Endurance of F-Stack Performance Advantage in Many-Process Environments

[victorstewart]

it's obviously common nowadays to have super high core count web server environments (say a dual Intel Xeon Gold with 28 cores, 56 threads).

but given that all F-Stack traffic must flow through a primary process, at what point does F-Stack's performance advantage over the host networking stack (say Clear Linux) begin to degrade?

and I assume that cost is never paid when going multi-thread vs multi-process?