Date: Fri, 14 Oct 2005 18:33:41 -0400 From: Garrett Wollman <wollman@csail.mit.edu> To: Bruce Evans <bde@zeta.org.au> Cc: current@freebsd.org Subject: Re: Call for performance evaluation: net.isr.direct (fwd) Message-ID: <17232.12869.659397.686840@khavrinen.csail.mit.edu> In-Reply-To: <20051015074316.T1260@epsplex.bde.org> References: <12907.1129286370@critter.freebsd.dk> <20051015074316.T1260@epsplex.bde.org>
next in thread | previous in thread | raw e-mail | index | archive | help
[Moved to -current as this is no longer net related.] <<On Sat, 15 Oct 2005 08:28:04 +1000 (EST), Bruce Evans <bde@zeta.org.au> said: > values of HZ like 1000 -- with hz > stathz it is easy to use a periodic > itimer to arrange to run about (hz - stathz) / hz of the time without ever > seeing a statclock tick. The whole point of statclock was that it was supposed to be an exponential distribution centered on stathz, not strictly periodic, precisely to prevent this problem. With the new APIC clock divider code, we should be able to do this fairly cheaply and at a higher rate than 128 Hz. It was originally developed on SPARC hardware so that architecture ought to be able to do it easily. -GAWollman
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?17232.12869.659397.686840>