Date: Sun, 30 Oct 2011 22:57:23 +1100 From: Adrian Chadd <adrian@freebsd.org> To: lev@freebsd.org Cc: arch@freebsd.org Subject: Re: [PATCH] fadvise(2) system call Message-ID: <CAJ-Vmo=MGkuaoq7=X8bCc-potxYsm22sCSqu3wMceaMNQAwOiw@mail.gmail.com> In-Reply-To: <1855260864.20111030155231@serebryakov.spb.ru> References: <201110281426.00013.jhb@freebsd.org> <4EAB550E.3060603@FreeBSD.org> <458756137.20111030024046@serebryakov.spb.ru> <CAJ-Vmo=D-b1SJ0WjJGGJ_WPK8FvE-fptDPRCGyMVhOm=FR8RYA@mail.gmail.com> <908954806.20111030130143@serebryakov.spb.ru> <CAJ-VmomdJnGqgRWH9zop=Opc=B6MxhCnL%2B1LSroUo3Np7LJNQg@mail.gmail.com> <1855260864.20111030155231@serebryakov.spb.ru>
next in thread | previous in thread | raw e-mail | index | archive | help
Right, that's why Squid/Lusca sucks much less there. The IO is done using a pool of threads. I bet in the torrent client case, massive speedups could be done with correct use of either IO threads, or the POSIX AIO library. I'm all for using fadvise() to hint to the OS about what's about to happen (and I'll start leveraging it myself once it hits the tree) but there are other ways to skin this cat.. Adrian
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?CAJ-Vmo=MGkuaoq7=X8bCc-potxYsm22sCSqu3wMceaMNQAwOiw>