Peng Li at Washington University

Peng Li
Department of Computer Science & Engineering
Washington University in St. Louis
Contact Information:
Campus Box 1045
One Brookings Drive
St. Louis, Missouri
63130-4899
Email: lip AT cse.wustl.edu
Phone: 314-935-4845 (O)

Current Work|Publications|Links

Currently I am a PhD student of Computer Science, co-advised by Prof. Jeremy Buhler and Prof. Roger D. Chamberlain. I also work with Prof. Kunal Agrawal. My research area is streaming computation. I am also interested in reconfigurable computing, etc.

Before joining Wash.U., I received B.S. in Computer Science and Technology from the University of Science and Technology of China (USTC), Hefei, P.R.China in 2005 and M.S. in Computer Science from the Graduate University of Chinese Academy of Sciences, Beijing in 2008.

Current Work

Auto-Pipe: A concurrency platform that aids in the design, evaluation and implementation of pipelined applications that are distributed across a set of heterogeneous devices including multiple processors, GPU and FPGAs. I am working to improve the flexibility and efficiency of data communication across computational nodes. In other words, I am working on pipes. (Hmm...try to be a good plumber.)
Deadlock Avoidance for Streaming Computation with Filtering: When computational nodes have the ability to filter, or discard, some of their inputs, a system with finite buffering is vulnerable to deadlock. We formalized a model of streaming computing systems with filtering and developed provably correct mechanisms to avoid deadlock. The approach relies on adding extra ``dummy'' tokens to the data streams and does not require global run-time coordination among nodes or dynamic resizing of buffers.

Publications

Peng Li, Kunal Agrawal, Jeremy Buhler, and Roger D. Chamberlain, "Deadlock Avoidance for Streaming Computations with Filtering", in 22nd ACM Symposium on Parallelism in Algorithms and Architectures (SPAA'10), June 2010, Santorini, Greece.
Peng Li, Kunal Agrawal, Jeremy Buhler, Roger D. Chamberlain, and Joseph M. Lancaster "Deadlock Avoidance for Streaming Applications with Split-Join Structure: Two Case Studies", in 21st IEEE International Conference on Application-specific Systems, Architectures and Processors (ASAP'10), July 2010, Rennes, France. (shorter paper)