Download Parallel and Distributed Computing Systems Lab.

Parallel and Distributed
Computing Systems Lab.
Ananth Grama
Associate Professor of Computer Sciences
Purdue University
http://www.cs.purdue.edu/people/ayg
Research Agenda
• Software infrastructure for large-scale parallel and
distributed computing.
• Algorithms for system as well as application
kernels.
• Compute-intensive applications in scientific and
commercial domains.
• Techniques for compression and analysis of
extremely large data-sets generated from
simulations and other sources.
Lab Members
Ph.D. Students:
Undergraduate Research
– Ioannis Ioannidis,
Students:
– Paul Ruth,
– Chris Daniels.
– Lei Shan,
– Robert Light,
– Mehmet Koyuturk.
M.S. Students:
– Ramakrishna Muralikrishna,
– Tzvetan Horozov,
– Min Li.
Sources of Research Funding
• Six current National Science Foundation research projects
(as PI or Co-PI) totaling over $1.6M.
• Equipment support from National Science Foundation and
Intel Corp. totaling over $2M.
• Research support from National Institutes of Health
($140K) for medical data analysis.
• Research grant from CERIAS/Lilly Foundation, $50K.
• Fellowship grants from the Department of Energy and
Department of Education.
Teaching Interests
• Parallel and Distributed Computing (CS525,
CS590D).
• Numerical Analysis (CS514).
• Data Structures (CS251) and Compilers
(CS352).
Recent Awards and Honors
• National Science Foundation CAREER
Award, 1998.
• Purdue University School of Science
Outstanding Assistant Professor Award,
1999.
Interdisciplinary Collaborations
• Prof. Mete Sozen, Civil Engg., Active Structures.
• Prof. Thomas Downar, Nuclear Engg., Reactor
Simulations.
• Prof. Bruce Craig, Statistics, Medical Data Analysis.
• Profs. Kent Fuchs and Rudolf Eigenmann, Elect. Engg.,
Systems Infrastructure.
• Prof. Morry Levy (Biology) and Jun Xie (Statistics),
Curriculum Development for Bioinformatics.
Professional Activities and
Affiliations
• Member, Sigma Xi, American Association
for the Advancement of Sciences.
• Conference and workshop program
committees and organization.
• Referee for international journals and
funding agencies.
• Guest editor and author of journals and
books, respectively.
Technical Contributions
• How do you use a large number of computers to
solve a single large problem?
– Parallel Algorithms.
• How do you program such computers?
– System software development.
• How do you solve specific problems in parallel?
– Application development, molecular dynamics,
astrophysical simulations, VLSI modeling, scattering
and inverse scattering problems.
Technical Contributions
• How do you handle extremely large data-sets
generated from these simulations and other
sources?
– Data compression and analysis.
• What are emerging paradigms in parallel and
distributed computing?
– Peer-to-peer networks for sharing data,
computation, and resources.
Parallel Algorithms and
Applications
Large-scale particle dynamics simulations.
Molecular Dynamics
Simulating the behavior of large complex molecules.
Protein Structure Estimation
Software Development
• Sharing data, services, and resources over
the network.
– Clients such as Napster provide mechanisms for data
sharing.
– Distributed clients such as Gnutella and Limewire do
not have a centralized server and therefore are more
scalable.
– How do we make a better Gnutella? Improved resource
location, adaptive (content-based) network topologies,
mechanisms for location and mapping of services,
support for offloading computations and remote
services.
Data Mining and Analysis
“People who buy diapers in the evening are also
likely to buy beer!” -- Put them in the same aisle.
Image Compression.
Pattern Matching Compression
JPEG (Current Standard) at same compression
Video Compression
Real-time mobile media handlers.