Search

Abstract: Fault-tolerant scheduling is an imperative step for large-scale computational Grid systems, as often geographically distributed nodes co-operate to execute a task. By and large, primary-backup approach is a common methodology used for fault tolerance wherein each task has a primary and a backup on two different processors. In this paper, we address the problem of how to schedule DAGs in Grids with communication delays so that service failures can be avoided in the presence of processors faults. The challenge is, that as tasks in a DAG have dependence on each other, a task must be scheduled to make sure that it will succeed when any of its predecessors fails due to a processor failure. We first propose a communication model and determine when communications between a backup and backups of its successors are necessary. Then we determine when a backup can start and its eligible processors so as to guarantee that every DAG can complete upon any processor failure. We develop two algorithms to schedule backups, which minimize response time and replication cost, respectively. We also develop a suboptimal algorithm which targets minimizing replication cost while not affecting response time. We conduct extensive simulation experiments to quantify the performance of the proposed algorithms. [Copyright &y& Elsevier]

Abstract: This paper is a natural continuation of Abbott et al. [Abbott, J., Bigatti, A., Kreuzer, M., Robbiano, L., 2000. Computing ideals of points. J. Symbolic Comput. 30, 341–356] further generalizing the Buchberger–Möller algorithm to zero-dimensional schemes in both affine and projective spaces. We also introduce a new, general way of viewing the problems which can be solved by the algorithm: an approach which looks to be readily applicable in several areas. Implementation issues are also addressed, especially for computations over where a trace-lifting paradigm is employed. We give a complexity analysis of the new algorithm for fat points in affine space over . Tables of timings show the new algorithm to be efficient in practice. [Copyright &y& Elsevier]

Abstract: This paper presents a general model to define, measure and predict the efficiency of applications running on heterogeneous parallel computer systems. Using this framework, it is possible to understand the influence that the heterogeneity of the hardware has on the efficiency of an algorithm. This methodology is used to compare an existing parallel genetic algorithm with a new adaptive parallel model. All the performance measurements were taken in a loosely coupled cluster of processors. [Copyright &y& Elsevier]

Pfair scheduling was proposed by Baruah, Cohen, Plaxton, and Varvel as a non-work-conserving way of optimally and efficiently scheduling periodic tasks on a multiprocessor. In this paper, we introduce a work-conserving variant of Pfair scheduling called “early-release” fair (ERfair) scheduling. We also present a new scheduling algorithm called PD2 and show that it is optimal for scheduling any mix of early-release and non-early-release asynchronous, periodic tasks. In contrast, almost all prior work on Pfair scheduling has been limited to synchronous systems. PD2 is an optimization of an earlier deadline-based algorithm of Baruah, Gehrke, and Plaxton called PD; PD2 uses a simpler tie-breaking scheme than PD to disambiguate equal deadlines. We present a series of counterexamples that suggest that, in general, the PD2 tie-breaking mechanism cannot be simplified. In contrast to this, we show that no tie-breaking information is needed on two-processor systems. [Copyright &y& Elsevier]