Path: utzoo!censor!geac!torsqnt!news-server.csri.toronto.edu!mailrus!uunet!mcsun!ukc!edcastle!dcl-cs!aber-cs!odin!pcg From: pcg@cs.aber.ac.uk (Piercarlo Grandi) Newsgroups: comp.unix.i386 Subject: Re: Adaptec 154x Tuning for SCO Unix Message-ID: Date: 5 Jul 90 14:31:29 GMT References: <440@ticipa.ti.com> Sender: pcg@aber-cs.UUCP Organization: Coleg Prifysgol Cymru Lines: 18 In-reply-to: john@ticipa.ti.com's message of 28 Jun 90 13:16:12 GMT In article <440@ticipa.ti.com> john@ticipa.ti.com (John Maline) writes: My simple I/O speed test was: time dd if=/dev/root of=/dev/null bs=10240 results: time user sys asynchronous (sync enb jumper off) 7:32.0 0.9 2:17.1 synchronous 7:29.3 1.2 2:09.7 synchronous, DMA=5.7 MB/s 7:29.3 1.0 2:10.1 But reading from the block device entails such colossal kernel overheads for cache management, as duly documented in the sys time, that it is hard to gauge raw IO bandwidth. On the other hand, 130 megs in 450 seconds comes to 200-300 KB sec., which is like Larry Snyder's results. -- Piercarlo "Peter" Grandi | ARPA: pcg%cs.aber.ac.uk@nsfnet-relay.ac.uk Dept of CS, UCW Aberystwyth | UUCP: ...!mcsun!ukc!aber-cs!pcg Penglais, Aberystwyth SY23 3BZ, UK | INET: pcg@cs.aber.ac.uk