Last week I ran the last planned benchmarks of Intel CPU performance on Clear Linux vs. Ubuntu with Intel having ceased development of Clear Linux following the restructuring at the company. In today’s article is a final look at how the AMD EPYC performance compares on Clear Linux relative to Ubuntu Linux and AlmaLinux. An AMD EPYC 9965 “Turin” dual socket server was used for showing the strong out-of-the-box performance on Intel’s Clear Linux even for this competing server processor.
Very strong Clear Linux performance on AMD hardware is no surprise as has been showcased on Phoronix over the years. Especially since Zen 4 with adding AVX-512 to AMD CPUs, the Clear Linux performance has been exceptional thanks to its AVX-512 and employing function multi-versioning (FMV) and other compiler optimizations for delivering a splendid out-of-the-box performance. Whether AMD Ryzen desktops/laptops or AMD EPYC server hardware, Intel’s Clear Linux was a great showcase of optimal out-of-the-box software potential on AMD platforms.
As a final run of Clear Linux on AMD, I ran some benchmarks of the final Clear Linux 43760 release compared to Ubuntu 25.04 for the very latest Ubuntu (non-LTS) state plus AlmaLinux 10.0 for the current performance on EPYC Turin for that Red Hat Enterprise Linux 10 derived community OS. The same AMD EPYC server platform was used for all of these fresh benchmarks: the AMD Volcano reference server, two AMD EPYC 9965 flagship Turin processors, 1.5TB of RAM, and 1TB Intel SSDPE2KX010T8 NVMe SSD.
With both AlmaLinux and Ubuntu Linux they were tested at their defaults plus an additional run with the “performance” governor for additional perspective, similar to last week’s Intel article. AlmaLinux 10 by default was using acpi-cpufreq with the schedutil governor while Ubuntu 25.04 was using amd-pstate-epp with the powersave governor. With the CPU frequency scaling governor adjustment being a common server change by administrators, I ran second runs of those operating systems with the “performance” governor to compare to Clear Linux’s default. That helps explain some of the performance differences observed.
It’s also worth noting that Clear Linux on the AMD EPYC 9965 actually runs at a disadvantage… Each AMD EPYC 9965 provides 192 cores / 384 threads for a combined 384 cores / 768 threads. But given the Intel focus and no current Intel offerings exceeding 512 cores/threads, Clear Linux maintains a 512 max NCPU kernel setting as part of its kernel build. Thus Clear Linux on this AMD EPYC server was running with 384 cores but only 512 threads enabled, which can impact some of the most scalable workloads tested but in any case as you’ll see Clear Linux still performs exceptionally well on this AMD server just as we’ve shown for years with Clear Linux on various AMD systems.
