Subject: Re: I am working on Application Section
Resent-Date: Mon, 08 Nov 1999 17:06:04 -0500
Resent-From: Geoffrey Fox <gcf@npac.syr.edu>
Resent-To: p_gcf@npac.syr.edu
Date: Mon, 8 Nov 1999 15:30:55 -0600
From: Linda Torczon <linda@rice.edu>
To: gcf@npac.syr.edu

Geoffrey -

Included below is the LaTeX source for the Applications
section of the 1998 CRPC Annual Report.

Linda

========================================================

\chapter{Research Accomplishments and Plans}

\section{Research} \label{research}

The Center for Research on Parallel Computation ({\small CRPC})
is committed to a program of research that is making
parallel computer systems truly usable.
{\small CRPC} research efforts focus on six major thrusts:
{\em Fortran Parallel Programming Systems,}
%directed by Ken Kennedy;
{\em Parallel Paradigm Integration,}
%directed by Mani Chandy;
{\em Linear Algebra,}
%directed by Jack Dongarra;
{\em Optimization and Automatic Differentiation,}
%directed by John Dennis;
{\em Parallel Algorithms for Physical Simulations,}
%co-directed by Dan Meiron and Mary Wheeler.
{and \em Applications.}
% directed by Geoffrey Fox;
The following subsections detail our progress in each of these areas
since October 1, 1997.
Because of space limitations, references to our own work and
that of others have been omitted.  Detailed citations can be found in the technical
papers listed in Appendix~E.

\subsection{Applications} \label{Applications-Accomplishments}

{\em STC Faculty:}  Geoffrey Fox (Director)

\noindent
Nearly every {\small CRPC} research thrust area
applies its emerging technologies in practical
proving grounds like Grand Challenge projects,
industrial applications, educational applications,
and projects with other institutional or agency
partners.  While {\small CRPC} involvement in Grand
Challenges is diminishing, applications are increasing
through the {\small CRPC}'s association with {\small NSF PACI}
partners, {\small DARPA}, {\small ASCI} Centers of Excellence,
the {\small DOD} modernization program, a
{\small DOE}/{\small AFOSR}/{\small NSF}/Boeing collaboration,
and other government agencies
and companies. The {\small NSF}'s high-performance connection
grants are expediting the applications of new technologies
by researchers at several {\small CRPC} sites. On a national
scale, {\small CRPC} Director Ken Kennedy's Co-chairmanship
of the Presidential Information Technology Advisory Committee
({\small PITAC}) supports the future development of all these
and other important national applications initiatives.

\paragraph{Grand Challenges.}
The Binary Black Hole Grand Challenge is in its final year.
This project has enabled the prediction of gravitational
waves formed from the collision of two black holes. {\small CRPC}
researchers at Syracuse and the University of Texas have
developed a portable Java visualization system and linkage of
{\small HPF} (Fortran90) with a Distributed Adaptive Grade
Hierarchy ({\small DAGH}). For more, see
{\tt http://www.npac.syr.edu/projects/bh/}.

\paragraph{ASCI.}
Through {\small DOE HPCC}, {\small DOE} 2000, and
{\small ASCI} funding, Los Alamos National Laboratory researchers
have built a C++ framework that provides high-level objects like arrays,
fields, meshes, and particle lists for high-performance simulations
on parallel architectures, the {\small POOMA}
(Parallel Object-Oriented Methods and Applications) Framework
({\tt http://www.cal.lanl.gov/Pooma}).
The first version, shipped to internal customers, has enabled important
Grand Challenge (Accelerators and Numerical Tokamak) and {\small ASCI}
(Multimaterial hydrodynamics) customers to field major new codes.
{\small POOMA} II is being written now, and will include a much
enhanced Array class and task parallelism.

\paragraph{NSF PACI Partners.}
{\small NPACI} support has enabled the University of
Tennessee and Oak Ridge National Laboratory ({\small ORNL}) to
apply {\small CRPC}'s NetSolve ({\tt http://www.cs.utk.edu/netsolve/})
in the MCell simulation program
developed at the Salk Institute. This program allows simulation
of microscopic cellular processes, such as how neurotransmitters
diffuse and activate receptors in synapses between different
cells. The Netsolve system allows users to distribute processing
workload and access computational resources across a network.
MCell, with the help of NetSolve, can now distribute its heavy
workload among several machines simultaneously. See
{\tt http://www.npaci.edu/online/v2.16/} under ``Research'' for
more details.

Alliance support  enables {\small CRPC} researchers at
Syracuse and {\small ANL} to use their {\small HPCC}
Commodity metacomputing approach to provide a high-level
interface to Globus applied to Quantum Simulation Grand
Challenges. This work brought about the first general-purpose
commodity front end to {\small HPCC} technology
({\tt http://www.npac.syr.edu/users/haupt/WebFlow/demo.html}).

\paragraph{DARPA.}
{\small CRPC} researchers at Caltech have been advancing
{\small DARPA}-funded Parallel C3I (Command, Control,
Communication, and Intelligence) applications using commodity
hardware and software. Using multithreaded libraries and
pre-compilers, they have developed parallel C3I applications
like aircraft route optimization. Important developments
include (1)~a multithreaded library on Windows and Posix used
successfully for maintaining sequential semantics and getting
excellent speedups and (2)~a source-to-source compiler on top
of Microsoft Visual Studio for Multithreaded C created to
allow users to develop sequential programs in C in Visual
Studio that are then executed in parallel. The second was
applied to two substantial C3I problems with excellent speedup.
See {\tt http://www.cs.caltech.edu/$\sim$threads}.

\paragraph{DOD Modernization.}
The {\small DOD} has sponsored the recently completed Rice/University
of Tennessee project, Sca{\small LAPACK}, a library for dense
linear algebra on distributed memory machines. Sca{\small LAPACK},
has been used to achieve a significant improvement in performance
for the Global and Basin-Scale Ocean Modeling and Prediction
Challenge Project described at
{\tt http://www.hpcm.dren.net/Htdocs/Challenge/FY98/index.html}.
{\small OCEANS} is an ocean circulation model that employs
second-order finite differences with an explicit leapfrog time
stepping scheme on a logically rectangular grid. The {\small OCEANS}
preprocessor constructs and inverts large systems of equations
that represent boundary conditions for the {\small OCEANS}
program. Use of Sca{\small LAPACK} routines resulted in a factor
of three-to-six speed up over the previous version of the code.

{\small CRPC} researchers at Tennessee are also collaborating
with {\small DOD} application scientists on a solution using iterative
methods that is expected to achieve an efficient parallel
implementation of the {\small DOD} {\small HPC CHSSI CEA}-3 Spectral
Domain {\small LO} Component Design Code. The application requires
solution of a sparse, complex, symmetric linear system. Although
library software does not currently exist for this case, sufficient
research has been done on the complex symmetric problem to allow
implementation of effective methods. This project is expected to
result in an efficient implementation of the {\small CHSSI}
{\small CEA}-3 code that will require considerably less memory
than the current implementation. The project is also expected to
produce sparse linear system solver methods and software for the
complex symmetric case that will benefit {\small CEA} and other
{\small DOD} users who need to solve similar problems. See
{\tt http://www.hpcm.dren.net/Htdocs/CTAs/CEA.html} for more
information about {\small CEA}-3.
See {\tt http://www.cs.utk.edu/$\sim$eijkhout/parpre.html} for more
information about parallel preconditioners for iterative methods.

{\small CRPC} researchers at the University of Texas are doing
environmental quality modeling for the {\small DOD} Modernization project
at {\small CEWES}.  They have been involved in the parallel
migration of the hydrodynamics code {\small ADCIRC} and the
reactive transport code {\small CE\_QUAL-ICM}. This work has
enabled {\small CEWES} to run 10-year simulations.

\paragraph{DOE/AFOSR/NSF/Boeing Collaboration.}
Funds for the following applications efforts by {\small CRPC}
researchers at Rice come from {\small DOE}, Boeing, {\small NSF},
and the Air Force Office of Sponsored Research ({\small AFOSR}):

\subparagraph{1)~Nozzle Design.} The {\small CRPC} is working with
a group in Boeing Commercial Airplane Group ({\small BCAG}) to
reduce the cycle time for designing nozzles, the inside parts
of engine housings.
%It takes about 3 hours on an {\small SGI}
%Challenge to get one function value.
The current length of a
design cycle is two weeks. {\small CRPC} researchers expect to
reduce that to approximately one day;

\subparagraph{2)~Planform Design.} The {\small CRPC} is working with
Boeing in designing planforms, the shape of the wing as viewed
from above.  This effort addresses difficult design problems
involving multiple objectives and nonlinear constraints;

\subparagraph{3)~Helicopter Rotor Blade Design.} The {\small CRPC}
is working with Boeing on problems involved in designing helicopter
rotor blades.  Wake simulation is the hard part of the problem, and
the effects at the tip of the rotor blade are the most difficult to
simulate. It takes about 4 hours on 16 fat nodes of an {\small SP}2.
The {\small CRPC} and Boeing collaborators built a new model that
reduces the cost to minutes and the bandwidth of the coupling
from 100s to 10s. The resulting software was the first to find the
best-known simulation-based design;

\subparagraph{4)~Model Management Framework Software.} {\small CRPC}
researchers are testing a C++ framework for managing the process of
optimizing extremely expensive functions. This is a reimplementation of
the {\small CRPC}'s F90 prototype that performed so well on the helicopter
rotor blade design example; and

\subparagraph{5)~Parts Nesting System.} {\small CRPC}-developed
{\small PDS} continues to be used at Boeing for Just-in-Time
manufacturing of aircraft parts.

\paragraph{Others.}
Caltech-, {\small LANL}-, and Syracuse-based {\small CRPC}
researchers are designing a computational infrastructure
to enable development and testing of models for earthquakes.
This provides a unique application of Parallel Fast Multipole
methods originally developed for astrophysics.
See {\tt http://www.npac.syr.edu/projects/gem/}.

Using multithreaded libraries and precompilers for parallel
execution of computational chemistry codes on multiprocessor
{\small PC}s running Windows, {\small CRPC} researchers at
Caltech have successfully implemented, from sequential code
to parallel implementation, several important computational
chemistry codes.

At Rice, researchers are applying Global Optimization Techniques
to the Least-Squares Phase Problem in X-Ray Crystallography,
thanks to {\small NSF} funding. Researchers expect to provide
better semi-local optimization techniques for obtaining an
atomic arrangement that fits the experimental intensity data.
Two techniques have been proposed: a modified Newton's method
and a weighting scheme. These techniques can not only be applied
to the phase problem where a zero or small residual in the
objective function is desired, but also to other application areas,
such as seismic problems, that can also be modeled as least-squares
problems.

The {\small NSF} is funding a Rice-based {\small CRPC} effort
to develop Successive Linear Programming ({\small SLP})
Software. As a result, the {\small COPILOT} package, which uses
successive linear programming and the user's choice of {\small LP}
solver, has been made more widely available. {\small COPILOT} has
been used to solve a problem in multidisciplinary optimization
from {\small NASA} Langley, and on standard benchmark sets.

\newpage
The Department of Energy and National Institutes of Health have
provided funding to {\small CRPC} researchers at Rice who actively
collaborate with Baylor College of Medicine researchers in the
Fitting Contrast Transfer Function in Electron Microscopy Project.
The project's goal is to best model the Fitting Contrast
Transfer Function of an electron microscope, a crucial step in
three-dimensional image reconstruction in computational biology.

A {\small CRPC} researcher at Rice, in collaboration with a
{\small CRPC} External Advisory Committee member from Sandia Labs,
is participating in the Sandia Labs {\small OPT}++ Engineering
Optimization System, funded by the Department of Energy. The
constraints version of {\small CRPC}-developed Parallel Direct
Search ({\small PDS}) has been incorporated into the Sandia Labs
{\small OPT}++ Engineering Optimization System.

With funding from the Pacific Northwest National Lab,
{\small CRPC} researchers at Syracuse University and
Argonne National Laboratory are developing new methodology,
algorithms, and tools for parallel computational chemistry.
The project has achieved record-sized {\small RI-MP}2 calculations
and a Global Array programming model extended to multi-dimensional arrays
({\tt http://www.npac.syr.edu/users/bernhold/comp\_chem/index.html}).

Additional {\small CRPC} applications efforts are discussed in
\S~\ref{optimization-applications}, \S~\ref{AD-applications},
\S~\ref{PAPS-applications}, and \S~\ref{PAPS-plans}.

\paragraph{VBNS Connections Enable Applications Process.}
{\small CRPC} sites at Caltech, Rice University, the
University of Tennessee, and the University
of Texas, and affiliated sites at Drexel University, Indiana University,
the University of Houston ({\small UH}), the University of Illinois,
and the University of Maryland have been awarded connections to the
very-high-speed Backbone Network Service (v{\small BNS}), an
experimental network launched in 1995 that connects national
supercomputing centers and universities across the country to
collaborate and share powerful computing and information resources.
For more information on the v{\small BNS}, see
{\tt http://www.cise.nsf.gov/} or {\tt http://www.vbns.net}.

Rice University, {\small UH}, and the Baylor College of Medicine
were among 13 original institutions awarded links to the v{\small BNS}
in 1996 as part of the Houston Area Computational Science Consortium
({\small HACSC}). This {\small CRPC} collaboration is using the
v{\small BNS} to conduct computationally intensive research projects
in areas like surgical simulation, 3-D modeling of molecular structures,
computational steering of hydrocarbon reservoir models, and aircraft
design. Rice researchers are developing the software infrastructure
that will make such distributed computer systems usable.

Initially connected at speeds of 45 megabits per second, the
{\small HACSC} established in December 1997 a fledgling Texas
Gigapop, a value-adding, three-layer meet point where customers
can meet providers.  This will eventually be the main connection
for the Houston schools to the Internet 2 consortium.

Researchers at the University of Texas are using the v{\small BNS}
for environmental simulations of groundwater transport in porous
media.  Ongoing collaborations with the {\small CRPC} and the U.S.
Army Corps of Engineers Waterways Experiment Station, as part of {\small DOD}
modernization work, are focused on the performance of water models
for distributed memory computers, such as the long-term environmental
impact of flows in bays and estuaries. The v{\small BNS} allows
timely access to large-scale data for applications that involve 3-D,
time-dependent simulations that can require hundreds of realizations
to complete a single analysis in a given region.

{\small CRPC} site Caltech was awarded a connection to the
v{\small BNS} through its affiliation with the Corporation for
Education Network Initiatives in California ({\small CENIC}), a
nonprofit benefit corporation formed by California universities
to achieve robust, high-capacity, Next Generation Internet
communications services for the higher education academic and
research communities. Its goals are to oversee and deploy a
cost-effective, state-of-the-art statewide communications
infrastructure for all post-secondary institutions in California,
and to support ready access to supercomputing facilities and
collaborations with federally sponsored research laboratories and
partners.

\newpage
{\small CRPC} affiliated site University of Maryland is using the
v{\small BNS} to connect to supercomputer sites as part of its
involvement with the National Partnership for Advanced Computational
Infrastructure ({\small NPACI}). The Maryland work focuses on
programming tools and environments and involves creating testbeds
to develop experimental tools and streamline their dissemination
and support. These tools are being evaluated on a scale that has not
been feasible previously.

Other {\small CRPC} affiliated sites are using the v{\small BNS} for
projects ranging from performance analysis tool development to stellar
simulations. Indiana University's link is being used for wide-area
metacomputing and tele-immersive ({\small CAVE}-to-{\small CAVE})
visualization experiments. Drexel University's newly-awarded connection
will be used for research in star cluster dynamics, numerical modeling,
computational astrophysics, advanced cluster computing, etc.
University of Illinois at Urbana-Champaign is using the v{\small BNS}
for research on performance tool development, I/O characterization,
visualization, and collaboration, and a wide range of other projects
related to the {\small NSF}'s Partnership for Advanced Computational
Infrastructure ({\small PACI}). Illinois is the headquarters for the
National Computational Science Alliance ({\small NCSA}), which, along
with {\small NPACI}, is a lead institution for {\small PACI}.

\paragraph{Presidential Information Technology Advisory Committee
(PITAC) Applications Plans.}
One of the goals set by the {\small PITAC}, of which {\small CRPC}
Director Ken Kennedy is Co-chair, is to demonstrate new applications
on the Next Generation Internet ({\small NGI}) that meet important
national goals and missions. Higher-speed, more-advanced networks
will enable a new generation of applications that support scientific
research, national security, distance education, environmental monitoring,
and health care. The principal agencies involved in this initiative are
the National Science Foundation, the Defense Advanced Research Projects
Agency, the Department of Energy, {\small NASA}, and the National
Institutes of Health. Other agencies may be involved in promoting
specific applications related to their missions. {\small CRPC}
applications, such as those described above, are helping to pave
the way for these {\small NGI} applications.

========================================================
>Did you send / Do you have what I did last year
>
>Geoffrey Fox  gcf@npac.syr.edu,   http://www.npac.syr.edu
>Director of NPAC and Professor of Physics and Computer Science
>Phones Cell 3152546387  Office 3154432163 Npac 3154431723 Fax 3154434741