Index to Research Displays

Computer Engineering
E-commerce
Education
Grid computing
High Performance Computing Applications
High Performance Computing Systems Research
Networking
Problem-Solving Environments and Portals
Security
Virtual Reality and Advanced Visualization

Supercomputing in the Memory

Peter M. Kogge, kogge@cse.nd.edu

Projects at Notre Dame such as HTMT have leveraged new technologies such as Processing-In-Memory to place supercomputer-level compute logic in the memory system, and create a new execution model that allows for efficient use of such logic in very highly scalable systems.

http://www.cse.nd.edu/~pim/

QCA Technology: Leaving Silicon Valley in the Dust

Craig Lent, lent@ee.nd.edu
Peter M. Kogge, kogge@cse.nd.edu

Quantum Cellular Automata (QCA) performs computation via Coulombic interaction, without current flow. When reduced to the molecular level this emerging technology may provide radically new architectures with orders of magnitude better characteristics than the best of future CMOS.

http://www.cse.nd.edu/~cse_proj/qca_design/

Computing with Energy Awareness

Peter Kogge, kogge@cse.nd.edu
Vincent Freeh, Vincent.W.Freeh.1@nd.edu
Jay Brockman, Jay.B.Brockman.1@nd.edu

As technology advances, our ability to leverage it will more and more be affected by our ability to cool the systems. The Morph project is developing new microarchitectural techniques that may significantly reduce the inherent energy cost of executing an instruction.

Back to Top

Rose-Hulman Ventures

Dee Reed, dee.reed@rhventures.org
Drew Gehman, drew.gehman@rhventures.org

Rose-Hulman Ventures is a technology-based business incubator and product-development center. Made possible by a grant from the Lilly Endowment, Rose-Hulman Ventures gives Rose-Hulman faculty, staff, and students opportunities to grow professionally in the fields of engineering, mathematics, and science, and encourages technologically adept graduates to remain in Indiana. A goal of the program is to attract, develop, and retain high-tech industries in Indiana so that those students, faculty, staff and others can apply their talents to enhancing Indiana's technological competitiveness and support the state's economic growth.

http://www.rhventures.org/

Back to Top

Information Technology in Undergraduate Education

Louis H. Turcotte, Louis.H.Turcotte@rose-hulman.edu

For more than a 125 years, Rose-Hulman has worked to provide the best undergraduate engineering, science, and mathematics education. Rose-Hulman is particularly intent that students receive a balanced education that enables them to obtain an excellent understanding of their area of specialty and an understanding of how their actions impact the world around us. Information technology is a key component of the educational experience at Rose-Hulman.

http://www.rose-hulman.edu/

Back to Top

Community Grids for Education and Computational Science

Geoffrey Fox, gcf@indiana.edu

We will describe the current Garnet collaborative environment which is being used to support distance education in conjuction with the Access Grid and lower-end Audio/Video systems. Combining Garnet with the portal Gateway gives an initial collaborative system for computational science research. We will also describe how the systems will be evolved to include peer to peer and Web service concepts.

http://communitygrids.iu.edu/

Middleware Architectures for Mobile Object Systems

Ananth Grama, ayg@purdue.edu
Paul Ruth, ruthp@purdue.edu
Muralikrishna Ramanathan, rmk@cs.purdue.edu
Tzvetan Horozov, horozov@purdue.edu

The large number of computing devices, ranging from handheld devices such as PDAs and cell-phones to large-scale wired computing platforms, provide a tremendous computational and information resource. The goal of this research is to develop middleware for (i) enabling resource and service sharing across peers; (ii) improving location mechanisms and network traffic in conventional peer-to-peer (P2P) systems by adapting network topology to reflect access patterns; and (iii) improving performance by optimally mapping and replicating services in proximity to peers using these services.

http://www.cs.purdue.edu/people/ayg/

Parasitic Computing

Albert-Laszlo Barabasi, Albert.L.Barabasi.1@nd.edu
Vincent Freeh, Vincent.W.Freeh.1@nd.edu
Hawoong Jeong, Hawoong.Jeong.1@nd.edu
Jay Brockman, Jay.B.Brockman.1@nd.edu

As a recent experiment has shown, it is possible to perform all the functions of a digital computer on the Web WITHOUT using classical CPUs at the Web nodes. The trick is in using features of the TCP/IP protocol.

http://www.nd.edu/~parasite/

GriPhyN

Randall Bramley, bramley@indiana.edu
Rob Gardner, rwg@indiana.edu

We are developing a portal (GRAPPA) for physicists to easily submit requests to run jobs on Condor flocks, and a testbed for bulk file transfer services between Fermilab and Indiana University using middleware products developed for grid applications.

http://lexus.physics.indiana.edu/griphyn/

PUNCH

Purdue University and Cantiga Systems
Nirav Kapadia, nirav@cantiga.com

PUNCH is a platform for Internet computing that turns the World Wide Web into a distributed computing portal. Users can access and run programs via standard Web browsers. Machines, data, applications, and other computing services can be located at different sites and managed by different entities. The PUNCH computing platform enables dynamic, on-demand creation of application and problem solving environments amongst assigned resources.

http://www.punch.purdue.edu/

Crystallography Grid Systems

John C. Huffman, huffman@indiana.edu
John C. Bollinger, jobollin@indiana.edu
Randall Bramley, bramley@indiana.edu
Donald F. (Rick) McMullen, mcmullen@indiana.edu

We will showcase several of our ongoing projects, including the XPort collaboratory and the distributed molecular database. XPort is a system that will provide "better than being there" tools for remote use of scientific instruments, including real-time data visualization and analysis.

http://www.iumsc.indiana.edu/

Back to Top

Stochastic Inlet Conditions for Large Eddy Simulation of a Fully-Turbulent Jet

David J. Glaze, glaze@purdue.edu
Steven H. Frankel, steven.h.frankel.1@purdue.edu

Simulating turbulent flow remains one of the great challenges of Computational Fluid Dynamics due to massive grid resolution requirements. A three-dimensional Large Eddy Simulation (LES) of a fully-turbulent jet is used to test the inlet condition methodologies, with experimental data used for validation. Results and some of the parallel computational issues will be discussed relating to the use of MPI with the compact finite difference solution technique, the grid filtering, and the Weighted Amplitude Wave Superposition (WAWS) inlet conditions.

http://widget.ecn.purdue.edu/~frankel/research.html

Enabling Technologies for Computational Molecular Biology

Jesús A. Izaguirre, Jesus.Izaguirre.1@nd.edu

MUSICO is an approximate molecular dynamics (MD) model that enables long simulations of MD by using a symplectic implicit-explict integrator and selective constraining. Also key is how computer scientists can greatly assist in the interfaces to set up complicated molecular simulations demanded by post-human genome computational biology.

http://www.nd.edu/~izaguirr/Career2001/node32.html

Computational Phylogenetics

Craig A. Stewart, stewart@iu.edu
David Hart, dhart@indiana.edu
Gary J. Olsen, gary@phylo.life.uiuc.edu

fastDNAml is a program for the maximum likelihood of phylogenetic trees from DNA sequence data. Maximum likelihood analysis of phylogenetic trees is extremely computationally intensive. fastDNAml is implemented in serial, PVM, and MPI versions.

http://www.indiana.edu/~rac/hpc/fastDNAml/index.html

The Advanced Computer Matrix for Engineering

David Moffett, dpm@purdue.edu
Garrett Jeong, gjeong@ecn.purdue.edu

Clusters are becoming the leading way to solve large, complex, time-consuming problems. The Advanced Computer Matrix for Engineering (ACME) is one means of implementing cluster technology with an extremely low budget. A personal cluster belonging to two researchers, ACME is currently being used to solve a 25-CPU-year transportation planning problem. In its idle time, it is working on large number factoring. Its next major project will be the structural simulation and analysis of asphaltic cement pavements.

http://acme.ecn.purdue.edu

Interactive Molecular Dynamics Simulations

Jesús A. Izaguirre, Jesus.Izaguirre.1@nd.edu

This demonstration will showcase interactive molecular dynamics (MD) simulations. We will demonstrate our framework for MD, ProtoMol, connected with the visualization program VMD from the University of Illinois. We will show enhanced haptic interfaces for the setup and steering of MD simulations. These simulations benefit from longer time step algorithms that speed up the simulation engine, from fast visualization using OpenGL, and from the capabilities of haptic interfaces.

http://www.nd.edu/~izaguirr/

Purdue Weather Models and Applications

Wen-Yih Sun, wysun@purdue.edu
Ben MacCall, maccall@purdue.edu
Ki-Hong Min, min@purdue.edu
Shu-Hua Chen, shachen@ucdavis.edu
Jiun-Dar Chern, jchern@dao.gsfc.nasa.gov
Wu-Ron Hsu (National Taiwan University),
Patrick A. Haines (ARL)

The Purdue atmospheric models based on geofluid dynamics equations have been applied to several phenomena, including severe weather, flooding, and drought in North America and Asia. These models have also been successfully applied to simulate lee vortices, downslope winds, turbulence, thermals, air pollution, and so on. Parallel versions of these models have been developed and run on Purdue's IBM SP system.

http://www.eas.purdue.edu/mml

Back to Top

Comparing Clusters and Supercomputers

Steven Gottlieb, sg@indiana.edu

PC clusters and supercomputers from vendors such as Cray, IBM, SGI and Sun have been used to study lattice QCD. This problem on a uniform grid of space-time points that employs sparse matrix algorithms such as conjugate gradient and bicgstab is easy to parallelize with MPI. Performance on a range of problem sizes, number of nodes, and computers is compared.

http://www.physics.indiana.edu/~sg/milc/benchmark.html

Advanced Compilation of Computational Applications

Ayon Basumallik, basumall@purdue.edu
Rudolf Eigenmann, eigenman@purdue.edu

We have created the Polaris compiler infrastructure, one of only two available research compiler infrastructures for experimentation with advanced HPC compilation techniques. We will demonstrate PCOMP, a compiler that translates applications written in the OpenMP API onto parallel machines that provide a shared address space. We will use the NETCARE network computing system. NETCARE makes it possible to run applications and tools on the Web.

http://www.ece.purdue.edu/ParaMount/
http://netcare.purdue.edu/Netcare/indexp.html

Middleware for Distributed Data

Beth Plale, plale@cs.indiana.edu

Extracting data from temporal event flows in large-scale applications is treated as extracting data from a persistent database using database queries, specifically SQL. This lets a user create optimized queries embedded in event flows efficiently enough to maintain high throughput for non-trivial queries in wide-area networks.

http://www.cs.indiana.edu/~plale/projects/

LAM/MPI

Andy Lumsdaine, lums@cs.indiana.edu
Jeff Squyres, jsquyres@indiana.edu
Brian Barrett, brbarret@indiana.edu

The LAM/MPI team is focusing on improving the ability of LAM to run on very large clusters. This includes a project called lamtree, which reduces the amount of time taken to bring up the run-time environment and tolerates errors during the boot sequence. Support for Myrinet hardware has reached a late beta stage, allowing for much higher speed communication on clusters. The LAM team is part of the OSCAR project, which seeks to provide an open source, unified, and easy to use cluster install and maintenance system.

http://www.lam-mpi.org/

Back to Top

Task Allocation for Wirelessly Networked Handheld Devices

Zhiyuan Li, li@cs.purdue.edu
Rong Xu, xur@cs.purdue.edu
Cheng Wang, wangc@cs.purdue.edu
Peifeng Ni, npf@cs.purdue.edu

This project investigates techniques to improve quality of service and battery energy usage on handheld devices. On such devices, it is often possible to improve program speed and to save the energy on the handheld by offloading its computation to the server. Our experimental results show that, even under an ordinary, uncontrolled, wireless LAN environment, the scheme can result in significant program speed improvement and energy-saving for more than half of the multimedia benchmark programs we tested.

http://www.cs.purdue.edu/homes/li/Mobile.html

Back to Top

Science Portals

Dennis Gannon, gannon@indiana.edu
Randall Bramley, bramley@indiana.edu

The science portal project is an NCSA/NSF funded effort to build a portable problem solving environment for Grid computing. The portal is a Web server installed on the user's desktop/laptop, designed to provide a "Notebook" interface to application scripting and data management.

http://www.extreme.indiana.edu/an/index.html

Back to Top

Secure Outsourcing

John R. Rice, jrr@cs.purdue.edu
Mikhail Atallah, mja@cerias.purdue.edu
John T. Korb, jtk@cs.purdue.edu

The Platform for Secure and Integrated Computing is a platform for cooperative computing where no one has to reveal their problems or data to anyone else. One can outsource work to remote computers and software with complete privacy; both the input data and answers are secure. Outsourcing provides resources: computing power for big problems, software for specialized or unusual applications, or access to proprietary data, software, and hardware. Security is achieved through disguises, which are application area dependent. Disguise is based on random but specialized changes in data and the problem.

http://www.cs.purdue.edu/people/jrr/

Management of Computer Security

Pascal Meunier, pmeunier@purdue.edu

Prevention and response in computer security respectively are two themes addressed by the Cassandra tool and the CERIAS Incident Response Database (CIRDB). The Cassandra tool allows you to create saved profiles of the services and applications running on your networks, typical (standard configurations) hosts, or important hosts. Cassandra can then notify you by e-mail of new vulnerabilities relevant to your profiles. The CERIAS Incident Response Database is a Web-based system intended to be used while responding to incidents. Its primary objective is assessing the cost of incidents for various uses and criteria. Therefore it supports classifications of incidents (including custom ones) and billing.

https://cassandra.cerias.purdue.edu/

Back to Top

iUniverse

Katy Börner, katy@indiana.edu
Mike Boyles, mjboyles@iupui.edu

The School of Library and Information Science at IUB in collaboration with the Advanced Visualization Laboratory recently installed an Active Worlds-powered universe server for Indiana University. The uniserver - named iUni - can host a large number of virtual worlds making Indiana University the first university in the US that provides free hosting of collaborative virtual worlds for faculty, staff, and graduate students.

http://iuni.slis.indiana.edu/

3D Image/Model database

Patrick Flynn, flynn@nd.edu
Kevin Bowyer, Kevin.W.Bowyer.2@nd.edu
Robert Stevenson, Robert.L.Stevenson.1@nd.edu

This database contains several thousand 3D images acquired from custom and commodity range scanners. It was created in 1988 at Michigan State University and has resided at Washington State University and Ohio State University and is now at Notre Dame. It is the premier archive of public-domain images for range image segmentation studies. The database also contains several hundred models of 3D objects obtained from various sources. All data files are in easily parsed text-based neutral formats.

http://www.nd.edu/~cvrl

Advanced Biometrics Research

Patrick Flynn, flynn@nd.edu
Kevin Bowyer, kwb@cse.nd.edu

This research is supported by DARPA, and involves development of databases, baseline algorithms, and advanced algorithms for human identification at a distance. Current efforts include large-scale image and video database collection, face recognition in intensity and infrared imagery, gait extraction and human identification from gait patterns, and shape-based biometrics involving the hand, the ear, and the face. Applications include authentication, security, and military operations.

http://www.nd.edu/~cvrl/

Back to Top