This is a big one. The only point to doing something like the thin wrapper is if we benchmark it and show that it's faster than fs2-tcp, at the very least on platforms which support epoll or io_uring. In the process, we can also figure out fun things like optimal EventLoopGroup tuning parameters and such.
This is a big one. The only point to doing something like the thin wrapper is if we benchmark it and show that it's faster than fs2-tcp, at the very least on platforms which support epoll or io_uring. In the process, we can also figure out fun things like optimal
EventLoopGroup
tuning parameters and such.