Re:Why HN was slow and how Rtm fixed it
by kujawa on Wednesday January 19, 2011 @02:37PM
(#2120941)
Read C10K? Both select() and poll() have this problem internally. You have to use one of the more advanced techniques available if you really want to scale. epoll(), kqueue() or friends.
[ Parent
| on HN ]
Based on RTM's description, it sounds like even select() would be a huge improvement. HN needs to solve the C100 problem before worrying about C10K.
Rtm's phrasing implies that MzScheme is making a syscall for each socket in turn, so it sounds like it isn't even using the basic select() or poll().
Re:Why HN was slow and how Rtm fixed it
by mahmoudimus on Wednesday January 19, 2011 @05:31PM
(#2121535)
I thought the same exact thing when I read rtm's description. I thought to myself select()/poll() is much more efficient, and it sounds like that's not what's implemented underneath.
poll() is slightly better than select(), because you only have to iterate over the file descriptors that were passed, rather than from 0 to nfds.