Date: Thu, 15 Sep 2005 00:00:53 +0200 From: Oliver Lehmann <lehmann@ans-netz.de> To: joseph.koshy@gmail.com Cc: current@freebsd.org Subject: Re: low(er) disk performance with sched_4bsd then with sched_ule Message-ID: <20050915000053.448f251b.lehmann@ans-netz.de> In-Reply-To: <84dead72050914135239514c49@mail.gmail.com> References: <20050914194612.15692485.lehmann@ans-netz.de> <43286E37.40203@samsco.org> <20050914222013.178dc4dc.lehmann@ans-netz.de> <84dead72050914135239514c49@mail.gmail.com>
next in thread | previous in thread | raw e-mail | index | archive | help
Joseph Koshy wrote: > On -current you could use hwpmc and pmcstat to get a quick idea > of what is happening to the system: > > 1. # kldload hwpmc > You'll need a kernel compiled with options HWPMC_HOOKS. > > 2. # pmcstat -S instructions -c 1 -S instructions -O /tmp/ule.out dd if=...etc > > This command line instructs pmcstat to setup sampling > on both CPUs 0 & 1. > > 3. # pmcstat -R /tmp/ule.out -g > Convert the hwpmc(4) log to gprof gmon.out profiles. > > 4. # gprof -l /boot/kernel/kernel > ./k8-fr-retired-x86-instructions/kernel.gmon > gprof.ule > > Repeat for the 4BSD case and compare. > > Current has a few bug fixes that need to be merged to 6.0; > but you should get by in a pinch by copying over the content > of sys/dev/hwpmc/ and usr.sbin/pmcstat/ to a 6.x tree > and rebuilding. I did this now for bs=64k count=32000. The results are here: http://pofo.de/tmp/gprof.ule http://pofo.de/tmp/gprof.4bsd Hope this helps -- Oliver Lehmann http://www.pofo.de/ http://wishlist.ans-netz.de/
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20050915000053.448f251b.lehmann>