RLSchert: An HPC Job Scheduler Using Deep Reinforcement Learning and Remaining Time Prediction
The job scheduler plays a vital role in high-performance computing platforms. It determines the execution order of the jobs and the allocation of resources, which in turn affect the resource utilization of the entire system. As the scale and complexity of HPC continue to grow, job scheduling is beco...
| Main Authors: | , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
MDPI AG
2021-10-01
|
| Series: | Applied Sciences |
| Subjects: | |
| Online Access: | https://www.mdpi.com/2076-3417/11/20/9448 |
| _version_ | 1797515463117242368 |
|---|---|
| author | Qiqi Wang Hongjie Zhang Cheng Qu Yu Shen Xiaohui Liu Jing Li |
| author_facet | Qiqi Wang Hongjie Zhang Cheng Qu Yu Shen Xiaohui Liu Jing Li |
| author_sort | Qiqi Wang |
| collection | DOAJ |
| description | The job scheduler plays a vital role in high-performance computing platforms. It determines the execution order of the jobs and the allocation of resources, which in turn affect the resource utilization of the entire system. As the scale and complexity of HPC continue to grow, job scheduling is becoming increasingly important and difficult. Existing studies relied on user-specified or regression techniques to give fixed runtime prediction values and used the values in static heuristic scheduling algorithms. However, these approaches require very accurate runtime predictions to produce better results, and fixed heuristic scheduling strategies cannot adapt to changes in the workload. In this work, we propose RLSchert, a job scheduler based on deep reinforcement learning and remaining runtime prediction. Firstly, RLSchert estimates the state of the system by using a dynamic job remaining runtime predictor, thereby providing an accurate spatiotemporal view of the cluster status. Secondly, RLSchert learns the optimal policy to select or kill jobs according to the status through imitation learning and the proximal policy optimization algorithm. Extensive experiments on real-world job logs at the USTC Supercomputing Center showed that RLSchert is superior to static heuristic policies and outperforms the learning-based scheduler DeepRM. In addition, the dynamic predictor gives a more accurate remaining runtime prediction result, which is essential for most learning-based schedulers. |
| first_indexed | 2024-03-10T06:45:50Z |
| format | Article |
| id | doaj.art-ee4f9ee1b3e944d487842a21c634c9e6 |
| institution | Directory Open Access Journal |
| issn | 2076-3417 |
| language | English |
| last_indexed | 2024-03-10T06:45:50Z |
| publishDate | 2021-10-01 |
| publisher | MDPI AG |
| record_format | Article |
| series | Applied Sciences |
| spelling | doaj.art-ee4f9ee1b3e944d487842a21c634c9e62023-11-22T17:18:47ZengMDPI AGApplied Sciences2076-34172021-10-011120944810.3390/app11209448RLSchert: An HPC Job Scheduler Using Deep Reinforcement Learning and Remaining Time PredictionQiqi Wang0Hongjie Zhang1Cheng Qu2Yu Shen3Xiaohui Liu4Jing Li5School of Computer Science and Technology, University of Science and Technology of China, Hefei 230026, ChinaSchool of Computer Science and Technology, University of Science and Technology of China, Hefei 230026, ChinaSchool of Computer Science and Technology, University of Science and Technology of China, Hefei 230026, ChinaSupercomputing Center, University of Science and Technology of China, Hefei 230026, ChinaSupercomputing Center, University of Science and Technology of China, Hefei 230026, ChinaSchool of Computer Science and Technology, University of Science and Technology of China, Hefei 230026, ChinaThe job scheduler plays a vital role in high-performance computing platforms. It determines the execution order of the jobs and the allocation of resources, which in turn affect the resource utilization of the entire system. As the scale and complexity of HPC continue to grow, job scheduling is becoming increasingly important and difficult. Existing studies relied on user-specified or regression techniques to give fixed runtime prediction values and used the values in static heuristic scheduling algorithms. However, these approaches require very accurate runtime predictions to produce better results, and fixed heuristic scheduling strategies cannot adapt to changes in the workload. In this work, we propose RLSchert, a job scheduler based on deep reinforcement learning and remaining runtime prediction. Firstly, RLSchert estimates the state of the system by using a dynamic job remaining runtime predictor, thereby providing an accurate spatiotemporal view of the cluster status. Secondly, RLSchert learns the optimal policy to select or kill jobs according to the status through imitation learning and the proximal policy optimization algorithm. Extensive experiments on real-world job logs at the USTC Supercomputing Center showed that RLSchert is superior to static heuristic policies and outperforms the learning-based scheduler DeepRM. In addition, the dynamic predictor gives a more accurate remaining runtime prediction result, which is essential for most learning-based schedulers.https://www.mdpi.com/2076-3417/11/20/9448high-performance computingRLSchertschedulingdeep reinforcement learningremaining runtime prediction |
| spellingShingle | Qiqi Wang Hongjie Zhang Cheng Qu Yu Shen Xiaohui Liu Jing Li RLSchert: An HPC Job Scheduler Using Deep Reinforcement Learning and Remaining Time Prediction Applied Sciences high-performance computing RLSchert scheduling deep reinforcement learning remaining runtime prediction |
| title | RLSchert: An HPC Job Scheduler Using Deep Reinforcement Learning and Remaining Time Prediction |
| title_full | RLSchert: An HPC Job Scheduler Using Deep Reinforcement Learning and Remaining Time Prediction |
| title_fullStr | RLSchert: An HPC Job Scheduler Using Deep Reinforcement Learning and Remaining Time Prediction |
| title_full_unstemmed | RLSchert: An HPC Job Scheduler Using Deep Reinforcement Learning and Remaining Time Prediction |
| title_short | RLSchert: An HPC Job Scheduler Using Deep Reinforcement Learning and Remaining Time Prediction |
| title_sort | rlschert an hpc job scheduler using deep reinforcement learning and remaining time prediction |
| topic | high-performance computing RLSchert scheduling deep reinforcement learning remaining runtime prediction |
| url | https://www.mdpi.com/2076-3417/11/20/9448 |
| work_keys_str_mv | AT qiqiwang rlschertanhpcjobschedulerusingdeepreinforcementlearningandremainingtimeprediction AT hongjiezhang rlschertanhpcjobschedulerusingdeepreinforcementlearningandremainingtimeprediction AT chengqu rlschertanhpcjobschedulerusingdeepreinforcementlearningandremainingtimeprediction AT yushen rlschertanhpcjobschedulerusingdeepreinforcementlearningandremainingtimeprediction AT xiaohuiliu rlschertanhpcjobschedulerusingdeepreinforcementlearningandremainingtimeprediction AT jingli rlschertanhpcjobschedulerusingdeepreinforcementlearningandremainingtimeprediction |