A Quick Look At Ubuntu 20.04 LTS vs. 20.10 With The Core i9 10900K

Written by Michael Larabel in Software on 20 October 2020 at 06:13 PM EDT. Page 1 of 1. 3 Comments.

With Ubuntu 20.10 due for release this week I have begun testing near-final Ubuntu 20.10 builds on many more systems in the lab. Larger than our normal distribution/OS comparisons, here is the culmination of running hundreds of benchmarks (366 tests to be exact) under both Ubuntu 20.04 LTS with all available updates and then again on the Ubuntu 20.10 development state while testing on Intel Comet Lake.

Aside from specific improvements for bleeding-edge hardware like Intel Tiger Lake performing better on Ubuntu 20.10 or when looking at cases like the Intel and Radeon graphics performance being better on Ubuntu 20.10 due to the newer Linux kernel and Mesa, for general CPU/system workloads the performance has largely been found to be similar to that of Ubuntu 20.04 LTS.

The other caveat is for workloads being built from source, Ubuntu 20.10 now ships with GCC 10 rather than GCC 9. GCC 10 doesn't normally yield any night-and-day differences in performance but in some cases for newer CPU microarchitectures there has been some improvements there or with features like LTO.

Core i9 10900K - Ubuntu 20.04 LTS vs. Ubuntu 20.10

First up were results on an Intel Core i9 10900K Comet Lake system with Ubuntu 20.04 LTS compared to Ubuntu 20.10 as of a few days ago. The hardware remained the same on both Ubuntu releases tested. Given the vast number of benchmarks carried out, here is the high-level side-by-side view:

Core i9 10900K - Ubuntu 20.04 LTS vs. Ubuntu 20.10

The TTSIOD renderer regressed with GCC 10 but aside from that the most impacted workloads that were much faster on Ubuntu 20.04 LTS, as in regressing with Ubuntu 20.10, was Intel's oneDNN library. Running Intel's oneDNN library was much slower on Ubuntu 20.10. But while there is the compiler version at play, in a recent autumn 2020 Linux distribution comparison we also found oneDNN on Ubuntu 20.10 performing slower than other Linux distributions -- including those with newer package versions, like GCC 10. That other distribution comparison was also done on a completely different platform (Core i9 10980XE Cascade Lake X). So for reasons unknown right now, Intel oneDNN is slower on Ubuntu 20.10 -- compared to Ubuntu 20.04 LTS and other distributions on multiple systems.. This also isn't an Intel CPU specific issue but have reproduced it similarly on a Ryzen 9 3900XT system between these Ubuntu releases.

Aside from oneDNN, the other big swings like Hackbench are due to the usual kernel churn and synthetic workloads. There are a number of healthy improvements in the graphics tests from GLMark2 to the Vulkan compute benchmarks thanks to Mesa 20.2. Especially for Radeon GPU owners, Ubuntu 20.10 is a big release with Mesa 20.2 for newer graphics cards especially but also if you use the RADV Vulkan driver where ACO is now the default shader compiler back-end.

In any case given the wide range of tests, if interested in diving deeper into any particular workloads of interest/concern to you, find all of these Core i9 10900K Ubuntu 20.04 LTS vs. 20.10 benchmarks via this OpenBenchmarking.org result file.

Core i9 10900K - Ubuntu 20.04 LTS vs. Ubuntu 20.10

If taking the geometric mean of the hundreds of tests run, Ubuntu 20.10 takes about a 4% hit on this particular system but obviously your mileage may vary based on hardware and the workloads you actually make use of on a day to day basis.

Additionally, some smaller comparisons using just the more popular benchmarks/workloads across multiple systems will be coming up once the official Ubuntu 20.10 images are released.

If you enjoyed this article consider joining Phoronix Premium to view this site ad-free, multi-page articles on a single page, and other benefits. PayPal or Stripe tips are also graciously accepted. Thanks for your support.


Related Articles
About The Author
Michael Larabel

Michael Larabel is the principal author of Phoronix.com and founded the site in 2004 with a focus on enriching the Linux hardware experience. Michael has written more than 20,000 articles covering the state of Linux hardware support, Linux performance, graphics drivers, and other topics. Michael is also the lead developer of the Phoronix Test Suite, Phoromatic, and OpenBenchmarking.org automated benchmarking software. He can be followed via Twitter, LinkedIn, or contacted via MichaelLarabel.com.