As the Linux 6.19 merge window winded down this weekend, I began running this development kernel on more systems. While there are some scheduler regressions currently with Linux 6.19 Git, for HPC workloads especially I am seeing some encouraging results using a flagship AMD EPYC 9965 2P server configuration.
Over the weekend with an AMD EPYC 9655 2P server configuration for a combined 384 cores / 768 threads I ran some benchmarks of Linux 6.18 stable against Linux 6.19 Git as of 12 December. The same kernel configuration was used besides accepting the new default Kconfig options of Linux 6.19. The two AMD EPYC 9965 processors were running on an AMD Volcano reference server pla…
As the Linux 6.19 merge window winded down this weekend, I began running this development kernel on more systems. While there are some scheduler regressions currently with Linux 6.19 Git, for HPC workloads especially I am seeing some encouraging results using a flagship AMD EPYC 9965 2P server configuration.
Over the weekend with an AMD EPYC 9655 2P server configuration for a combined 384 cores / 768 threads I ran some benchmarks of Linux 6.18 stable against Linux 6.19 Git as of 12 December. The same kernel configuration was used besides accepting the new default Kconfig options of Linux 6.19. The two AMD EPYC 9965 processors were running on an AMD Volcano reference server platform.
All of the software and hardware was kept the same throughout testing besides swapping out the kernel version in use while the rest of the OS was comprised of the Ubuntu 25.10 defaults.
From there I ran a large variety of benchmarks for seeing the AMD EPYC 9965 2P server performance across a range of mostly real-world Linux server workloads. Tests on more hardware and different workloads relevant to different classes of systems will come over the weeks ahead. Linux 6.19-rc1 released today in working toward the stable Linux 6.19 debut around early February.