Active from 2000 - 2002
To obtain best performance in a multi-processor chip, on-chip resources such as cache memory and off-chip bandwidth must be managed dynamically. In this project, ESP is used to evolve neural networks to decide how to continuously reallocate the available cache banks to processors. Trace-based simulations show that significant performance increases over static allocation are possible this way.
Faustino Gomez Postdoctoral Alumni tino [at] idsia ch
A Neuroevolution Method For Dynamic Resource Allocation On A Chip Multiprocessor 2001
Faustino J. Gomez, Doug Burger, and Risto Miikkulainen, In Proceedings of the {INNS-IEEE} International Joint Conference on Neural Networks, pp. 2355-2361, Piscataway, NJ 2001. IEEE.