{"title":"基于改进的乐观成本矩阵的异构系统列表调度算法","authors":"","doi":"10.1016/j.future.2024.107576","DOIUrl":null,"url":null,"abstract":"<div><div>In heterogeneous computing systems, efficient task-scheduling methods are paramount for enhancing computational performance. However, the existing algorithm exhibits certain deficiencies, notably its oversight of load balancing concerns and inadequate emphasis on the out-degree property of tasks. To address these issues, a novel list scheduling algorithm is proposed, Average Earliest Finish Time (AEFT), which proficiently allocates task flows onto heterogeneous processors. The AEFT algorithm primarily consists of two key stages: (1) prioritizing tasks to determine the distribution of task priorities and (2) assigning optimal processors for tasks with given priorities. By leveraging its specific topology, the AEFT algorithm minimizes the scheduling length of task flows. Simultaneously, a prediction mechanism in determining task prioritization and selecting processors stages is proposed to reduce the scheduling time of task flows. In addition, in the processor selection stage, AEFT algorithm considers the out-degree characteristics of tasks, ameliorating situations of processor load imbalance. The AEFT algorithm demonstrates superior performance compared to prior list scheduling algorithms concerning makespan, speedup, and the percentage of occurrences of better solutions, as evidenced by experiments conducted on randomly generated and real-application graphs. Specifically, for <span><math><mi>t</mi></math></span> tasks and <span><math><mi>p</mi></math></span> processors, the AEFT algorithm achieves a time complexity of <span><math><mrow><mi>O</mi><mrow><mo>(</mo><msup><mrow><mi>t</mi></mrow><mrow><mn>2</mn></mrow></msup><mi>p</mi><mo>)</mo></mrow></mrow></math></span>.</div></div>","PeriodicalId":55132,"journal":{"name":"Future Generation Computer Systems-The International Journal of Escience","volume":null,"pages":null},"PeriodicalIF":6.2000,"publicationDate":"2024-10-28","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":"{\"title\":\"Heterogeneous system list scheduling algorithm based on improved optimistic cost matrix\",\"authors\":\"\",\"doi\":\"10.1016/j.future.2024.107576\",\"DOIUrl\":null,\"url\":null,\"abstract\":\"<div><div>In heterogeneous computing systems, efficient task-scheduling methods are paramount for enhancing computational performance. However, the existing algorithm exhibits certain deficiencies, notably its oversight of load balancing concerns and inadequate emphasis on the out-degree property of tasks. To address these issues, a novel list scheduling algorithm is proposed, Average Earliest Finish Time (AEFT), which proficiently allocates task flows onto heterogeneous processors. The AEFT algorithm primarily consists of two key stages: (1) prioritizing tasks to determine the distribution of task priorities and (2) assigning optimal processors for tasks with given priorities. By leveraging its specific topology, the AEFT algorithm minimizes the scheduling length of task flows. Simultaneously, a prediction mechanism in determining task prioritization and selecting processors stages is proposed to reduce the scheduling time of task flows. In addition, in the processor selection stage, AEFT algorithm considers the out-degree characteristics of tasks, ameliorating situations of processor load imbalance. The AEFT algorithm demonstrates superior performance compared to prior list scheduling algorithms concerning makespan, speedup, and the percentage of occurrences of better solutions, as evidenced by experiments conducted on randomly generated and real-application graphs. Specifically, for <span><math><mi>t</mi></math></span> tasks and <span><math><mi>p</mi></math></span> processors, the AEFT algorithm achieves a time complexity of <span><math><mrow><mi>O</mi><mrow><mo>(</mo><msup><mrow><mi>t</mi></mrow><mrow><mn>2</mn></mrow></msup><mi>p</mi><mo>)</mo></mrow></mrow></math></span>.</div></div>\",\"PeriodicalId\":55132,\"journal\":{\"name\":\"Future Generation Computer Systems-The International Journal of Escience\",\"volume\":null,\"pages\":null},\"PeriodicalIF\":6.2000,\"publicationDate\":\"2024-10-28\",\"publicationTypes\":\"Journal Article\",\"fieldsOfStudy\":null,\"isOpenAccess\":false,\"openAccessPdf\":\"\",\"citationCount\":\"0\",\"resultStr\":null,\"platform\":\"Semanticscholar\",\"paperid\":null,\"PeriodicalName\":\"Future Generation Computer Systems-The International Journal of Escience\",\"FirstCategoryId\":\"94\",\"ListUrlMain\":\"https://www.sciencedirect.com/science/article/pii/S0167739X24005405\",\"RegionNum\":2,\"RegionCategory\":\"计算机科学\",\"ArticlePicture\":[],\"TitleCN\":null,\"AbstractTextCN\":null,\"PMCID\":null,\"EPubDate\":\"\",\"PubModel\":\"\",\"JCR\":\"Q1\",\"JCRName\":\"COMPUTER SCIENCE, THEORY & METHODS\",\"Score\":null,\"Total\":0}","platform":"Semanticscholar","paperid":null,"PeriodicalName":"Future Generation Computer Systems-The International Journal of Escience","FirstCategoryId":"94","ListUrlMain":"https://www.sciencedirect.com/science/article/pii/S0167739X24005405","RegionNum":2,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"Q1","JCRName":"COMPUTER SCIENCE, THEORY & METHODS","Score":null,"Total":0}
引用次数: 0
摘要
在异构计算系统中,高效的任务调度方法对提高计算性能至关重要。然而,现有算法存在一些缺陷,特别是忽略了负载平衡问题,对任务的出度属性重视不够。为了解决这些问题,我们提出了一种新型列表调度算法--平均最早完成时间(AEFT),它能将任务流有效地分配到异构处理器上。AEFT 算法主要包括两个关键阶段:(1) 对任务进行优先级排序,以确定任务优先级的分布;(2) 为具有给定优先级的任务分配最佳处理器。通过利用其特定的拓扑结构,AEFT 算法最大限度地减少了任务流的调度长度。同时,在确定任务优先级和选择处理器阶段,提出了一种预测机制,以减少任务流的调度时间。此外,在处理器选择阶段,AEFT 算法考虑了任务的出度特征,改善了处理器负载不平衡的情况。在随机生成的图和实际应用图上进行的实验证明,与之前的列表调度算法相比,AEFT 算法在时间跨度、速度提升和更好解决方案的出现率方面都表现出更优越的性能。具体来说,对于 t 个任务和 p 个处理器,AEFT 算法的时间复杂度为 O(t2p)。
Heterogeneous system list scheduling algorithm based on improved optimistic cost matrix
In heterogeneous computing systems, efficient task-scheduling methods are paramount for enhancing computational performance. However, the existing algorithm exhibits certain deficiencies, notably its oversight of load balancing concerns and inadequate emphasis on the out-degree property of tasks. To address these issues, a novel list scheduling algorithm is proposed, Average Earliest Finish Time (AEFT), which proficiently allocates task flows onto heterogeneous processors. The AEFT algorithm primarily consists of two key stages: (1) prioritizing tasks to determine the distribution of task priorities and (2) assigning optimal processors for tasks with given priorities. By leveraging its specific topology, the AEFT algorithm minimizes the scheduling length of task flows. Simultaneously, a prediction mechanism in determining task prioritization and selecting processors stages is proposed to reduce the scheduling time of task flows. In addition, in the processor selection stage, AEFT algorithm considers the out-degree characteristics of tasks, ameliorating situations of processor load imbalance. The AEFT algorithm demonstrates superior performance compared to prior list scheduling algorithms concerning makespan, speedup, and the percentage of occurrences of better solutions, as evidenced by experiments conducted on randomly generated and real-application graphs. Specifically, for tasks and processors, the AEFT algorithm achieves a time complexity of .
期刊介绍:
Computing infrastructures and systems are constantly evolving, resulting in increasingly complex and collaborative scientific applications. To cope with these advancements, there is a growing need for collaborative tools that can effectively map, control, and execute these applications.
Furthermore, with the explosion of Big Data, there is a requirement for innovative methods and infrastructures to collect, analyze, and derive meaningful insights from the vast amount of data generated. This necessitates the integration of computational and storage capabilities, databases, sensors, and human collaboration.
Future Generation Computer Systems aims to pioneer advancements in distributed systems, collaborative environments, high-performance computing, and Big Data analytics. It strives to stay at the forefront of developments in grids, clouds, and the Internet of Things (IoT) to effectively address the challenges posed by these wide-area, fully distributed sensing and computing systems.