your laptop should be linux-amd (more info: https://aur.archlinux.org/pkgbase/linux-amd)

Thanks for all the great testing you guys do, your site was a great help while I was looking for my laptop.

[Eumaios]

Michael thank you very much for such an extensive and informative benchmark comparison.

Can anyone suggest why Clear shows only average performance in the Perl benchmarks, as well as consuming more power and producing more heat? Clear seems otherwise uniformly strong, both in performance and efficiency.

Thank you,
Eric

[Vorpal]

I did some performance analysis of the zstd issue and ended up filing an upstream bug report (https://github.com/facebook/zstd/issues/3163).

The TL;DR is that zstd provides multiple build systems (CMake, Meson or plain Make are the ones relevant for Linux). For some odd reason, zstd only scales properly with number of threads when using the make build system. Arch Linux uses CMake however, unlike ubuntu. (I have not checked what Clear Linux etc use).