DeLoc: A Locality and Memory-Congestion-Aware Task Mapping Method for Modern NUMA Systems
The mapping of tasks to processor cores, called task mapping, is crucial to achieving scalable performance on multicore processors. On modern NUMA (non-uniform memory access) systems, the memory congestion problem could degrade the performance more severely than the data locality problem because hea...
| Main Authors: | , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
IEEE
2020-01-01
|
| Series: | IEEE Access |
| Subjects: | |
| Online Access: | https://ieeexplore.ieee.org/document/8949493/ |
| _version_ | 1819170008101552128 |
|---|---|
| author | Mulya Agung Muhammad Alfian Amrizal Ryusuke Egawa Hiroyuki Takizawa |
| author_facet | Mulya Agung Muhammad Alfian Amrizal Ryusuke Egawa Hiroyuki Takizawa |
| author_sort | Mulya Agung |
| collection | DOAJ |
| description | The mapping of tasks to processor cores, called task mapping, is crucial to achieving scalable performance on multicore processors. On modern NUMA (non-uniform memory access) systems, the memory congestion problem could degrade the performance more severely than the data locality problem because heavy congestion on shared caches and memory controllers could cause long latencies. Conventional work on task mapping mostly focuses on improving the locality of memory accesses. However, our previous work showed that on modern NUMA systems, maximizing the locality can degrade the performance due to memory congestion. In this work, we propose a task mapping method that addresses the locality and the memory congestion problems to improve the performance of parallel applications. In the proposed method, first, the spatial and temporal communication behaviors of the applications are analyzed from the time-series dataset of communications among the parallel tasks. Then, a data clustering technique is employed to detect groups of tasks that potentially cause the memory congestion. Finally, this information is used to compute the task mapping to improve the locality and reduce the memory congestion. We also provide a set of metrics to describe the communication behaviors and to evaluate if the target application can benefit from our method. The proposed method is evaluated with the NPB and PARSEC applications on a real NUMA system and a multicore simulator. A detailed analysis of the sources of performance gain is also provided. Experimental results show that our method can achieve up to a 61% performance improvement compared with the state-of-the-art locality-based method. |
| first_indexed | 2024-12-22T19:28:33Z |
| format | Article |
| id | doaj.art-e2eb797b73d74a889c1900589bdc08c8 |
| institution | Directory Open Access Journal |
| issn | 2169-3536 |
| language | English |
| last_indexed | 2024-12-22T19:28:33Z |
| publishDate | 2020-01-01 |
| publisher | IEEE |
| record_format | Article |
| series | IEEE Access |
| spelling | doaj.art-e2eb797b73d74a889c1900589bdc08c82022-12-21T18:15:10ZengIEEEIEEE Access2169-35362020-01-0186937695310.1109/ACCESS.2019.29637268949493DeLoc: A Locality and Memory-Congestion-Aware Task Mapping Method for Modern NUMA SystemsMulya Agung0https://orcid.org/0000-0001-9521-2177Muhammad Alfian Amrizal1https://orcid.org/0000-0003-1124-5137Ryusuke Egawa2https://orcid.org/0000-0001-8966-867XHiroyuki Takizawa3https://orcid.org/0000-0003-2858-3140Graduate School of Information Sciences, Tohoku University, Sendai, JapanResearch Institute of Electrical Communication, Tohoku University, Sendai, JapanCyberscience Center, Tohoku University, Sendai, JapanCyberscience Center, Tohoku University, Sendai, JapanThe mapping of tasks to processor cores, called task mapping, is crucial to achieving scalable performance on multicore processors. On modern NUMA (non-uniform memory access) systems, the memory congestion problem could degrade the performance more severely than the data locality problem because heavy congestion on shared caches and memory controllers could cause long latencies. Conventional work on task mapping mostly focuses on improving the locality of memory accesses. However, our previous work showed that on modern NUMA systems, maximizing the locality can degrade the performance due to memory congestion. In this work, we propose a task mapping method that addresses the locality and the memory congestion problems to improve the performance of parallel applications. In the proposed method, first, the spatial and temporal communication behaviors of the applications are analyzed from the time-series dataset of communications among the parallel tasks. Then, a data clustering technique is employed to detect groups of tasks that potentially cause the memory congestion. Finally, this information is used to compute the task mapping to improve the locality and reduce the memory congestion. We also provide a set of metrics to describe the communication behaviors and to evaluate if the target application can benefit from our method. The proposed method is evaluated with the NPB and PARSEC applications on a real NUMA system and a multicore simulator. A detailed analysis of the sources of performance gain is also provided. Experimental results show that our method can achieve up to a 61% performance improvement compared with the state-of-the-art locality-based method.https://ieeexplore.ieee.org/document/8949493/High-performance computinglocalitymemory congestionNUMAprocess mappingtask mapping |
| spellingShingle | Mulya Agung Muhammad Alfian Amrizal Ryusuke Egawa Hiroyuki Takizawa DeLoc: A Locality and Memory-Congestion-Aware Task Mapping Method for Modern NUMA Systems IEEE Access High-performance computing locality memory congestion NUMA process mapping task mapping |
| title | DeLoc: A Locality and Memory-Congestion-Aware Task Mapping Method for Modern NUMA Systems |
| title_full | DeLoc: A Locality and Memory-Congestion-Aware Task Mapping Method for Modern NUMA Systems |
| title_fullStr | DeLoc: A Locality and Memory-Congestion-Aware Task Mapping Method for Modern NUMA Systems |
| title_full_unstemmed | DeLoc: A Locality and Memory-Congestion-Aware Task Mapping Method for Modern NUMA Systems |
| title_short | DeLoc: A Locality and Memory-Congestion-Aware Task Mapping Method for Modern NUMA Systems |
| title_sort | deloc a locality and memory congestion aware task mapping method for modern numa systems |
| topic | High-performance computing locality memory congestion NUMA process mapping task mapping |
| url | https://ieeexplore.ieee.org/document/8949493/ |
| work_keys_str_mv | AT mulyaagung delocalocalityandmemorycongestionawaretaskmappingmethodformodernnumasystems AT muhammadalfianamrizal delocalocalityandmemorycongestionawaretaskmappingmethodformodernnumasystems AT ryusukeegawa delocalocalityandmemorycongestionawaretaskmappingmethodformodernnumasystems AT hiroyukitakizawa delocalocalityandmemorycongestionawaretaskmappingmethodformodernnumasystems |