A static mapping heuristics to map parallel applications to heterogeneous computing systems

Abstract

SUMMARYIn order to minimize the execution time of a parallel application running on a heterogeneously distributed computing system, an appropriate mapping scheme is needed to allocate the application tasks to the processors. The general problem of mapping tasks to machines is a well-known NP-hard problem and several heuristics have been proposed to approximate its optimal solution. In this paper we propose a static graph-based mapping algorithm, called Heterogeneous Multi-phase Mapping (HMM), which permits suboptimal mapping of a parallel application onto a heterogeneous computing distributed system by using a local search technique together with a tabu search meta-heuristic. HMM allocates parallel tasks by exploiting the information embedded in the parallelism forms used to implement an application, and considering an affinity parameter, that identifies which machine in the heterogeneous computing system is most suitable to execute a task. We compare HMM with some leading techniques and with an exhaustive mapping algorithm. We also give an example of mapping of two real applications using HMM. Experimental results show that HMM performs well demonstrating the applicability of our approach.

Analysis of Different Variants in Round Robin Algorithms for Load Balancing in Cloud Computing

“…M. Katevenis, S. Sidiropoulos, and C. Courcoubetis have propsed Weighted round-robin cell multiplexing in a general-purpose ATM switch chip. [18] M. Shreedhar and G. Varghese have discussed the efficient fair queuing using deficit round robin. [19] Bhathiya Wickrema etal.…”

Alea – Grid Scheduling Simulation Environment

“…We are proposing a complex and extensible simulation environment unlike to frequently used ad-hoc simulators [2,12,24,3] that is able to model various situations such as different types of jobs and applications or different Grid topology and then evaluate proposed solutions and algorithms. The goal is to use the simulator to design Grid schedulers with differing scheduling techniques and test their behaviour in different but fully controlled conditions.…”

“…Currently we are focusing on local search based algorithms [11] and dispatching rules [13,22]. Various techniques were applied in related problems such as job scheduling on single machine [20,17], identical parallel machines [24,2] or heterogeneous parallel machines [21,3]. Unfortunately, these solutions are usually tested only in static situation, i.e., when all the jobs are known to the scheduler before execution.…”

“…Unfortunately, these solutions are usually tested only in static situation, i.e., when all the jobs are known to the scheduler before execution. One of our main interest is the dynamic aspect of scheduling [25,4,16], i.e., scheduling of jobs arriving during the system run [3,12], sometimes referred to as incremental scheduling [16] or on-line reasoning [4].…”

Comparison Of Multi-Criteria Scheduling Techniques

“…This allows us to use advanced scheduling algorithms [16,8] such as local search methods [8] to optimize the schedule. Due to their computational cost, these approaches were mostly applied to static problems, assuming that all the jobs and resources are known in advance which allows to create schedule for all jobs at once [2][3]. CCS [10] as well as GORBA [19] are both advanced resource management systems that use schedule instead of a queue(s) to schedule workflows (GORBA), or sequential and parallel jobs while supporting the advanced reservations (CCS).…”