In this quick follow-up video, I revisit my MikroTik CRS504-4XQ-IN 100GbE switch testing because I realized my original iperf3 approach wasn’t the cleanest way to show what this switch can do. In the review video I used iperf3 and had to run multiple parallel processes since iperf3 isn’t multi-threaded. What I didn’t know at the time is that iperf2 actually is multi-threaded (thanks to a viewer comment that mysteriously disappeared), so I wanted to redo the tests the right way and share the results. My setup is simple: a Dell R730 as the client on the left, and a Supermicro box as the server on the right, both on 100GbE. I also backed out all the kernel/sysctl tuning I’d been experimenting with—no special kernel parameters, no jumbo frame games—just default settings. With a basic run I saw about 31.2 Gbps, but with 5 parallel threads (-P 5) I hit 94.1 Gbps, and the CPU graphs looked healthy. On a longer 60-second run, the switch interface graph sat right around 100 Gbps, which is exactly what I wanted to see. Bottom line: this CRS504 is fast, and you don’t need a bunch of OS tuning to get real 100GbE performance in a multi-stream, server-like workload.