I, too, read the article incorrectly the first time round. What it's saying is that for an EPYC CPU, performance under systemd is _better_. For a Xeon CPU, perf

I, too, read the article incorrectly the first time round. What it's saying is that for an EPYC CPU, performance under systemd is _better_. For a Xeon CPU, performance under systemd is _worse_. So, the EPYC's architectural differences won't explain why an EPYC CPU runs slower under systemd, because it's the opposite.