[mythtv-users] Asus Pundit vs. Shuttle XPC SK41G

Will Dormann wd at pobox.com
Thu Jul 24 17:05:27 EDT 2003


At 11:22 AM 7/24/2003 -0700, you wrote:
>I just checked this with our main technical guy and he assures me that
>this is still the case.  

Well, then we'll just have to agree to disagree then.   This could
*possibly* be the case with some mis-behaved or ancient drives, and/or a
brain-dead EIDE controller, but as a general rule there should be no
problem with having the HD and CD-ROM device on the same controller.  Make
sure you've got a good 80-conductor cable, but also take a look at
jumpering.  I've heard of some devices "perferring" master/slave as opposed
to cable select and vice-versa.
Anyway, check this Seagate doc.  It clearly indicates that ATA/66 and
ATA/33 devices can share the same cable, each operating at their own max
speed.
http://www.seagate.com/support/kb/disc/ultra_ata_66.html#8

On my Pundit system, I have an ATA-33 optical drive and an ATA-100 HD on
the same controller.   hdparm -t shows that the *sustained* transfer rate
for the HD is over 40MB/sec.  I'm not sure of a way to test burst speed
offhand, but regardless, if the sustained rate is over 33MB/sec I'm sure
the drive is not being downgraded to ATA-33 speed due to the optical drive.

Maybe I should tack on a "YMMV" to my statement, but to flat-out say that
an ATA33 optical drive on the same cable will slow your HD down to ATA33
speeds is wrong.

-WD

I trust his opinion, since he often spends his
>days debugging hard drive issues, and mentioned that we deal with this
>particular issue in many of the servers that we build because the low
>profile CD drives we use are only ATA-33, and often need to share the
>channel with one of the server's main drives - forcing it down to the
>slower speed.
>
>The problem basically comes down to the fact that the master and slave
>drives need to communicate with each other (so they know which info is
>meant for which drive, presumably), as well as with the controller, so
>they all have to speak the same language.  Thus, if your optical drive
>doesn't support UDMA, or doesn't support DMA, the controller can't speak
>UDMA or DMA or whatever, and it reverts to a slower protocol. 
>
>Anyway, I'm sure there is technical documentation for this online
>somewhere, but as I said, I've seen this first-hand in my own pundit. 
>The hard drive (and thus the machine itself) runs noticeably slower with
>an ATA-33 cdrw plugged into it.
>
>-Chris
>
>_______________________________________________
>mythtv-users mailing list
>mythtv-users at snowman.net
>http://lists.snowman.net/cgi-bin/mailman/listinfo/mythtv-users
>
>
>


More information about the mythtv-users mailing list