[Beowulf] torus versus (fat) tree topologies
Many of your questions may have already been answered in earlier discussions or in the FAQ. The search results page will indicate current discussions as well as past list serves, articles, and papers.
Underwood, Keith D kdunder at sandia.govFri Nov 12 19:45:27 PST 2004
- Previous message: [Beowulf] CCGrid2005: Call for Tutorial Proposals
- Next message: [Beowulf] CFP: MDC'05
- Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]
Yep, that's the standard solution. Red Storm has the Z dimension in the cabinet (and, so, is a torus) and we choose not to connect X or Y in a torus (for a couple of reasons), but it could hypothetically be done using this type of scheme. (i.e. you use the same type of interleaving, but you do it at the cabinet/row level). Keith > BlueGene's trick is to attach every other node to the end, and then > come back with the unused ones. So: > > Nodes: 0 1 2 3 4 > > Connections: 0 -> 2 -> 4 -> 3 -> 1 -> 0 > > No long cable. This is probably a classic solution not invented by IBM. > > -- greg > > _______________________________________________ > Beowulf mailing list, Beowulf at beowulf.org > To change your subscription (digest mode or unsubscribe) visit http://www.beowulf.org/mailman/listinfo/beowulf >
- Previous message: [Beowulf] CCGrid2005: Call for Tutorial Proposals
- Next message: [Beowulf] CFP: MDC'05
- Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]
More information about the Beowulf mailing list
