after I’ve added support for Clear Linux containers to distrobox, I thought that it would be a nice experiment to show the performance benefits of Clear Linux even on non-Clear distributions.
In this article on Phoronix it’s discussed the use of distroboxes to augment a Clear Linux host’s software availablility. Now I wanted to check the benefits of running a Clear Linux distrobox on a normal host distribution.
So here is a series of benchmarks performed with the Phoronix Test Suite.
Bear in mind I have an old laptop with no dGPU so I didn’t do anything graphical and due to my poor connection I wasn’t able to download the bigger tests
I’ve tested on a Fedora 35 host (all latest updates installed as of 20/01/2022), included also a container of Fedora 35 to show a comparison of eventual Podman overheads (little to none )
Full benchmark here: https://openbenchmarking.org/result/2201215-NE-CLEARLINU53
Clear Linux finishes first 90% of the times, the F35 container also is in the range of measurement error with the host, showing that running in Podman is not influential for these benchmarks.
In almost all tests Clear Linux shows a mild to significant performance gain (5% up to 15%, except libavif test with 50%+), even without the kernel optimizations/patches that you’d have running a Clear Linux host.