Seminar 20th May 2010 noon University of Southampton, Mountbatten 53/4025
Center for Adaptive Supercomputing Software
Dr David Halgin and Dr Jace Mogill
Pacific Northwest National Laboratory, US
- Web page
- http://cass-mt.pnl.gov/
- Categories
- Data Management, Distributed computing, HPC, Scientific Computing
- Submitter
- Petrina Butler
from http://www.cray.com/products/xmt/
Many fundamental science, national security, and business problems require fast, scalable solutions to highly unstructured analytical problems. There are worldwide operational deployments of instruments to detect the proliferation of weapons of mass destruction, monitor terrorist cells, and track the movement of illicit goods and services. In the next 15 years 30% of battle-space defense forces will be autonomous with each advanced robotic device carrying dozens of sophisticated sensors collecting, processing, analyzing and transmitting large amounts of data. American economic competitiveness will depend increasingly on the timely analysis of many Petabytes of data collected in diverse computing clouds charting the social and economic behavior of consumers.
Most analytical problems can be expressed as graph algorithms. Typically, these algorithms comprise integer operations, irregular data access patterns, low computation to memory access ratios, and high levels of fine grain parallelism that pass data and synchronize frequently. Conventional parallel computer systems optimized to run large-scale floating point intensive simulations are inadequate, and more suitable shared-memory, multithreaded architectures such as the Cray XMT are needed.
The Center for Adaptive Supercomputing Software at Pacific Northwest National Laboratory is funded by the Department of Defense to study data-intensive, analytic problems. The Center conducts research in parallel graph algorithms, programming languages, runtime systems, and multithreaded architectures. In this talk, I will survey our current research projects, describe the Cray XMT architecture, and discuss several of the scalable graph algorithms we have developed.
A light lunch and refreshments will be provided.