SMART drive in PC is useful. We need it. Why PC give us an option to disable it?

Lee wrote: : SMART drive in PC is useful and I believe everybody needs it. Why PC : give us an option to disable it? So some non-professional person may : happen to disable it and lose some benefits.

The usual - when it does not work (correctly) or f**ks up something else; Get rid of it.

BTW, of what interest is this in comp.arch.EMBEDDED?

--
  ******************************************************
  Never ever underestimate the power of human stupidity.
  -Robert Anson Heinlein

		GeirFRS@INVALID.and.so.forth
  ******************************************************

You're right regarding the first, but seriously wrong about the second statement. Not "everybody" actually "needs" smartdrv --- some applications will even stop to work correctly if smartdrv is used. Caching disrupts the sequence of events between what the software thinks it's writing to the harddisk, and what actually gets written.

[And what ever might have made you think this is an on-topic question for this newsgroup completely evades me.]

--
Hans-Bernhard Broeker (broeker@physik.rwth-aachen.de)
Even if all the snow were burnt, ashes would remain.

Brain in skull is useful and I believe everybody needs it. Why human consciousness give us an option to disable it? So some foolish person may happen to disable it and lose some benefits.

Not here; I turned my brain off long ago and haven't missed it a bit. No more guilt, fear, embarassment, etc.

On another serious note.

I read the original post a bit differently. I read the SMART drive as the application that monitors the health of the drive for predictive failure analysis. This was added a few years ago to the ATA drive interface spec.

One reason to disable it is for performance for some applications like video streaming where you cannot giveup time for the monitor software to communicate with the drive to get status. Otherwise I have used it quite a bit. The white paper on SMART describes in many cases a drive failure can be predicted 24-48 hours before a fail.

Regards, Nitro

Interesting, I didn't know there was a performance hit. My understanding was that the feature was implemented entirely in the drive's firmware, no host-side support required except the ability to send the "turn it on" command.

There are actually quite a few ways that SMART can be used, including "only report when asked but don't do anything extra" to "perform a long offline test that preempts my use of the drive for an hour or so".

When it's enabled, there probably is a minor effect on performance as, for example, when there is a sector-read CRC error there is extra activity by the SMART firmware to make note of the error in addition to the normal activity of the regular firmware to execute a re-seek of the sector. This aspect of SMART should be virtually invisible to the user.

However, there are options to enable various background self tests. These could steal a noticeable number of cycles.

If you have Linux (I rather suspect that Lewin does (I have his book) but for the info of others...), look at the "smartctl" command. It gives a lot more control and information than any Windows based utility I've seen (which doesn't mean, of course, that there's nothing out there).

For Windows users, check out the free utility DIskCheckup from

Handy to read the SMART info but it doesn't have the ability to launch self tests, etc.

--
Rich Webb   Norfolk, VA