Bounds and algorithms for a practical task allocation model (extended abstract)

A parallel computational model is defined which addresses I/O contention, latency, and pipe-lined message passing between tasks allocated to different processors and can be used for modeling parallel task-allocation on a network of workstations or on a multi-stage interconnected parallel machine. To study the performance bounds more closely, basic properties are developed for when the precedence constraints form a directed tree. Approximation algorithms are presented for scheduling directed one-level task trees on an unlimited number of processors and also on a fixed number of k, k > 1, processors, where the approximation ratios are 3 and 3 + k−2/k, respectively. The case of equal task execution times is also considered.