Skip to content. Skip to navigation
McGill Home SOCS Home
Personal tools
You are here: Home People Profile

Carl Tropper


photo

Email: carl AT cs DOT mcgill DOT ca
Home Page: http://www.cs.mcgill.ca/~carl/
Office: MC112N
Phone: +1-514-398-3743
Fax: +1-514-398-3883
Address:

Research Description

Professor Tropper's primary research interest is in the area of parallel discrete event simulation and more generally in parallel computing. Discrete event simulations are of fundmental importance in designing VLSI systems, computer networks, manufacturing systems among a plethora of systems. As these systems get larger (hundreds of millions of gates on a chip, multi core computers, the Intenet...) the demand for memory and compute cycles to perform these simulations has outstripped the increase in the number of trasistors which can be etched on a chip. Hence the objective of research in parallel simulation is to take advantage of parallel and distributed computing platforms (e.g. clusters of computers) in order to speed up these simulations and cope with memory demands. Developing synchronization techniques and load balancing algorithms constitute the heart of the research agenda. Hybrid simulators, the combination of continuous and discrete event simulations (such as astrophysical and weather simulations) are equally in need of parallel platforms and should benefit from the techniques developed for discrete event simulations.

Professor Tropper has been active in this area from its inception and has developed a number of synchronization and load balancing algorithms for both conservative and optimistic parallel simulations. In recent years his focus has been on the use of optimistic protocols for VLSI simulations. His research group has developed a distributed Verilog simulation environment environment for the purpose of experimenting with synchronization and partitioning mechanisms under an optimistic (Time Warp) protocol. In the area of hybrid simulations his group is attempting to integrate optimistic discrete event protocols with continuous solutions for astrophysical simulations. The use of reverse computation is a key component of this reseach.

Research Interests

Research Labs

Teaching

Selected Publications (click link in front of each publication to see bibtex in ASCII format)

[1] Lijun Li, C. T. A design-driven partitioning algorithm for distributed verilog simulation. Conference on Parallel and Distributed Simulation, IEEE, June 2007.
[2] Wang, J., and Tropper, C. Optimizing the time warp protocol with learning automata. In European Modeling and Simulation Conference, October 2007, p. 7 pages.
[3] Wang, J., and Tropper, C. Optimizing time warp simulation with reinforcement learning techniques. In Winter Simulation Conference, December 2007, p. 8 pages.
[4] Li, L., and Tropper, C. Design driven partitioning in verilog. In SCS Simulation Jouranl Special Issue, December 2007. In Press.
[5] Lijun Li, C. T. Event reconstruction vs dynamic checkpointing. IEEE,ACM, IEEE, May 2004, pp. 37-44. Workshop on Parallel and Distributed Simulation.
[ .pdf ]
[6] Lijun Li, Hai Huang, C. T. Towards distributed verilog simulation. International Journal of Simulation: Systems,Science and Technology, September 2003, v. 3, n. 3-4, pp. 44-55.
[ .pdf | .pdf ]
[7] Herve Avril, C. T. On checkpointing and rolling back in time warp. IEEE Transactions of Parallel and Distributed Systems, November 2001, v. 12, n. 11, pp. 1105-1122.
[ .pdf | .pdf ]

Last Update:   2012/08/30 14:59:12.724 GMT-4