Author: Robert Hyatt
Date: 07:25:36 07/12/02
Go up one level in this thread
On July 12, 2002 at 02:17:59, Dann Corbit wrote:
>On July 12, 2002 at 02:14:57, Jonas Cohonas wrote:
>
>>On July 11, 2002 at 23:49:02, Nolan Denson wrote:
>>
>>>On July 11, 2002 at 23:31:08, K. Burcham wrote:
>>>
>>>>
>>>>
>>>>http://www.angstrom.com/products/racer500_quad.htm
>>>
>>>I am not sure if that system would work like a true Quad motherboard ... like
>>>the one that SuperMicro has for the Intel processors. This system seems to be a
>>>cluster setup. I don't think many chess programs can take advantage of a
>>>cluster setup.
>>
>>It looks like it is a real quad:
>>
>>(snip)
>>a. 4 CPUs in 1U form factor
>>b. Can rack up to 168 Athlon MP 2100+ CPUs in a 42U rack
>>(snip)
>>
>>1 unit does not equal a cluster right? "Can rack up to 168 Athlon MP 2100+ CPUs
>>in a 42U rack" this sounds like the cluster of 42 quad MP's.
>>
>>Am i missing something?
>
>It's a bit ambiguous.
>
>It is very clear that the cluster is not SMP.
>
>It is possible that the 4 CPU machine has 4 independent CPU's loosely affiliated
>via message passing of some sort.
>
>Probably stuff that is sold for Beowulf clusters.
>
>The setups are pretty bizarre. A high end machine with a giant pile {320 GIG}
>of IDE disk!? SHUDDER!
It looks like their basic building block is a dual-cpu machine. Or, in this
case, two of those in a 1U chassis. Compact, but a dual-node cluster with
each node having two cpus, rather than a quad-processor node.
No idea why they would think that IDE is the way to go for a disk server,
however. Makes it worthless, IMHO.
This page took 0 seconds to execute
Last modified: Thu, 15 Apr 21 08:11:13 -0700
Current Computer Chess Club Forums at Talkchess. This site by Sean Mintz.