** This is the quasi-official and semi-temporary T13 email list server. **
Hi Ron,
Just allow them to do a taskfile io seek for a given range (capacity) and
divide by total time and call that ugly a benchmark.
That is the granularity one is allowed to have, being a host.
One does not need to know the story of the big-lie or majic about the
black box of voodoo....all one needs is to know that you get back what you
requested and the time it took.
It would be nice to know that there were possible barrier locks on
operations but that is not going to happen.
Andre Hedrick
Linux ATA Development
On Fri, 18 May 2001, Ron Stephens wrote:
> Date: Fri, 18 May 2001 11:55:43 -0700
> From: Ron Stephens <[EMAIL PROTECTED]>
> Reply-To: [EMAIL PROTECTED]
> To: [EMAIL PROTECTED]
> Subject: [temp t13] Re: ARRRGH: Drive performance
>
> ** This is the quasi-official and semi-temporary T13 email list server. **
>
> At 10:36 AM 5/18/2001, you wrote:
> >Do you know the logical LBA to physical LBA mapping for a drive?
> >Do you know the zone layout? Do you know if the drive does
> >small seeks faster than a head switch (most do!)? Do you
> >realize that with write caching on all write commands have
> >virtually zero "command overhead"? Do you realize that many
> >drives can execute 5, 10, maybe a 100 commands all in hardware
> >without the drive's firmware getting involved? Do you know that
> >many drives will recognize the seek and read/write patterns of
> >the most popular "performance measurement" software and adjust
> >their caching operations so they look better while running that
> >test? If you issue a bunch of seek commands without reading or
> >writing data do you know if the drive is really doing the seek
> >you think it should? Do you know if it is doing any seek at all?
> >It might not even be spinning while you think you are doing those
> >seek commands!
>
> HALE HALE HALE!!! People are NOT supposed to know our secrets!!
> If you tell everybody our secrets, then everybody will know our secrets!
> TSK TSK TSK.
>
> Now, we'll have to redo all our performance measurement detection and adjustment
>code!
>
> =ron
> (just kidding!)
>
> But, Hale is right: there is a whole lot of (1) hardware assist, (2) data location
> mangling due to defect management, (3) data caching, including write-back!,
> (4) cyl/head/zone/slip stuff going on depending on the TYPE of drive AND quality
> of drive, and (5) cache management adjustments based on past behaviour! [for example,
> if suddenly "writes" prevail, more and more cache may be assigned to writing...]
>
> I'd have to say that just looking from the outside interface, a lot of it would
>appear to
> be non-deterministic in a short term test.
>
>
>
> --
> If you have any questions or wish to unsubscribe send a
> message to Hale Landis, [EMAIL PROTECTED] To post to
> this list server send your message to [EMAIL PROTECTED]
>
> For questions concerning Thistle Grove Industries or TGI's
> list services please send email to [EMAIL PROTECTED]
>
>
>
--
If you have any questions or wish to unsubscribe send a
message to Hale Landis, [EMAIL PROTECTED] To post to
this list server send your message to [EMAIL PROTECTED]
For questions concerning Thistle Grove Industries or TGI's
list services please send email to [EMAIL PROTECTED]
