Subject: Re: 1.4.2 Observations
To: Laine Stump <lainestump@rcn.com>
From: Dave Olson <olson@bengaltech.com>
List: tech-kern
Date: 03/30/2000 20:38:00
Laine Stump wrote: 
|  I thought the deal with thermal recalibration and AV drives was that AV
|  drives are designed to be able to do the recalibration without pausing,

Not without pausing, they just do a better job of doing recal on the
fly during other seeks.  If you keep doing long linear reads, eventually
most AV drives will still have to stop and recal, but in that case they
break it up into multiple pieces.

They also try to be intelligent, and if power consumption is more or
less steady, they can postpone the recal until seek errors start occurring
above some threshold, which might well be never if the drive is in a
thermally stable environment.

These days most of the higher end drives use the same algorithms, they
just tend to be a bit more biased towards performance, rather than trying
to keep performance "the same all the time".  The higher end IBM drives
have been doing this for at least 3 years now, and under reasonable
operating conditions could postpone the recal for up to 2 hours waiting
for an "idle period", even back then.

ide drives have to do recal just the same as scsi drives; it's a fundamental
issue of the drive in some parts of the power/thermal/performance curve.


Dave Olson
Owned by 6 cats, owner of none...
Personal:  olson@bengaltech.com        Work:  olson@geocast.com
           http://www.bengaltech.com          http://www.geocast.com