From owner-freebsd-current Mon Jan 11 05:16:57 1999 Return-Path: Received: (from majordom@localhost) by hub.freebsd.org (8.8.8/8.8.8) id FAA01316 for freebsd-current-outgoing; Mon, 11 Jan 1999 05:16:57 -0800 (PST) (envelope-from owner-freebsd-current@FreeBSD.ORG) Received: from st-lcremean.tidalwave.net (st-lcremean.tidalwave.net [208.213.203.186]) by hub.freebsd.org (8.8.8/8.8.8) with ESMTP id FAA01307 for ; Mon, 11 Jan 1999 05:16:54 -0800 (PST) (envelope-from lee@st-lcremean.tidalwave.net) Received: (from lee@localhost) by st-lcremean.tidalwave.net (8.9.1/8.8.8) id IAA00646; Mon, 11 Jan 1999 08:16:15 -0500 (EST) (envelope-from lee) Message-ID: <19990111081614.A624@tidalwave.net> Date: Mon, 11 Jan 1999 08:16:14 -0500 From: Lee Cremeans To: "Kevin G. Eliuk" Cc: current@FreeBSD.ORG Subject: Re: ide_pci.c Reply-To: lcremean@tidalwave.net References: <19990111002530.A1443@tidalwave.net> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii X-Mailer: Mutt 0.93.2i In-Reply-To: ; from Kevin G. Eliuk on Sun, Jan 10, 1999 at 11:49:33PM -0800 X-OS: FreeBSD 3.0-CURRENT X-Evil: microsoft.com Sender: owner-freebsd-current@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.ORG On Sun, Jan 10, 1999 at 11:49:33PM -0800, Kevin G. Eliuk wrote: > On Mon, 11 Jan 1999, Lee Cremeans almost wrote: > > > dd if=/dev/rwd0 of=/dev/null bs=1024k count=50 > > > > on a quiet (single-user) system. I haven't done any filesystem-throughput > > tests yet...this is the raw throughput through the controller. > > The patches allowed me to build the kernel, and I don't get the same > results at boot time. With the final results of your test show similar > to yours. > > Take a note of the "dmesg" output in my case and compare them to the raw > read/write. I'm not sure what to make of this really, ... other than > this motherboard and chipset really suk :-( > [snip] > wdc0 at 0x1f0-0x1f7 irq 14 on isa > wdc0: unit 0 (wd0): , 32-bit, multi-block-16 > wd0: 8010MB (16406208 sectors), 16276 cyls, 16 heads, 63 S/T, 512 B/S > wdc0: unit 1 (wd1): , DMA, 32-bit, multi-block-32 ^^^^^ This drive sucks almost as much as the board. :/ Maxtor 7000-series drives are notoriously slow, and the earliest (like the 7060 and 7120, made way back in 1991-1992) had reliability problems. Enable DMA on the DiamondMax 8GB, with "flags 0xa0ffa0ff". You'll like what you see. > wd0 > > -------Sequential Output-------- ---Sequential Input-- --Random-- > -Per Char- --Block--- -Rewrite-- -Per Char- --Block--- --Seeks--- > Machine MB K/sec %CPU K/sec %CPU K/sec %CPU K/sec %CPU K/sec %CPU /sec %CPU > CATHERNA 64 2698 77.0 7241 42.8 2730 19.4 2828 74.7 7980 31.2 95.2 4.4 > > wd1 > -------Sequential Output-------- ---Sequential Input-- --Random-- > -Per Char- --Block--- -Rewrite-- -Per Char- --Block--- --Seeks--- > Machine MB K/sec %CPU K/sec %CPU K/sec %CPU K/sec %CPU K/sec %CPU /sec %CPU > CATHERNA 64 3338 96.5 4120 21.8 1678 13.4 3584 95.7 4485 19.2 50.5 2.1 ^^^^ ^^^^ ^^^^ ^^^ Here's the biggest thing DMA buys you -- more CPU time. Notice that it takes more CPU to move single characters than to move huge blocks -- this is because large chunks of data are what DMA shines at. -- +--------------------------------------------------------------------+ | Lee Cremeans -- Manassas, VA, USA (WakkyMouse on DALnet and WTnet)| | lcremean@tidalwave.net| http://st-lcremean.tidalwave.net/~lee | To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-current" in the body of the message