Date: Thu, 15 Jul 1999 12:21:04 -0700 From: "David Schwartz" <davids@webmaster.com> To: "Tani Hosokawa" <unknown@riverstyx.net> Cc: <lh@aus.org>, <chat@FreeBSD.ORG> Subject: RE: Known MMAP() race conditions ... ? Message-ID: <000201becef7$2e790680$021d85d1@youwant.to> In-Reply-To: <Pine.LNX.4.10.9907151211190.10288-100000@avarice.riverstyx.net>
next in thread | previous in thread | raw e-mail | index | archive | help
> > If you're going to make a threaded program, you should do > it right. A > > half-assed threaded program may even be worse than a well-designed > > one-process-per-request program. > Again, off in theory. Fact is, the threaded server is faster than the > per-process model. And, it's not just going to go away. Really? If that's true (at least on some set of platforms) then take a small step forward. At least it will decrease Apache's memory consumption. I'm surprised it's actually faster. I honestly can't think of any reason it would be. You avoid the overhead of having to fork, but Apache pre-forks and reuses its children so this should be negligible. The kernel has fewer processes to schedule, but the overhead of having that same number of threads to schedule should be roughly equivalent. Perhaps it's because some of the context switches that were previously between processes are now between threads of the same process. But that shouldn't be a major difference anyway. And I would think that would be swamped by thread contention for things like memory allocation. It will probably make Apache run worse on FreeBSD though. If it pushes FreeBSD to a better threads implementation, I'm all for it. DS To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-chat" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?000201becef7$2e790680$021d85d1>