maddie@pnet01.cts.com (Tom Schenck) (08/14/89)
I was looking through a few of my back issues of Scietific American, and I noticed an article on the Connection Machine... I was wondering if anyone on the net has heard anything of this machine? It was apparently designed and built in the mid-1980's, and I haven't heard much else on it. ------------------------------------------------------------------------------- UUCP: {cbosgd, hplabs!hp-sdd, sdcsvax, nosc}!crash!maddie | Disclaimer : ARPA: crash!pnet01!maddie@nosc.mil | The only company INET: maddie@pnet01.CTS.COM | who thinks like me, [ Some day I'll come up with a nifty .signature... ] | is owned by me. -------------------------------------------------------------------------------
adam@mit-amt.MEDIA.MIT.EDU (Adam Glass) (08/14/89)
maddie@pnet01.cts.com (Tom Schenck) writes: > I was looking through a few of my back issues of Scietific American, and I > noticed an article on the Connection Machine... I was wondering if anyone on > the net has heard anything of this machine? It was apparently designed and > built in the mid-1980's, and I haven't heard much else on it. So far as I know, the Connection Machine is a massively parallel machine, having 16384 (or thereabouts) processors (I think it would be deemed "fine grain" by most people). It is made by Thinking Machines (Corporation?) here in Cambridge, MA. Again, this is just a guess, but I believe that we here at the Media Lab own the only one, though this doesn't sound likely. Anything beyond that and you're going to have to ask someone who knows more about the beastie. Adam -- "Offer me anything I ask for..." | email: adam@media-lab.media.mit.edu | "Anything you want." +---------------------------+---------+ "I want my father back, you son of a bitch." - The Princess Bride | Sigh... |
bv3456@leah.albany.edu (Victor @ The Concrete Museum) (08/15/89)
In <476@mit-amt.MEDIA.MIT.EDU>, adam@mit-amt.MEDIA.MIT.EDU (Adam Glass) writes, >maddie@pnet01.cts.com (Tom Schenck) writes: >> I was looking through a few of my back issues of Scietific American, and I >> noticed an article on the Connection Machine... I was wondering if anyone on >> the net has heard anything of this machine? It was apparently designed and >> built in the mid-1980's, and I haven't heard much else on it. > >So far as I know, the Connection Machine is a massively parallel machine, >having 16384 (or thereabouts) processors (I think it would be deemed "fine >grain" by most people). It is made by Thinking Machines (Corporation?) here >in Cambridge, MA. Again, this is just a guess, but I believe that we here >at the Media Lab own the only one, though this doesn't sound likely. Off the top of my head: I believe the CM was designed by Dan Hillis, who is the author of a book entitled, "The Connection Machine", published by the MIT Press. I think the CM can acutally have up to 65536 processors (!), and must be connected to a front-end, usually a VAX or a Sun. There is a C compiler for it (C*), which has some extensions (as well as limitations). - victor
cfreese@super.ORG (Craig F. Reese) (08/16/89)
In article <1957@leah.Albany.Edu> bv3456@leah.albany.edu (Victor @ The Concrete Museum) writes: > > >In <476@mit-amt.MEDIA.MIT.EDU>, adam@mit-amt.MEDIA.MIT.EDU (Adam Glass) writes, >>maddie@pnet01.cts.com (Tom Schenck) writes: >>> I was looking through a few of my back issues of Scietific American, and I >>> noticed an article on the Connection Machine... I was wondering if anyone on >>> the net has heard anything of this machine? It was apparently designed and >>> built in the mid-1980's, and I haven't heard much else on it. >> >>So far as I know, the Connection Machine is a massively parallel machine, >>having 16384 (or thereabouts) processors (I think it would be deemed "fine >>grain" by most people). It is made by Thinking Machines (Corporation?) here >>in Cambridge, MA. Again, this is just a guess, but I believe that we here >>at the Media Lab own the only one, though this doesn't sound likely. > >Off the top of my head: I believe the CM was designed by Dan Hillis, >who is the author of a book entitled, "The Connection Machine", published >by the MIT Press. I think the CM can acutally have up to 65536 processors (!), >and must be connected to a front-end, usually a VAX or a Sun. There is a C >compiler for it (C*), which has some extensions (as well as limitations). > > - victor The CM is a commercial machine that is manufactured & sold by Thinking Machines Inc. of Cambridge Mass. It comes in various forms: 2K, 4K, 8K, 16K, 32K and 64K fine grained (single bit) processors. As of the last I heard there were about 35 installed at various government, commercial, and achademic institutions. The CM-2 (the current model) has a number of enhanced features of the CM-1 (the one you probably read about). Some of these include. The DataVault mass storage system. The FrameBuffer high speed/resolution graphics interface. There is also a floating point option that you can get (for big bucks!) which crams upto 2000 Weitek chips into the 64K box. In some cases this can really fly. Within the last year or so there was a pretty good article in IEEE computer describing the CM-2. There was also an article in Comm. ACM a while back that talked about some programming applications. craig ----------------- Craig F. Reese Email: cfreese@super.org Institute for Defense Analyses/ Supercomputing Research Center 17100 Science Dr. Bowie, MD 20715-4300
jdm@h.cs.wvu.wvnet.edu (James D Mooney,205K,7,2913548) (08/17/89)
From article <197@crash.cts.com>, by maddie@pnet01.cts.com (Tom Schenck): > > I was looking through a few of my back issues of Scietific American, and I > noticed an article on the Connection Machine... I was wondering if anyone on > the net has heard anything of this machine? It was apparently designed and > built in the mid-1980's, and I haven't heard much else on it. The Connection Machine has been a commercial product for several years. There are at least two current models, CM-1 and CM-2, each with potentially as many as 64K real processors. For a fairly recent description see the article by Tucker and Robertson in IEEE Computer Magazine, August 1988. Recently an offer was posted on comp.parallel to provide internet access to a CM-2 at Thinking Machines Corporation. Accounts were being offered on request; however, I requested one and was placed on a waiting list. For information contact David Ray (ray@think.com). Jim Mooney Dept. of Stat. & Computer Science (304) 293-3607 West Virginia University Morgantown, WV 26506 INTERNET: jdm@a.cs.wvu.wvnet.edu