Date: Sun, 19 Sep 1999 11:48:06 -0700 (PDT) From: "Rodney W. Grimes" <freebsd@gndrsh.dnsmgr.net> To: dillon@apollo.backplane.com (Matthew Dillon) Cc: phk@critter.freebsd.dk (Poul-Henning Kamp), dg@root.com, grog@lemis.com (Greg Lehey), cvs-committers@FreeBSD.org, cvs-all@FreeBSD.org Subject: Re: User block device access (was: cvs commit: src/sys/miscfs/specfs spec_vnops.c src/sys/sys vnode.h src/sys/kern vfs_subr.c) Message-ID: <199909191848.LAA55782@gndrsh.dnsmgr.net> In-Reply-To: <199909191837.LAA73569@apollo.backplane.com> from Matthew Dillon at "Sep 19, 1999 11:37:23 am"
next in thread | previous in thread | raw e-mail | index | archive | help
... > : > :Yes indeed there are: lack of funds for much needed HW upgrades. > :This is actually a 1.8G Quantum Prodrive I inherited when a former > :instance of freefall had a disk upgrade in early 1995 and currently > :the only scsi disk I have to test with. > : > :-- > :Poul-Henning Kamp FreeBSD coreteam member > > What kind of cpu is this? If the difference is this great, then > the memory-copy overhead is what is killing your test due to a slow > cpu. Since you are not using the buffered block device the way it > is supposed to be used (i.e. your test is resulting in a 100% cache > miss ratio), your results are a foregone conclusion. They aren't > really proving anything other then the fact that the buffered block > device is caching the data (incuring an extra copy) while the raw > device is not. This does not make the raw device somehow magically > better. There are benchmarks, damn benchmarks, and lies. You pick, but what I have seen of this thread doesn't have a ``b'' in it :-). -- Rod Grimes - KD7CAX - (RWG25) rgrimes@gndrsh.dnsmgr.net To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe cvs-all" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?199909191848.LAA55782>