allan@cpsc.UCalgary.CA (Jeff Allan) (02/20/89)
In article <4156@hubcap.UUCP> I wrote: >I am interested in identifying BBN Butterfly sites and establishing a forum >for discussing experiences, problems, etc. >Anyone interested in participating in this exchange of information can either >post their particulars to this newsgroup or email me directly. >I will post a follow-up summary to this group if the response warrants. >If anyone would prefer not to their particulars posted to the net should >indicate so in their response. > >I'll get things started by telling you about our site. >The University of Calgary has a Butterfly Plus with 12 nodes. >We use it for research in parallel processing and simulation. >We've had for about a year now. I received about a dozen replies representing about eight sites. These are summarized below, followed by a long listing of the original replies. %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% Summary %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% Brian J Palmer Bolt Beranek and Newman Inc., Cambridge MA bpalmer%BBN.COM%UcEdu.UNCA.AdhocNet.CA@Ucnet.UNCA.AdhocNet.CA Scott Whitman (slim@cis.ohio-state.edu) Ohio State Computer Science Department Dick Peskin Rutgers University peskin%caip.rutgers.edu%UcEdu.UNCA.AdhocNet.CA@Ucnet.UNCA.AdhocNet.CA (R. L. Peskin) Neil B. Cohen BBN-ACI, Washington D.C. "Neil B. Cohen" <uunet!BBN.COM!nbc> Randy Michelsen Los Alamos National Laboratory rem%manihi%LANL.GOV%UcEdu.UNCA.AdhocNet.CA@Ucnet.UNCA.AdhocNet.CA (Randy E. Michelsen) David Kotz Department of Computer Science, Duke University, Durham, NC 27706 decvax!duke!dfk Karsten Georgia Tech. schwan@gatech.edu Michael L. Scott University of Rochester (716) 275-7745 scott@cs.rochester.edu scott%rochester@CSNET-RELAY {decvax, allegra, cmcl2}!rochester!scott Floyd Wofford Ballistic Research Laboratory floyd@smoke.brl.mil Yiannis Ohio State University yiannis%cis.ohio-state.edu%UcEdu.UNCA.AdhocNet.CA@Ucnet.UNCA.AdhocNet.CA David BBN Advanced Computers, Inc., Cambridge MA mailrus!yeager.bbn.com!djoslin@gatech.uucp jr BBN Systems and Technologies Corporation, Cambridge MA jr@bbn.com or bbn!jr %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% Full responces follow %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% >From bpalmer%BBN.COM%UcEdu.UNCA.AdhocNet.CA@Ucnet.UNCA.AdhocNet.CA Fri Jan 20 11:52:10 1989 To: allan@cpsc.ucalgary.ca Subject: Re: BBN Butterfly Users Newsgroups: comp.parallel In-Reply-To: <4156@hubcap.UUCP> Organization: Bolt Beranek and Newman Inc., Cambridge MA Date: Fri, 20 Jan 89 11:17:02 -0500 From: bpalmer%BBN.COM%UcEdu.UNCA.AdhocNet.CA@Ucnet.UNCA.AdhocNet.CA Status: RO Hi, I would also like to see a BBN Butterfly forum get started. I work in BBN's Newport,RI office where we've been working on a large simulation for the last 3 years. At present, we're using a 28 node Plus with 16 BVME's. Brian J Palmer >From slim%cis.ohio-state.edu%UcEdu.UNCA.AdhocNet.CA@Ucnet.UNCA.AdhocNet.CA Fri Jan 20 13:33:41 1989 Date: Fri, 20 Jan 89 13:23:33 EST From: slim%cis.ohio-state.edu%UcEdu.UNCA.AdhocNet.CA@Ucnet.UNCA.AdhocNet.CA Message-Id: <8901201823.AA20131@tut.cis.ohio-state.edu> To: allan@cpsc.ucalgary.ca Subject: Re: BBN Butterfly Users Newsgroups: comp.parallel In-Reply-To: <4156@hubcap.UUCP> Organization: Ohio State Computer Science Department Cc: Status: RO Just thought you would want to include us on your list. I'm at the Ohio State University Computer Science Department and we have a Butterfly GP1000 Running Mach. We are doing research in operating systems, graphics, and AI on the machine (it has 10 nodes). You might talk with BBN for other sites but some others include: Duke, Rochester, and University of Maryland to name a few. Scott Whitman (slim@cis.ohio-state.edu) -- Scott Whitman, Graduate Research Assistant in Computer Graphics slim@cis.ohio-state.edu or ...!cbosgd!osu-cis!slim The Ohio State University; 2036 Neil Ave. Columbus OH USA 43210-1277 >From peskin%caip.rutgers.edu%UcEdu.UNCA.AdhocNet.CA@Ucnet.UNCA.AdhocNet.CA Fri Jan 20 16:19:29 1989 Date: Fri, 20 Jan 89 18:13:09 EST From: peskin%caip.rutgers.edu%UcEdu.UNCA.AdhocNet.CA@Ucnet.UNCA.AdhocNet.CA (R. L. Peskin) Message-Id: <8901202313.AA23429@caip.rutgers.edu> To: allan@cpsc.ucalgary.ca Subject: Re: BBN Butterfly Users Keywords: Butterfly Sites, Users' group In-Reply-To: USENET article <4156@hubcap.UUCP> Status: RO Rutgers Univ has a 32 node GP1000. Used for parallel computing research. --dick peskin <peskin@caip.rutgers.edu> >From alberta!mnetor!uunet!BBN.COM!nbc Sun Jan 22 10:30:14 1989 Date: Sat, 21 Jan 89 20:18:31 EST From: "Neil B. Cohen" <uunet!BBN.COM!nbc> To: alberta!calgary!allan@seismo.CSS.GOV Subject: Butterfly Users group Status: RO Jeff, I work for BBN-ACI on the Butterfly in the Washington D.C. area. The people at the Naval Research Lab (NRL) tried to start a local users group this past summer. They had a few meetings, with some interesting speakers - one of the speakers was Darrin West from U. Calgary. You might be interested in contacting Steve Matney at NRL about your user's group forum. His email address used to be matney@nrl-css.arpa. If you can't get him, try Jeff Handy (handy@nrl-css.arpa or handy@itd.nrl.navy.mil). As a BBN employee, I tried to help Steve get a user's group going, but at the same time, I believe that such a group should be run by and for the customers - BBN will help where possible, but we want to stay out of the way too. If there is anything I can do for you - please let me know. Neil B. Cohen (nbc) PHONE: 703-848-4877 UUCP: ...{ihnp4, seismo}!bbn.com!nbc DOMAIN: nbc@bbn.com Murphy was an optimist! >From rem%manihi%LANL.GOV%UcEdu.UNCA.AdhocNet.CA@Ucnet.UNCA.AdhocNet.CA Mon Jan 23 08:33:37 1989 Date: Mon, 23 Jan 89 08:21:17 MST From: rem%manihi%LANL.GOV%UcEdu.UNCA.AdhocNet.CA@Ucnet.UNCA.AdhocNet.CA (Randy E. Michelsen) Message-Id: <8901231521.AA00595@manihi.lanl.gov> To: allan@cpsc.ucalgary.ca Subject: Butterfly Status: RO Jeff -- We recently acquired a 6 node GP1000 for (parallel) simulation research. Are you associated with the TimeWarp activities (we have some interest in that paradigm here)? Thanks, Randy Michelsen Los Alamos National Laboratory 505-667-0789 >From dfk%cs.duke.edu%UcEdu.UNCA.AdhocNet.CA@Ucnet.UNCA.AdhocNet.CA Sun Jan 22 19:07:14 1989 Date: Sun, 22 Jan 89 20:48:05 EST From: dfk%cs.duke.edu%UcEdu.UNCA.AdhocNet.CA@Ucnet.UNCA.AdhocNet.CA Message-Id: <8901230148.AA20344@romeo.cs.duke.edu> To: allan@cpsc.UCalgary.CA Subject: Re: BBN Butterfly Users In-Reply-To: your article <4156@hubcap.UUCP> News-Path: duke!mcnc!rutgers!mailrus!uflorida!gatech!hubcap!allan Status: RO > I am interested in identifying BBN Butterfly sites and establishing a forum > for discussing experiences, problems, etc. BBN used to support a mailing-list for Butterfly users. Check with them to see if they still have such a thing. I don't think so. We have a 32-processor GP1000, which just had the first real release of Mach booted this weekend. We have beta-tested Mach for several months. In early 1987 we got a Butterfly 1 which was later upgraded to a 32-processor Butterfly Plus running Chrysalis 4.0. However, this is not used much and will soon merge with the above to make a 64-node GP1000, whenever 64-node Mach is available. We use the two for parallel Systems research and for VLSI simulation research, and for graduate-level OS classes. We hope it will also be used for scientific computing applications and research, but that is starting slowly. Include me, but I can't speak for any other users here. David Kotz Department of Computer Science, Duke University, Durham, NC 27706 ARPA: dfk@cs.duke.edu CSNET: dfk@duke UUCP: decvax!duke!dfk >From schwan%wayward%gatech.edu%UcEdu.UNCA.AdhocNet.CA@Ucnet.UNCA.AdhocNet.CA Mon Jan 23 11:54:40 1989 Date: Mon, 23 Jan 89 13:35:13 EST From: schwan%wayward%gatech.edu%UcEdu.UNCA.AdhocNet.CA@Ucnet.UNCA.AdhocNet.CA (Karsten Schwan) Message-Id: <8901231835.AA01575@wayward.gatech.edu> To: allan@cpsc.UCalgary.CA Subject: Butterfly Status: R We are now getting a 32-node BBN Butterfly here at Georgia Tech. I had a machine at my previous place of employment (Ohio State), as well. In fact, there is a 12 node machine there still. Karsten schwan@gatech.edu >From scott%cs.rochester.edu%UcEdu.UNCA.AdhocNet.CA@Ucnet.UNCA.AdhocNet.CA Mon Jan 23 11:55:25 1989 Date: Mon, 23 Jan 89 09:39:54 EST From: scott%cs.rochester.edu%UcEdu.UNCA.AdhocNet.CA@Ucnet.UNCA.AdhocNet.CA Message-Id: <8901231439.AA06398@sirius.cs.rochester.edu> To: allan@cpsc.UCalgary.CA Subject: Re: BBN Butterfly Users Newsgroups: comp.parallel In-Reply-To: <4156@hubcap.UUCP> Organization: U of Rochester, CS Dept, Rochester, NY Cc: Status: R The University of Rochester got its first Butterfly, a 3-node Butterfly 1 with a discrete component switch in the fall of 1984. In the summer of 1985 we purchased a 128-node machine, 8 spare nodes, and an extra 16-way switch, which we configured as a 120-node machine and a 16-node machine. In 1986 we upgraded the 16-node machine to the 68020 daughterboards. In January of 1988 we upgraded the 3-node machine to be a Butterfly Plus. Later that spring we sliced off one quarter of the switch of the 120-node machine (leaving it with 96 nodes) and traded in the remaining nodes to get a 24-node Plus. So now we have a 96-node Butterfly 1, a 16-node Butterfly 1 with daughterboards (and a VME adaptor), a 3-node Butterfly Plus, and a 24-node Butterfly Plus. Our work with the Butterfly 1's is documented in an article in last summer's PPEALS conference. The Pluses are being used primarily for operating systems research, notably the Psyche multiprocessor O.S. project and the Platinum testbed for experimentation with non-uniform access time memory management issues. Psyche is described in a paper at last summer's ICPP. Platinum is described in a Rochester TR. -- Michael L. Scott University of Rochester (716) 275-7745 scott@cs.rochester.edu scott%rochester@CSNET-RELAY {decvax, allegra, cmcl2}!rochester!scott >From floyd%BRL.MIL%UcEdu.UNCA.AdhocNet.CA@Ucnet.UNCA.AdhocNet.CA Tue Jan 24 09:32:47 1989 Date: Tue, 24 Jan 89 10:34:24 EST From: floyd%BRL.MIL%UcEdu.UNCA.AdhocNet.CA@Ucnet.UNCA.AdhocNet.CA To: allan@cpsc.ucalgary.ca Message-Id: <8901241034.aa15989@SMOKE.BRL.MIL> Status: R To: allan@cpsc.UCalgary.CA Subject: Re: BBN Butterfly Users Newsgroups: comp.parallel In-Reply-To: <4156@hubcap.UUCP> Organization: Ballistic Research Laboratory Cc: I am an employee of the U.S. Army Ballistic Research Laboratory. We currently have a 16 node (processor) Butterfly. It is idle due to the lack of a file server. We are in the process of an upgrade to make this machine standalone. I am not aware of what stage the procurement process is in for the upgrade. I have reason to believe that I shall be the administrator and a possible user for this machine. It seems that I have been tasked to do so. I am interested in your idea to keep the users in touch. I currently do not have a project which I believe would exploit the power of the architecture. This is where I would like to hear from some of the community. I am an EE with a major in signal processing. I have been doing some recent reading in neural networks. It seems that most of the implementations are on non-parallel machines (Connection not included). I am not sure I want to implement a NN on 15 68020s, but I have not seen much which points to parallel architecture implementations. This may be an absurd notion, but I would be interesten if work is being done here. I would be interested in seeing some references to parallel algorithm implementation of signal processing techniques such as spectral analysis, adaptive filtering/beamforming or some pattern recognition/image processing work. Until this upgrade is complete I would request that mention of my interest be withheld. With the upgrade I will be fully involved with work on the machine. As of this moment it is vapor. Floyd Wofford floyd@smoke.brl.mil >From yiannis%cis.ohio-state.edu%UcEdu.UNCA.AdhocNet.CA@Ucnet.UNCA.AdhocNet.CA Tue Jan 24 19:37:54 1989 Date: Tue, 24 Jan 89 21:29:11 EST From: yiannis%cis.ohio-state.edu%UcEdu.UNCA.AdhocNet.CA@Ucnet.UNCA.AdhocNet.CA Message-Id: <8901250229.AA04341@junk.cis.ohio-state.edu> To: allan@cpsc.ucalgary.ca Subject: Bfly newsgroup Status: R We (at Ohio State University) have a Butterfly GP1000 with 10 nodes running Mach O/S. We have designed and implemented a thread library which originally was running under chrysalis O/S but now have been succesfully ported under Mach (the Bfly Mach doesn't provide threads). We would be definitely very intersted in a Bfly newsgroup (I am currently writing my thesis on the thread library I mentioned above). Yiannis >From utai!gatech!mailrus!BBN.COM!yeager.bbn.com!djoslin Fri Jan 27 09:56:05 1989 To: BBN.COM!cpsc.UCalgary.CA!allan@mailrus.uucp Subject: Re: BBN Butterfly Users Newsgroups: comp.parallel In-Reply-To: <4156@hubcap.UUCP> Organization: BBN Advanced Computers, Inc., Cambridge MA Cc: yeager.bbn.com!djoslin@mailrus.uucp Date: Fri, 27 Jan 89 09:49:11 EST From: mailrus!yeager.bbn.com!djoslin@gatech.uucp Status: R I'd be interested in what you find out. I work in the OS group doing Mach Unix on the Butterfly. As you might assume, we have a number of Butterfly machines here, and we all think that they are wonderful. :-) David >From jr%CHIPS.BBN.COM%UcEdu.UNCA.AdhocNet.CA@Ucnet.UNCA.AdhocNet.CA Wed Feb 1 13:38:09 1989 To: allan@cpsc.ucalgary.ca Subject: Re: BBN Butterfly Users Newsgroups: comp.parallel In-Reply-To: <4156@hubcap.UUCP> Organization: BBN Systems and Technologies Corporation, Cambridge MA Cc: Date: Wed, 1 Feb 89 14:30:45 EST From: jr%CHIPS.BBN.COM%UcEdu.UNCA.AdhocNet.CA@Ucnet.UNCA.AdhocNet.CA Sender: jr%CHIPS.BBN.COM%UcEdu.UNCA.AdhocNet.CA@Ucnet.UNCA.AdhocNet.CA Status: RO Have you gotten a response? If not, write back and I'll put you in touch with someone at BBN ACI. -- /jr jr@bbn.com or bbn!jr %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% Jeff Allan University of Calgary Calgary, Alberta, Canada uunet!ubc-cs!calgary!allan (403) 289-7877 (res)