{"page":"353-367","day":"02","project":[{"name":"Elastic Coordination for Scalable Machine Learning","_id":"268A44D6-B435-11E9-9278-68D0E5697425","grant_number":"805223","call_identifier":"H2020"}],"publication_identifier":{"isbn":["9781450392044"]},"language":[{"iso":"eng"}],"date_created":"2022-04-17T22:01:46Z","title":"Multi-queues can be state-of-the-art priority schedulers","article_processing_charge":"No","department":[{"_id":"DaAl"}],"ec_funded":1,"abstract":[{"lang":"eng","text":"Designing and implementing efficient parallel priority schedulers is an active research area. An intriguing proposed design is the Multi-Queue: given n threads and m ≥ n distinct priority queues, task insertions are performed uniformly at random, while, to delete, a thread picks two queues uniformly at random, and removes the observed task of higher priority. This approach scales well, and has probabilistic rank guarantees: roughly, the rank of each task removed, relative to remaining tasks in all other queues, is O (m) in expectation. Yet, the performance of this pattern is below that of well-engineered schedulers, which eschew theoretical guarantees for practical efficiency.\r\n\r\nWe investigate whether it is possible to design and implement a Multi-Queue-based task scheduler that is both highly-efficient and has analytical guarantees. We propose a new variant called the Stealing Multi-Queue (SMQ), a cache-efficient variant of the Multi-Queue, which leverages both queue affinity---each thread has a local queue, from which tasks are usually removed; but, with some probability, threads also attempt to steal higher-priority tasks from the other queues---and task batching, that is, the processing of several tasks in a single insert / remove step. These ideas are well-known for task scheduling without priorities; our theoretical contribution is showing that, despite relaxations, this design can still provide rank guarantees, which in turn implies bounds on total work performed. We provide a general SMQ implementation which can surpass state-of-the-art schedulers such as OBIM and PMOD in terms of performance on popular graph-processing benchmarks. Notably, the performance improvement comes mainly from the superior rank guarantees provided by our scheduler, confirming that analytically-reasoned approaches can still provide performance improvements for priority task scheduling."}],"oa":1,"date_published":"2022-04-02T00:00:00Z","oa_version":"Preprint","conference":{"name":"PPoPP: Sympopsium on Principles and Practice of Parallel Programming","location":"Seoul, Republic of Korea","end_date":"2022-04-06","start_date":"2022-04-02"},"type":"conference","publisher":"Association for Computing Machinery","_id":"11180","year":"2022","doi":"10.1145/3503221.3508432","quality_controlled":"1","author":[{"full_name":"Postnikova, Anastasiia","first_name":"Anastasiia","last_name":"Postnikova"},{"id":"2F4DB10C-F248-11E8-B48F-1D18A9856A87","last_name":"Koval","first_name":"Nikita","full_name":"Koval, Nikita"},{"full_name":"Nadiradze, Giorgi","id":"3279A00C-F248-11E8-B48F-1D18A9856A87","last_name":"Nadiradze","first_name":"Giorgi"},{"full_name":"Alistarh, Dan-Adrian","last_name":"Alistarh","id":"4A899BFC-F248-11E8-B48F-1D18A9856A87","orcid":"0000-0003-3650-940X","first_name":"Dan-Adrian"}],"external_id":{"arxiv":["2109.00657"],"isi":["000883318200025"]},"date_updated":"2023-08-03T06:48:35Z","user_id":"4359f0d1-fa6c-11eb-b949-802e58b17ae8","related_material":{"record":[{"status":"public","relation":"research_data","id":"13076"}]},"status":"public","acknowledgement":"We would like to thank the anonymous reviewers for their useful comments. This project has received funding from the European Research Council (ERC) under the European Union’s Horizon 2020 research and innovation programme (grant agreement No 805223 ScaleML).","isi":1,"publication_status":"published","scopus_import":"1","main_file_link":[{"open_access":"1","url":" https://doi.org/10.48550/arXiv.2109.00657"}],"citation":{"ista":"Postnikova A, Koval N, Nadiradze G, Alistarh D-A. 2022. Multi-queues can be state-of-the-art priority schedulers. Proceedings of the 27th ACM SIGPLAN Symposium on Principles and Practice of Parallel Programming. PPoPP: Sympopsium on Principles and Practice of Parallel Programming, 353–367.","mla":"Postnikova, Anastasiia, et al. “Multi-Queues Can Be State-of-the-Art Priority Schedulers.” <i>Proceedings of the 27th ACM SIGPLAN Symposium on Principles and Practice of Parallel Programming</i>, Association for Computing Machinery, 2022, pp. 353–67, doi:<a href=\"https://doi.org/10.1145/3503221.3508432\">10.1145/3503221.3508432</a>.","ama":"Postnikova A, Koval N, Nadiradze G, Alistarh D-A. Multi-queues can be state-of-the-art priority schedulers. In: <i>Proceedings of the 27th ACM SIGPLAN Symposium on Principles and Practice of Parallel Programming</i>. Association for Computing Machinery; 2022:353-367. doi:<a href=\"https://doi.org/10.1145/3503221.3508432\">10.1145/3503221.3508432</a>","chicago":"Postnikova, Anastasiia, Nikita Koval, Giorgi Nadiradze, and Dan-Adrian Alistarh. “Multi-Queues Can Be State-of-the-Art Priority Schedulers.” In <i>Proceedings of the 27th ACM SIGPLAN Symposium on Principles and Practice of Parallel Programming</i>, 353–67. Association for Computing Machinery, 2022. <a href=\"https://doi.org/10.1145/3503221.3508432\">https://doi.org/10.1145/3503221.3508432</a>.","apa":"Postnikova, A., Koval, N., Nadiradze, G., &#38; Alistarh, D.-A. (2022). Multi-queues can be state-of-the-art priority schedulers. In <i>Proceedings of the 27th ACM SIGPLAN Symposium on Principles and Practice of Parallel Programming</i> (pp. 353–367). Seoul, Republic of Korea: Association for Computing Machinery. <a href=\"https://doi.org/10.1145/3503221.3508432\">https://doi.org/10.1145/3503221.3508432</a>","ieee":"A. Postnikova, N. Koval, G. Nadiradze, and D.-A. Alistarh, “Multi-queues can be state-of-the-art priority schedulers,” in <i>Proceedings of the 27th ACM SIGPLAN Symposium on Principles and Practice of Parallel Programming</i>, Seoul, Republic of Korea, 2022, pp. 353–367.","short":"A. Postnikova, N. Koval, G. Nadiradze, D.-A. Alistarh, in:, Proceedings of the 27th ACM SIGPLAN Symposium on Principles and Practice of Parallel Programming, Association for Computing Machinery, 2022, pp. 353–367."},"publication":"Proceedings of the 27th ACM SIGPLAN Symposium on Principles and Practice of Parallel Programming","month":"04"}