| Author: | He, Qingqiang |
| Title: | Intra-task interference analysis for real-time scheduling of DAG tasks |
| Advisors: | Guan, Nan (COMP) Lyu, Mingsong (COMP) |
| Degree: | Ph.D. |
| Year: | 2023 |
| Subject: | Real-time data processing System design Electronic data processing Hong Kong Polytechnic University -- Dissertations |
| Department: | Department of Computing |
| Pages: | xvi, 166 pages : color illustrations |
| Language: | English |
| Abstract: | Most parallel applications can be modeled as DAG (directed acyclic graph) tasks. Real-time scheduling and analysis of DAG tasks are of critical importance to both academia and industry. When scheduling DAG tasks, parallel vertices inside a DAG task can interfere with each other, which is called the intra-task interference. This work focuses on analyzing the intra-task interference for the real-time scheduling of DAG tasks. By exploring the internal structure of DAG tasks, we propose new techniques and scheduling approaches to improve the system schedulability. In detail, this thesis presents the following contributions. 1. Intra-Task Priority Assignment. We are the first of introducing the technique of intra-task priority assignment to explore the execution order of eligible vertices within a DAG task. In this work, we show that this intra-task vertex execution order has a large impact on system schedulability and propose to control the execution order by vertex-level priority assignment. We develop analysis techniques to bound the worst-case response time for the proposed scheduling strategy and design heuristics for proper priority assignment to improve system schedulability as much as possible. We also extend the proposed technique to support arbitrary intra-task priority assignment and the conditional DAG task model. 2. The Technique of Long Paths. We are the first of introducing the technique of long paths to bound the response time of DAG tasks. In 1969, Graham developed a well-known response time bound for a DAG task using the total workload and the longest path of the DAG, which has been widely applied to solve many scheduling and analysis problems of DAG-based task systems. This work presents a new response time bound for a DAG task using the total workload and the lengths of multiple long paths of the DAG, instead of the longest path in Graham’s bound. Our new bound theoretically dominates and empirically outperforms Graham’s bound. 3. The Degree of Parallelism. We are the first of introducing the degree of parallelism to analyze the intra-task interference. The degree of parallelism of DAG tasks is an important characterization in scheduling. This work revisits the definition and the computing algorithms for the degree of parallelism of DAG tasks, and clarifies some misunderstandings regarding the degree of parallelism which exist in real-time literature. Based on the degree of the parallelism, we propose a real-time scheduling approach for DAG tasks, which is quite simple but rather effective and outperforms the state-of-the-art by a considerable margin. All the above proposed techniques are extended to the general setting of scheduling multiple DAG tasks. Comprehensive evaluations using randomly generated DAG tasks or realistic parallel benchmarks demonstrate the superiority of the proposed methods over the state-of-the-art. |
| Rights: | All rights reserved |
| Access: | open access |
Copyright Undertaking
As a bona fide Library user, I declare that:
- I will abide by the rules and legal ordinances governing copyright regarding the use of the Database.
- I will use the Database for the purpose of my research or private study only and not for circulation or further reproduction or any other purpose.
- I agree to indemnify and hold the University harmless from and against any loss, damage, cost, liability or expenses arising from copyright infringement or unauthorized usage.
By downloading any item(s) listed above, you acknowledge that you have read and understood the copyright undertaking as stated above, and agree to be bound by all of its terms.
Please use this identifier to cite or link to this item:
https://theses.lib.polyu.edu.hk/handle/200/12527