interesting iocp benchmark

Marc Lehmann schmorp at
Sun Jan 30 18:04:29 CET 2011

Just saw this (libevent2 performance linux vs. windows).

Linux (virtual machine) : 1220 Mb/s (cf. 1.45 Gb/s using select directly)
Windows (select)        :   96 Mb/s (cf.  100 Mb/s using select directly)
Windows (IOCP)          :  125 Mb/s

Libevent2 can use iocps because if offers a special file handle
abstraction which can take advantage of iocps.

However, barring some phenomenal shortcoming/bug in libevent2, it seems
that windows throughput doesn't improve that much with iocps.

Of course, this benchmark *probably* uses few fd's, so it doesn't actually
test scalability of select vs. iocps, but then, it seems it's entirely not
worth it to support windows, at leatsf or high-throughput applications.

Just looked interesting...

                The choice of a       Deliantra, the free code+content MORPG
      -----==-     _GNU_    
      ----==-- _       generation
      ---==---(_)__  __ ____  __      Marc Lehmann
      --==---/ / _ \/ // /\ \/ /      schmorp at
      -=====/_/_//_/\_,_/ /_/\_\

More information about the libev mailing list