[MUD-Dev] Sockets

[Mark Gritter]

Mark Gritter writes:
> Leaving aside the implementation-specific question of what the fixed
> overhead is... polling overhead would seem to scale with the number of
> _possible_ I/O events, while threading overhead increases with the
> number of _actual_ I/O events.

Duh.  Usually the two are proportional.  :)  I shouldn't try thinking
hard late at night.

How about this: as a function of "window size", threads look better at 
small polling times and select() looks better at large polling times--- 
when it can report several events at once.

Mark Gritter
mark at erdos.stanford.edu


_______________________________________________
MUD-Dev maillist  -  MUD-Dev at kanga.nu
http://www.kanga.nu/lists/listinfo/mud-dev