32bit slots and riser cards

Robert G. Brown rgb at phy.duke.edu
Fri Aug 29 10:17:09 EDT 2003


On Fri, 29 Aug 2003, Bogdan Costescu wrote:

> One other thing that turned me off was that in a system composed of only
> Intel components (SE7501WV2 mainboard and SR2300 case) the 64bit PCI slots
> on the mainboard allow inserting of the Myrinet card (but didn't try to
> see if it works), while the riser cards that came with the case do not,
> allowing only 3.3V ones - so the riser imposes additional limitations...

One last possible solution you can consider if you're using 2U cases and
don't mind ugly is that MANY of the cards you might want to add nowadays
are half-height cards on full height backplates.  Usually the backplate
is held on by two little screws.  The half height cards will snap into a
regular PCI slot normally (vertically) and still permit the case to
close with no riser at all.  The two negatives are there there are no
"half height riser backplates" that I know of, so the back of each
chassis will be open to the air, which may or may not screw around with
cooling airflow in negative ways, and the fact that you can't "screw the
cards down".  Both of these can be solved (or ignored) with a teeny bit
of effort, although you'll probably prefer to just get a riser that
meets your needs -- there are risers with a key that fits in the AGP
slot, risers with 32 bit keys, risers with 64 bit keys.. shop around.
Be aware that some of the risers you can buy don't work properly (why I
can't say, given that they appear to be little more than bus extenders
with keys to grab power and timing/address lines).

At a guess this won't help you with an old myrinet card as it is
probably full height, but if you get desperate and it's not, you could
likely make this work.

   rgb

> 
> -- 
> Bogdan Costescu
> 
> IWR - Interdisziplinaeres Zentrum fuer Wissenschaftliches Rechnen
> Universitaet Heidelberg, INF 368, D-69120 Heidelberg, GERMANY
> Telephone: +49 6221 54 8869, Telefax: +49 6221 54 8868
> E-mail: Bogdan.Costescu at IWR.Uni-Heidelberg.De
> 
> _______________________________________________
> Beowulf mailing list, Beowulf at beowulf.org
> To change your subscription (digest mode or unsubscribe) visit http://www.beowulf.org/mailman/listinfo/beowulf
> 

Robert G. Brown	                       http://www.phy.duke.edu/~rgb/
Duke University Dept. of Physics, Box 90305
Durham, N.C. 27708-0305
Phone: 1-919-660-2567  Fax: 919-660-2525     email:rgb at phy.duke.edu



_______________________________________________
Beowulf mailing list, Beowulf at beowulf.org
To change your subscription (digest mode or unsubscribe) visit http://www.beowulf.org/mailman/listinfo/beowulf



More information about the Beowulf mailing list