Date: Sun, 2 Feb 2003 17:24:37 -0500 From: Bill Vermillion <bv@wjv.com> To: stable@FreeBSD.ORG Subject: Re: at33 vs others? Message-ID: <20030202222437.GA83950@wjv.com> In-Reply-To: <bulk.48016.20030202122052@hub.freebsd.org> References: <bulk.48016.20030202122052@hub.freebsd.org>
next in thread | previous in thread | raw e-mail | index | archive | help
> Date: Sun, 2 Feb 2003 11:10:59 +0100 (CET) > From: Alban Hertroys <dalroi@solfertje.student.utwente.nl> > Subject: Re: ata33 vs others? > > On 31 Jan, Eric Timme punched keys in this particular order: > > After wrestling with an 80gb hd and an Abit BX6-R2 for an afternoon and being > > pleasantly surprised that Abit's last bios release would allow my board to > > detect and use an 80gb hd I came to the sad realization that the computer > > only supported ata33. > > That shouldn't matter much if it is the only drive on the IDE channel. > > Modern harddrives still perform under 40MB/s last time I checked (the > IBM 60GXP series could do 37.5MB/s), and that is at maximum. > > The ATA100 standard was necessery because two disks on a channel could > in some cases press the required data throughput of the controller > over 66MB/s (2x37.5=75), though I'm sure it maxes out quite a bit under > 100MB/s. > > I don't know how much overhead data on an IDE channel has (like 'packet > headers' or some equivalent) and we now have an ATA133 standard that > would be kind of absurd with the above reasoning. > Maybe someone can shine a light on this? > I just updated an iNTEL 1100R rack mount from an ATA33 [don't recall the drive] to a Promise 100/133 controller and a Maxtor 120GB 7200 RPM ATA133 I saw significant increase. I used the old iozone2 [in ports] as I've used it for years and it measure r/w performance through the file system. The old system, using a 200MB fiel for test showed 9,166,619 B/sec write and 9,422,339B/sec read. With the new combo on a fully fresh filesystem I used a 500MB file for test. I got 21,408,22B/sec write and 46,091,252B/sec write. Since this was so much faster I repeated the test with a 1GB file. That gave me this: 20,677,512B/sec write and 45,554,310Bsec read. I just compiiled 4.7-P3 [CPU is 650MHz] and came in about 55 minutes and the kernel compile took just about 9 minutes. The faster drive and controller make a significant perfomance difference. Bill To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-stable" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20030202222437.GA83950>