A performance comparison of processor allocation and job scheduling algorithms for mesh-connected multiprocessors

Abstract

Due to its simplicity, regularity and suitability for VLSI implementation, the mesh topology for multiprocessors has drawn considerable attention. Several processor allocation strategies for mesh-connected multiprocessors have been proposed in recent years. In this paper; we present the results of a perform" study of all the proposed strategies known to authors. Originally each of these allocation strategies was proposed for use with First-Come-FirstServed job scheduling. In this paper we also propose and evaluate new variants of these strategies using the Scan scheduling discipline. We find Scan to signijicantly improve the performance of all the allocation strategies. A wide range of workloads and system sizes are considered. In addition, we compare the overheads of these algorithms and study the effects of overheads on pe~ormance.

A Novel Approach to Comparative Analysis and Performance Evaluation of Energy Aware Multi-Layered Job Scheduling in Heterogeneous Cloud

“…Multi-level Queue Scheduling Algorithm: In this type of procedure; the initialized queue is demarcated into n number of queues; with different tasks in multi level scheduling [15]. The whole exercise involves the use of a simulator for assembling scheduling tasks and strategies.…”

A compendium of processor allocation strategies for two‐dimensional mesh connected systems

“…This scheme is most appropriate for large, sparse, 2-D meshes where processor utilization requirement is low due to the nature of the task requests. For algorithmic details see [ 11, and see [2] for an enhancement to the method. In Section 3 we present a completely new and much more competitive buddy strategy.…”

“…For the submesh allocations shown in Figure 3 and a request for a 3 x 4 submesh, there are two candidate submesh bases: < 2,3 > and < 3,3 >. With bV (2,3,4,6)=8 and bv (3,3,5,6)=9, the HBV will allocate submesh (3,3,5,6).…”

The Effects of Precedence and Priority Constraints on the Performance of Scan Scheduling for Hypercube Multiprocessors

“…Scan scheduling achieves this performance advantage by carefully ordering job executions in such a way that the hypercube is efficiently ''packed,'' reducing external fragmentation. Though this paper focuses on hypercube systems, Scan has been shown to have similar advantages for mesh-connected multiprocessors [1].An important weakness in these results is that, while they are based on the assumption that jobs have no scheduling constraints, such constraints are not uncommon in practice. As a result, a job scheduler may not have complete freedom to set job execution order, so it may have limited ability to affect performance.…”