cheap clusters

Robert G. Brown rgb at phy.duke.edu
Thu Apr 5 19:52:10 EDT 2001


On Thu, 5 Apr 2001, Mark Hahn wrote:

> > Oh, and one more point.  I still don't know what the memory limitations
> > of e.g. AMD's 760 chipset will be, but I do know that VIA's is 1.5 GB.
>
> the issue is how many rows/banks the chipset supports;
> size a function of this (and the prevailing dram density).
>
> VIA's KT266 spec states 6/8-row support for 3/4G (unregistered/registered),
> but that's assuming 512Mb density, which doesn't seem to be shipping.
> KT133a has a similar spec (6, no support for registered), but assumes
> 256Mb, and therefore a limit of 1.5GB.  note that KT266 also does ECC.
> AMD761 seems to support 4G with double-sided, registered, x4 256 Mb chips.

This is all very useful information, thanks.  I was using the very
pragmatic information of the motherboards available through my local
vendor (two or three manufacturers) plus what I could scrounge easily on
the web.  It looks like practically speaking (within my sample), even
makers of 266 FSB systems seem to be putting only three slots on the
motherboards.  For example, the Abit KT7a has VIA KT133a/VIA 686B and
claims to take 512's, but only will hold 1.5 GB.  Ditto for the Asus
A7V133.

Asus A7M266 uses AMD 761 and will hold 2 GB DDR but one or two retail
websites I visited indicated that it was wait in line available at best.
I'll try a google search on KT133a and see if I can find one that has
more than 3 slots.  Hmm, looks like gigabyte has a decent selection of
boards including all of the above.  I'll see if I can get some quotes
via my vendor.

Double thanks,

    rgb

-- 
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