Dynamic Scheduling in an Implicit Parallel System

Penny is a system that exploits ne grained paral lelism in an AKL program During execution a set of workers processes share a set of tasks that are cre ated dynamically To achieve good parallel speedup a worker must be assigned a new task as soon as it be comes idle Since no compiler support nor user anno tations are available to guide the execution the Penny system needs a very e cient dynamic scheduler We have developed a con gurable scheduler in or der to experiment with di erent approaches We eval uated and analyzed the approaches by running a small set of benchmarks for which statistics and performance are reported