Last week Friday the unfortunate news came down that Intel was discontinuing their Clear Linux project effective immediately. For the past ten years Intel software engineers have been crafting Clear Linux as a high performance distribution that is extensively optimized for x86_64 processors via aggressive compiler tuning, various patches to the Linux kernel and other packages, and a variety of other optimizations throughout the operating system. For years Clear Linux has led Linux x86_64 performance not only on Intel desktop/mobile/server hardware but on AMD systems too. Here is a final look at the Clear Linux performance on the Intel side compared to the performance of the latest Ubuntu 25.04 release.
Ending the Clear Linux project was a heavy hit to the open-source world and for those caring about Linux software optimizations but not too unexpected given Intel’s financial difficulties building up over the years. Clear Linux assets remain available but are no longer being maintained. Before falling too far out of date, I decided to run some final benchmarks of Clear Linux. In this article are tests on the Intel side while I will also have similar benchmarks on AMD too for showing how Intel’s Linux software optimizations even helped its primary rival.
While Clear Linux runs on all modern Intel platforms even low-power notebooks, Clear Linux is heavily optimized for AVX-512 and other newer ISA features of Intel processors. So with wanting to test on an AVX-512 enabled system, it meant going for an Intel Xeon server as what was one of the primary targets anyhow for Clear Linux development. It would have been great to showcase the final Clear Linux benchmarks on their current-generation Granite Rapids processors but sadly my Avenue City reference server remains non-operational and haven’t been able to source any new motherboard or server… Not even able to find any Granite Rapids motherboard retail in-stock for adapting a new GNR server build. The unfortunate sign of the times and why there haven’t been any new Granite Rapids Linux benchmarks on Phoronix since the end of last year.
So in evaluating other Intel Xeon server options in the lab and wanting a retail platform to showcase real-world CPU power use of Clear Linux, I settled on Intel Xeon Max (Sapphire Rapids) using the Supermicro Hyper SuperServer SYS-221H-TNR / X13DEM. That Supermicro Xeon Max server continues running well with solid performance and reliable so that was used for carrying out these final benchmarks. The Supermicro Hyper SuperServer SYS-221H-TNR / X13DEM was running with dual Xeon Max 9468 processors, 512GB of RAM, and 7.6TB INTEL SSDPF2KX076TZ NVMe SSD.
Clear Linux 43760 in its final state shipped with the Linux 6.15.5 kernel, GCC 15.1.1, Python 3.13, and other up-to-date software components for this rolling-release distribution.
For a baseline to evaluate the Clear Linux performance was Ubuntu 25.04 as the newest Ubuntu Linux release from Canonical. Ubuntu 25.04 ships with Linux 6.14, GCC 14.2, Python 3.13, and other newer software versions than what is available with Ubuntu 24.04 LTS.
Ubuntu 25.04 on this server defaults to the intel_cpufreq scaling driver with the “schedutil” governor. Ubuntu still frustratingly defaults to schedutil or powersave governors even on servers… A rather silly default many will argue. As that can be a significant difference right there while Clear Linux defaults to intel_cpufreq performance, I carried out Ubuntu 25.04 runs on this Intel Xeon Max server both with Ubuntu 25.04 at its defaults and then again at the defaults but having switched from the scheduler utilization CPU frequency scaling governor over to the performance governor in matching Clear Linux to rule out that difference.
The combined dual socket Xeon Max 9468 power consumption was also monitored for each test in seeing the impact Clear Linux’s performance optimizations have on the CPU power consumption.
