2 ms·
> although I'm not sure you'd have any control over the kernel threads You don't have control from the application, but if you're really trying to get the most
by toast0 3y ago
> although I'm not sure you'd have any control over the kernel threads
You don't have control from the application, but if you're really trying to get the most performance from your networking heavy application, you want to set up your server where the NIC rx and tx queues are cpu pinned, and the application thread that processes any given socket is pinned to the same cpu that the kernel is using for that socket. Then there's no cross cpu traffic for that socket (if all else goes well).
Hopefully all the memory used is NUMA local for the CPU and the NIC as well. That gets trickier with multiple socket systems, although some NICs do have provisions to connect to two PCIe slots so they can be NUMA local to two sockets.