Date: Sun, 7 Sep 1997 00:12:32 -0500 (EST) From: "John S. Dyson" <toor@dyson.iquest.net> To: michaelv@MindBender.serv.net (Michael L. VanLoon -- HeadCandy.com) Cc: sos@sos.freebsd.dk, mal@algonet.se, current@FreeBSD.ORG Subject: Re: lousy disk perf. under cpu load (was IDE vs SCSI) Message-ID: <199709070512.AAA00465@dyson.iquest.net> In-Reply-To: <199709070407.VAA04801@MindBender.serv.net> from "Michael L. VanLoon -- HeadCandy.com" at "Sep 6, 97 09:07:02 pm"
next in thread | previous in thread | raw e-mail | index | archive | help
Michael L. VanLoon -- HeadCandy.com said: > > Why is it necessary to bring this up over and over again. > Because I think that people sometimes think that the world changes from time to time. SCSI is really great in large high-end systems for sure. EIDE isn't the joke that IDE was 4-5yrs ago though. > > It's an Asus Triton-1 board (P55TP4N) with a Cyrix 6x86 P166+, 64MB > RAM, running NetBSD-1.2.1. "No load" means all the standard system > processes are running, along with a few X apps, but nothing using any > real CPU time. Loadit was a simple program I wrote (appended at the > bottom), which simply generated a constant load of one process, each. > Yep, I don't understand the fall off that others have seen. In fact, we have more complaints about I/O not being counted properly in system load and tasks doing I/O appear to have too high a priority. We have some mods that help the situation, but there are tradeoffs. > > I don't mean this in a condescending way, but I'd really like to see > this same kind of test run against four ccd-striped EIDE drives, > running in both PIO mode, and in DMA mode. Anyone have four drives > they could test it with? I only have a couple, and one is committed > elsewhere. > > Here are the results from the tests I ran: > > time dd if=/dev/rccd0f of=/dev/null bs=64k count=4096, no load: > > 268435456 bytes transferred in 35 secs (7669584 bytes/sec) > 0.026u 1.726s 0:35.27 4.9% 0+0k 6+1io 10pf+0w > > time dd if=/dev/rccd0f of=/dev/null bs=64k count=4096, 1 loadit: > > 268435456 bytes transferred in 35 secs (7669584 bytes/sec) > 0.021u 1.727s 0:35.46 4.9% 0+0k 0+1io 0pf+0w > > time dd if=/dev/rccd0f of=/dev/null bs=64k count=4096, 4 loadits: > > 268435456 bytes transferred in 35 secs (7669584 bytes/sec) > 0.021u 1.715s 0:35.17 4.9% 0+0k 0+1io 0pf+0w > > Here are my dd results for my EIDE 4GB Caviar drive -- NO STRIPING, running FreeBSD-current: I wonder how it would be to add a Promise EIDE controller, and run one drive per EIDE interface??? I do have one of those Promise controllers, and will probably add support in FreeBSD soon. Maybe I'll try ccd then :-). Sorry that I didn't have time for a scientific measurement, but I would be interested in running some packaged benchmarks. (BTW, the command overhead for my 4GB Caviar is about 80-100usecs also... Older Caviars get about 200-400usecs.) My 2GB Hawk with an NCR controller gets about 800usecs. IDE isn't a toy any more, but it also isn't for every application either. No load: dd if=/dev/rwd1 of=/dev/null bs=64k count=1600 1600+0 records in 1600+0 records out 104857600 bytes transferred in 10.848874 secs (9665298 bytes/sec) One loadit: ./loadit & dd if=/dev/rwd1 of=/dev/null bs=64k count=1600 1600+0 records in 1600+0 records out 104857600 bytes transferred in 10.864632 secs (9651279 bytes/sec) Two loadits: ./loadit & dd if=/dev/rwd1 of=/dev/null bs=64k count=1600 1600+0 records in 1600+0 records out 104857600 bytes transferred in 10.841087 secs (9672240 bytes/sec) Four loadits: ./loadit & ./loadit & dd if=/dev/rwd1 of=/dev/null bs=64k count=1600 1600+0 records in 1600+0 records out 104857600 bytes transferred in 10.835611 secs (9677129 bytes/sec)
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?199709070512.AAA00465>