News for hackers. Stuff that matters.
Re:Let's Do Some Science: Zed's Test Stack for poll vs. epoll by lhorie on Tuesday August 03, 2010 @11:26PM (#1573561)
I'd be interested to see if the idea of switching between epoll and poll at 60% ATR would have too much overhead as to be impractical.

Re: > 60% active fd's being real-life or not, I don't really see what's the problem in wanting to handle that case in a project whose goal is basically to keep humming along nicely regardless of what you throw at it.

Re:Let's Do Some Science: Zed's Test Stack for poll vs. epoll by pphaneuf on Tuesday August 03, 2010 @11:54PM (#1573601)
Adding/removing from the poll() array is more or less free, but with epoll, you have to use epoll_ctl(), so you'd definitely want an hysteresis to avoid bouncing things around too much.
Re:Let's Do Some Science: Zed's Test Stack for poll vs. epoll by zedshaw on Wednesday August 04, 2010 @01:12AM (#1573709)
Interesting, you're like the 3rd or 4th person who's mentioned hysteresis.
Re:Let's Do Some Science: Zed's Test Stack for poll vs. epoll by rbanffy on Wednesday August 04, 2010 @09:37AM (#1574444)
It's a fairly common idea.
Re:Let's Do Some Science: Zed's Test Stack for poll vs. epoll by pphaneuf on Wednesday August 04, 2010 @10:36PM (#1576973)
As in "only the 3rd or 4th" (so I'm possibly clever), or as in "a bunch of people"? I don't know your sample size. :-)