Impressive: "With the NGINX workload [Falkor/AMD] handled almost the same amount of requests as the Skylake server. [Falkor/AMD] managed to get 214 requests/watt vs the Skylake’s 99 requests/watt"
It's not entirely clear what the power consumption is actually measuring. If it's not whole-system draw then it could be highly misleading. What is considered "CPU power" and what doesn't can vary (like does it include the memory controller? Or PCI-E controller? etc...)