Re: Absolutely horrid IDE performance...

Mark Lord (
Wed, 02 Dec 1998 19:49:40 +0000

David wrote:

> try doing a heavy dump of one or multiple IDE disks and watch your system
> performance. repeat the task with scsi. scsi takes more housekeeping but
> is smarter. of course the term smarter is my opinion :)

Looks fine to me. Much faster than similar operations
on all SCSI-based unix boxes I work with.
But then, most of that performance is due to the IDE disks
being used under Linux rather than HP-UX, FreeBSD, Solaris, ..

Just to dispell any lingering myths, the biggest single
differences between regular SCSI (non differential, non fibre channel
and IDE drives are:

(1) IDE drives are significantly cheaper to purchase
(2) SCSI drives usually have faster spin rates,
eg. 10000rpm (SCSI) vs. 7200rpm (IDE)
or. 7200rpm (SCSI) vs. 5400rpm (IDE)

Point (2) accounts for nearly all of the performance difference
between SCSI and IDE systems. When people post, say, bonnie results
for a 7200rpm SCSI system, they turn out to be nearly exactly 33%
faster than an otherwise similar (half the cost) 5400rpm IDE system
(note that 7200rpm is 133% of 5400rpm).

Decent IDE performance is still possible, though.
Here is bonnie from my own 2-IDE-disk RAID0.
Most SCSI systems should be able to come close to this.

>Seeker 1...Seeker 2...Seeker 3...start 'em...done...done...done...
> -------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
> 256 17374 95.5 32681 33.2 10706 24.3 15683 68.4 23710 24.3 214.6 1.8


- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to Please read the FAQ at