Energy‐efficient scheduling algorithms for batch‐of‐tasks (BoT) applications on heterogeneous computing systems

Abstract

One of the major design constraints of a heterogeneous computing system is optimal scheduling, that is, mapping of tasks on the processing nodes in order to optimize the QoS parameters. Because of the huge energy consumption by computing resources, negative environmental effects and reduced system reliability, energy has unavoidably been added as a new parameter to the list of QoS parameters. Energy optimization in scheduling strategies along with makespan makes it an even more challenging combinatorial optimization problem. This work proposes two energy-aware scheduling algorithms G1 and G2 to schedule a batch-oftasks, made of a collection of independent tasks, on heterogeneous processors in order to minimize the makespan and the energy consumption. The proposed algorithms schedule tasks based on weighted aggregation cost function to the appropriate processors followed by task migration phase designed to further minimize the makespan and the energy consumption. The study evaluates the performance of the proposed algorithms with some of the peers, that is, MinMin, MINSuff on account of makespan, energy consumption, flowtime, and utilization. An experimental study reveals that the proposed algorithm (G2) consistently performs better under various test conditions. the reduced reliability of the system. The increased temperature leads to system failure resulting in high monetary cost as well as environmental concerns. Therefore, reducing energy consumption of computing systems has become a must with the appeal to develop energy-efficient computing units, storage, networking, and scheduling models that consume energy and other resources as minimum as possible [4][5][6][7].Scheduling is the indispensable process of resource management in computing paradigms and one of the potential components that can be very much helpful in order to optimize the energy consumption of resources. In general, optimal scheduling has been proven to be NP-complete, and it is intractable to realize an optimal scheduling model in the presence of heterogeneity of jobs as well as resources, limited memory bandwidth, and many other constraints [8]. Job scheduling on heterogeneous parallel processors in order to minimize the makespan as well as energy consumption is one of the classical combinatorial optimization problems [9,10]. The batch-of-tasks (BoT) consists of many independent tasks belonging to different applications and can be run simultaneously to utilize the resources as much as possible. The scheduling of BoT applications on HCS is very common as it has been used by many streams, namely, optimization, cryptographic analysis, high energy physics, engineering, biochemistry, drug discovery, weather forecasting, simulation applications to name a few. The wide applicability of BoT applications on HCS makes it an important area for research, and therefore, a large number of algorithms have been developed for single as well as multi-objective problems. In order to deal with a bi-objective problem, weighted aggregation is one of the widely ac...

Energy‐efficient quantum‐inspired stochastic Q‐HypE algorithm for batch‐of‐stochastic‐tasks on heterogeneous DVFS‐enabled processors

“…Being a bi‐objective problem, energy‐aware scheduling requires determining the solutions with maximum possible optimized makespan and energy consumption . Due to NP‐complete nature, a large number of energy‐aware heuristics and meta‐heuristics have been reported in the literature for different types of applications as well as execution environment …”

“…The proposed algorithms have been compared to MINSuff and MinMin on account of flow‐time, energy makespan, utilization, and makespan. The proposed algorithm (G2) performs better than peers …”

A scheduling selection process for energy‐efficient task execution on DVFS processors

“…In contrast, our approach makes the assumption that each task is executed with a fixed frequency, but different tasks may use different frequencies when executed on different processors. Two energy‐aware scheduling algorithms for independent tasks have been proposed in the work of Sajid et al The algorithms schedule tasks to the processors based on a weighted aggregation cost function, and then, a task migration phase follows. Many other scheduling algorithms for DVFS environments have been proposed in the literature (see other works).…”

Off-Line Time Aware Scheduling of Bag-of-Tasks on Heterogeneous Distributed System

“…Heterogeneous distributed system has emerged as a commonly system for handling large scale scientific and commercial problems in various fields, such as image processing, signal processing, pattern matching in text, and so on [2], [9], [12]- [14]. For the sake of improving the performance of the system, many task scheduling algorithms for heterogeneous or homogeneous distributed system have been proposed in the past decades, e.g., [30], [33], [41], [44]. In order to obtain high performance and fast scheduling, Wang et al [41] proposed a multi-objective bi-level programming model for energy and locality aware multi-job scheduling in heterogeneous system.…”

“…In order to obtain high performance and fast scheduling, Wang et al [41] proposed a multi-objective bi-level programming model for energy and locality aware multi-job scheduling in heterogeneous system. Sajid et al developed two scheduling algorithms to schedule a BoT (bag-of-tasks, BoT) on heterogeneous system so as to minimize the makespan and the energy consumption [33]. A representative algorithm is…”