Path: utzoo!attcan!utgpu!jarvis.csri.toronto.edu!mailrus!cs.utexas.edu!usc!ucsd!ucsdhub!hp-sdd!ncr-sd!ncrcae!hubcap!Larry
From: lkaplan@BBN.COM (Larry Kaplan)
Newsgroups: comp.parallel
Subject: Re: scalability of n-cubes, meshes
Keywords: Butterfly TC2000
Message-ID: <7280@hubcap.clemson.edu>
Date: 1 Dec 89 16:49:25 GMT
Sender: fpst@hubcap.clemson.edu
Lines: 18
Approved: parallel@hubcap.clemson.edu

In article <7252@hubcap.clemson.edu> Donald.Lindsay@MATHOM.GANDALF.CS.CMU.EDU writes:
>
>...On an NCUBE-2,
>the header is 32 bits, sent serially: the latency is 32+ bit times
>(44, to be precise). A byte-wide path, such as BBN uses, could make
>the same decision in 4+ clocks. (And both could be faster if the
>header's likeliest command mode was decodable from, say, the first
>half of the header.) 

Actually, the current generation BBN Butterfly Switch does better than this.
By "eating" the addressing bits used by a particular switch element, the
relevant addressing bits for the next switch element arrive there in the FIRST
byte.  So routing decisions are made after 1 clock.
_______________________________________________________________________________
				 ____ \ / ____
Laurence S. Kaplan		|    \ 0 /    |		BBN Advanced Computers
lkaplan@bbn.com			 \____|||____/		10 Fawcett St.
(617) 873-2431			  /__/ | \__\		Cambridge, MA  02238


Brought to you by Super Global Mega Corp .com