From owner-chemistry@ccl.net Mon Nov 5 01:28:01 2007 From: "Don Steiger don.steiger * gmail.com" To: CCL Subject: CCL: Massive parallelization. Message-Id: <-35556-071105003725-28277-aoybm8YKoraG/1IlD8zxMA---server.ccl.net> X-Original-From: "Don Steiger" Date: Mon, 5 Nov 2007 00:37:21 -0500 Sent to CCL by: "Don Steiger" [don.steiger~~gmail.com] I am looking for a program that would make a good test case for parallelization. I want to try out some parallelization ideas that uses mpi-2 for intra-machine message passing, and Globus for inter-machine message passing (i.e. over the internet). In addition, instead of using a passive synchronization, which mpi programs normally use, I want to try an approach that uses a dedicated processor to coordinate the processing performed by the individual processors. I'm hoping that this kind of parallelization scales well when the number of processors are large and spread across multiple machines. There are three key requirements for an appropriate test case: 1) the algorithm has an underlying parallel structure, 2) the scalar version of the code is reasonably bug free, and 3) the code must be either very well documented, or reasonably simple to figure out, or somebody familiar with the source be willing to take some time to provide guidance. If you have, or know of, a scalar program that you would like to port to a parallel environment, and want somebody else to do the work then let me know. Don don.steiger,gmail.com From owner-chemistry@ccl.net Mon Nov 5 03:14:01 2007 From: "Andreas Bick andreas.bick/a\scienomics.com" To: CCL Subject: CCL: Massive parallelization. Message-Id: <-35557-071105025807-5587-up8sq4Sww+3xLge/1mgXww-x-server.ccl.net> X-Original-From: Andreas Bick Content-Transfer-Encoding: 7bit Content-Type: text/plain; charset=ISO-8859-1; format=flowed Date: Mon, 05 Nov 2007 08:57:48 +0100 MIME-Version: 1.0 Sent to CCL by: Andreas Bick [andreas.bick . scienomics.com] Hi Don, I think LAMMPS, a classical MD code (http://lammps.sandia.gov/) fulfills your criteria. Regards Andreas >Sent to CCL by: "Don Steiger" [don.steiger~~gmail.com] >I am looking for a program that would make a good test case for >parallelization. I want to try out some parallelization ideas that uses mpi-2 >for intra-machine message passing, and Globus for inter-machine message passing >(i.e. over the internet). In addition, instead of using a passive >synchronization, which mpi programs normally use, I want to try an approach >that uses a dedicated processor to coordinate the processing performed by the >individual processors. I'm hoping that this kind of parallelization scales well >when the number of processors are large and spread across multiple machines. > >There are three key requirements for an appropriate test case: > > 1) the algorithm has an underlying parallel structure, > > 2) the scalar version of the code is reasonably bug free, and > > 3) the code must be either very well documented, or > reasonably simple to figure out, or somebody familiar with > the source be willing to take some time to provide guidance. > >If you have, or know of, a scalar program that you would like to port to a >parallel environment, and want somebody else to do the work then let me know. > > >Don >don.steiger- -gmail.com> > > > > > -- Dr. Andreas Bick Scienomics SARL Sales and Marketing mobile: +491621309945 From owner-chemistry@ccl.net Mon Nov 5 06:24:00 2007 From: "=?ISO-8859-1?Q?Sina_T=FCreli?= sinatureli(-)gmail.com" To: CCL Subject: CCL: Massive parallelization. Message-Id: <-35558-071105050417-27073-ZimWDxG6+muC48MoJ2hOcA-,-server.ccl.net> X-Original-From: "=?ISO-8859-1?Q?Sina_T=FCreli?=" Content-Type: multipart/alternative; boundary="----=_Part_11362_20306597.1194257018569" Date: Mon, 5 Nov 2007 12:03:38 +0200 MIME-Version: 1.0 Sent to CCL by: "=?ISO-8859-1?Q?Sina_T=FCreli?=" [sinatureli(-)gmail.com] ------=_Part_11362_20306597.1194257018569 Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: 7bit Content-Disposition: inline NAMD can be run on parallel machines and is free source http://www.ks.uiuc.edu/Research/namd/ On 11/5/07, Andreas Bick andreas.bick/ascienomics.com < owner-chemistry**ccl.net> wrote: > > > Sent to CCL by: Andreas Bick [andreas.bick . scienomics.com] > Hi Don, > > I think LAMMPS, a classical MD code (http://lammps.sandia.gov/) fulfills > your criteria. > > Regards > > Andreas > > >Sent to CCL by: "Don Steiger" [don.steiger~~gmail.com] > >I am looking for a program that would make a good test case for > >parallelization. I want to try out some parallelization ideas that uses > mpi-2 > >for intra-machine message passing, and Globus for inter-machine message > passing > >(i.e. over the internet). In addition, instead of using a passive > >synchronization, which mpi programs normally use, I want to try an > approach > >that uses a dedicated processor to coordinate the processing performed by > the > >individual processors. I'm hoping that this kind of parallelization > scales well > >when the number of processors are large and spread across multiple > machines. > > > >There are three key requirements for an appropriate test case: > > > > 1) the algorithm has an underlying parallel structure, > > > > 2) the scalar version of the code is reasonably bug free, and > > > > 3) the code must be either very well documented, or > > reasonably simple to figure out, or somebody familiar with > > the source be willing to take some time to provide > guidance. > > > >If you have, or know of, a scalar program that you would like to port to > a > >parallel environment, and want somebody else to do the work then let me > know. > > > > > >Don > >don.steiger- -gmail.com> > > > > > > > > > > > > > -- > Dr. Andreas Bick > Scienomics SARL > Sales and Marketing > mobile: +491621309945> > > > ------=_Part_11362_20306597.1194257018569 Content-Type: text/html; charset=ISO-8859-1 Content-Transfer-Encoding: 7bit Content-Disposition: inline
NAMD can be run on parallel machines and is free source
 
On 11/5/07, Andreas Bick andreas.bick/ascienomics.com <owner-chemistry**ccl.net> wrote:

Sent to CCL by: Andreas Bick [andreas.bick . scienomics.com]
Hi Don,

I think LAMMPS, a classical MD code (http://lammps.sandia.gov/) fulfills
your criteria.

Regards

Andreas

>Sent to CCL by: "Don  Steiger" [ don.steiger~~gmail.com]
>I am looking for a program that would make a good test case for
>parallelization.  I want to try out some parallelization ideas that uses mpi-2
>for intra-machine message passing, and Globus for inter-machine message passing
>(i.e. over the internet).  In addition, instead of using a passive
>synchronization, which mpi programs normally use, I want to try an approach
>that uses a dedicated processor to coordinate the processing performed by the
>individual processors. I'm hoping that this kind of parallelization scales well
>when the number of processors are large and spread across multiple machines.
>
>There are three key requirements for an appropriate test case:
>
>           1) the algorithm has an underlying parallel structure,
>
>           2) the scalar version of the code is reasonably bug free, and
>
>           3) the code must be either very well documented, or
>              reasonably simple to figure out, or somebody familiar with
>              the source be willing to take some time to provide guidance.
>
>If you have, or know of, a scalar program that you would like to port to a
>parallel environment, and want somebody else to do the work then let me know.
>
>
>Don
>don.steiger- -gmail.com>
>
>
>
>
>


--
Dr. Andreas Bick
Scienomics SARL
Sales and Marketing
mobile: +491621309945



E-mail to subscribers: CHEMISTRY**ccl.net or use:
     http://www.ccl.net/cgi-bin/ccl/send_ccl_message

E-mail to administrators: CHEMISTRY-REQUEST**ccl.net or use
     http://www.ccl.net/cgi-bin/ccl/send_ccl_message

Subscribe/Unsubscribe:
     http://www.ccl.net/chemistry/sub_unsub.shtml

Before posting, check wait time at: http://www.ccl.net

Job: http://www.ccl.net/jobs
Conferences: http://server.ccl.net/chemistry/announcements/conferences/

Search Messages: http://www.ccl.net/htdig  (login: ccl, Password: search)
     http://www.ccl.net/spammers.txt

RTFI: http://www.ccl.net/chemistry/aboutccl/instructions/




------=_Part_11362_20306597.1194257018569--