Stella C. S. Porto

Static and Dynamic Processor Scheduling Disciplines in Heterogeneous Parallel Architectures

Most parallel jobs cannot be fully parallelized. In a homogeneous parallel machine-one in which all processors are identical-the serial fraction of the computation has to be executed at the speed of any of the identical processors, limiting the speedup that can be obtained due to parallelism. In a heterogeneous architecture, the sequential bottleneck can be greatly reduced by running the sequential part of the job or even the critical tasks in a faster processor. This paper uses Markov chain based models to analyze the performance of static and dynamic processor assignment policies for heterogeneous architectures. Parallel jobs are assumed to be described by acyclic directed task graphs. A new static processor assignment policy, called Largest Task First Minimum Finish Time (LTFMFT), is introduced. The analysis shows that this policy is very sensitive to the degree of heterogeneity of the architecture, and that it outperforms all other policies analyzed. Three dynamic assignment disciplines are compared and it is shown that, in heterogeneous environments, the disciplines that perform better are those that consider the structure of the task graph, and not only the service demands of the individual tasks. The performance of heterogeneous architectures is compared with cost-equivalent homogeneous ones taking into account different scheduling policies. Finally, static and dynamic processor assignment disciplines are compared in terms of performance.

A TABU SEARCH APPROACH TO TASK SCHEDULING ON HETEROGENEOUS PROCESSORS UNDER PRECEDENCE CONSTRAINTS

Parallel programs may be represented as a set of interrelated sequential tasks. When multiprocessors are used to execute such programs, the parallel portion of the application can be speeded up by an appropriate allocation of processors to the tasks of the application. Given a parallel application defined by a task precedence graph, the goal of task scheduling (or processor assignment) is thus the minimization of the makespan of the application. In a heterogeneous multiprocessor system, task scheduling consists of determining which tasks will be assigned to each processor, as well as the execution order of the tasks assigned to each processor. In this work, we apply the tabu search metaheuristic to the solution of the task scheduling problem on a heterogeneous multiprocessor environment under precedence constraints. The topology of the Mean Value Analysis solution package for product form queueing networks is used as the framework for performance evaluation. We show that tabu search obtains much better results, i.e., shorter completion times, improving from 20 to 30% the makespan obtained by the most appropriate algorithm previously published in the literature.

Parallel tabu search message-passing synchronous strategies for task scheduling under precedence constraints

This paper presents parallelization strategies for a tabu search algorithm for the task scheduling problem on heterogeneous processors under task precedence constraints. Parallelization relies exclusively on the decompostion of the solution space exploration. Four different parallel strategies are proposed and implemented on an asynchronous parallel machine under PVM: the master-slave model, with two different schemes for improved load balancing, and the single-program-multiple-data model, with single-token and multiple-token message passing schemes. The comparative analysis of these strategies shows that the tabu search approach for this problem is very suitable to the parallelization of the neighborhood search, with efficiency results almost always close to one for problems over a certain size.

Performance evaluation of a parallel tabu search task scheduling algorithm

Abstract This paper presents the solution quality analysis of a parallel tabu search algorithm for the task scheduling problem on heterogeneous processors under precedence constraints. We evaluate the achieved makespan reduction of different parallel applications, relatively to the results obtained by the best greedy algorithm in the literature, as a function of parameters such as problem size, system heterogeneity, and number of processors. Our results show that the parallel tabu search algorithm leads to much better solutions than the greedy algorithm in many cases where the latter is not capable of profiting from the inherent application parallelism and system heterogeneity.

Processor assignment in heterogeneous parallel architectures

It has been already demonstrated that cost-effective multiprocessor designs may be obtained by combining in the same architecture processors of different speeds (heterogeneous architecture) so that the serial and critical portions of the application may benefit from a fast single processor. The paper presents a systematic way to build static heuristic scheduling algorithms for such environments. Several algorithms are proposed and their performances are compared through simulation. One of the proposed algorithms is shown to achieve substantial performance gains as the degree of heterogeneity of the architecture increases.<<ETX>>

STATIC HEURISTIC PROCESSOR ASSIGNMENT IN HETEROGENEOUS MULTIPROCESSORS

It has been already demonstrated that cost-effective multiprocessor designs may be obtained by combining in the same architecture processors of different speeds (heterogeneous architecture) so that the serial and critical portions of the application may benefit from a fast single processor. In such an environment, the problem of assigning tasks to processors becomes a very important one. This papers presents a systematic way to build static heuristic scheduling algorithms. Using this strategy, several algorithms are proposed and their performance are compared through simulation. One of the proposed algorithms is shown to achieve substantial performance gains as the degree of heterogeneity of the architecture increases.

An object-oriented approach to a parallel tabu search algorithm for the task scheduling problem

This work presents a parallel object-oriented tabu search (TS) algorithm for static task scheduling. The scheduling problem and the TS method are separately modeled under an object-oriented approach. The TS parallelization follows a strategy based on multi-search threads and the algorithm is fully implemented using the Java language. Besides providing a new scheduling algorithm, this work contributes to demonstrate: the strength of object-orientation also in this field of applications; the adaptability of TS to asynchronous parallelization; the significance of diversification in TS algorithms; and the potential of the Java language in implementing highly portable object-oriented parallel software.

Processor assignment in heterogeneous message passing parallel architectures

The authors propose new scheduling algorithms for loosely coupled message passing heterogeneous multiprocessors. These algorithms are extensions to previous work on scheduling in heterogeneous environments by D.A. Menasce and V. Almeida (1992). It is assumed that parallel jobs are structured as task graphs and that tasks communicate with each other by exchanging messages at synchronization points. A Markov chain based analyzer was built to obtain parallel application execution times for each of the algorithms, and their performances are compared.<<ETX>>

Selected Algorithmic Techniques for Parallel Optimization

The use of parallel algorithms for solving computationally hard problems becomes attractive as parallel systems, consisting of a collection of powerful processors, offer large computing power and memory storage capacity. Even though parallelism will not be able to overdue the assumed worst case exponential time or memory complexity of those problems (unless an exponential number of processors is used) [11], the average execution time of heuristic search algorithms which find good suboptimal solutions for many hard problems is polynomial. Consequently, parallel systems, possibly with hundreds or thousands of processors, give us the perspective of efficiently solving relatively large instances of hard problems.

Using Synthetic Workloads for Parallel Task Scheduling Improvement Analysis

This paper presents an experimental validation of makespan improvements of two scheduling algorithms: a greedy construction algorithm and a tabu search based algorithm. Synthetic parallel executions were performed using the scheduled graph costs. These synthetic executions were performed on a real parallel machine (IBM SP). The estimated and observed response times improvements are very similar, representing the low impact of system overhead on makespan improvement estimation. This guarantees a reliable cost function for static scheduling algorithms and confirms the actual better results of the tabu search meta-heuristic applied to scheduling problems.