A layer-wise frequency scaling for a neural processing unit
DC Field | Value | Language |
---|---|---|
dc.contributor.author | Chung, Jaehoon | - |
dc.contributor.author | Kim, HyunMi | - |
dc.contributor.author | Shin, Kyoungseon | - |
dc.contributor.author | Lyuh, Chun-Gi | - |
dc.contributor.author | Cho, Yong Cheol Peter | - |
dc.contributor.author | Han, Jinho | - |
dc.contributor.author | Kwon, Youngsu | - |
dc.contributor.author | Gong, Young-Ho | - |
dc.contributor.author | Chung, Sung Woo | - |
dc.date.accessioned | 2023-03-22T02:40:03Z | - |
dc.date.available | 2023-03-22T02:40:03Z | - |
dc.date.created | 2023-03-22 | - |
dc.date.issued | 2022-10 | - |
dc.identifier.issn | 1225-6463 | - |
dc.identifier.uri | http://scholarworks.bwise.kr/ssu/handle/2018.sw.ssu/43430 | - |
dc.description.abstract | Dynamic voltage frequency scaling (DVFS) has been widely adopted for run-time power management of various processing units. In the case of neural processing units (NPUs), power management of neural network applications is required to adjust the frequency and voltage every layer to consider the power behavior and performance of each layer. Unfortunately, DVFS is inappropriate for layer-wise run-time power management of NPUs due to the long latency of voltage scaling compared with each layer execution time. Because the frequency scaling is fast enough to keep up with each layer, we propose a layer-wise dynamic frequency scaling (DFS) technique for an NPU. Our proposed DFS exploits the highest frequency under the power limit of an NPU for each layer. To determine the highest allowable frequency, we build a power model to predict the power consumption of an NPU based on a real measurement on the fabricated NPU. Our evaluation results show that our proposed DFS improves frame per second (FPS) by 33% and saves energy by 14% on average, compared with DVFS. | - |
dc.language | 영어 | - |
dc.language.iso | en | - |
dc.publisher | WILEY | - |
dc.relation.isPartOf | ETRI JOURNAL | - |
dc.title | A layer-wise frequency scaling for a neural processing unit | - |
dc.type | Article | - |
dc.identifier.doi | 10.4218/etrij.2022-0094 | - |
dc.type.rims | ART | - |
dc.identifier.bibliographicCitation | ETRI JOURNAL, v.44, no.5, pp.849 - 858 | - |
dc.identifier.kciid | ART002888433 | - |
dc.description.journalClass | 1 | - |
dc.identifier.wosid | 000852984500001 | - |
dc.identifier.scopusid | 2-s2.0-85137937352 | - |
dc.citation.endPage | 858 | - |
dc.citation.number | 5 | - |
dc.citation.startPage | 849 | - |
dc.citation.title | ETRI JOURNAL | - |
dc.citation.volume | 44 | - |
dc.contributor.affiliatedAuthor | Gong, Young-Ho | - |
dc.identifier.url | https://onlinelibrary.wiley.com/doi/10.4218/etrij.2022-0094 | - |
dc.type.docType | Article | - |
dc.description.isOpenAccess | Y | - |
dc.subject.keywordAuthor | dynamic frequency scaling | - |
dc.subject.keywordAuthor | neural processing unit | - |
dc.subject.keywordAuthor | power model | - |
dc.relation.journalResearchArea | Engineering | - |
dc.relation.journalResearchArea | Telecommunications | - |
dc.relation.journalWebOfScienceCategory | Engineering, Electrical & Electronic | - |
dc.relation.journalWebOfScienceCategory | Telecommunications | - |
dc.description.journalRegisteredClass | scie | - |
dc.description.journalRegisteredClass | scopus | - |
dc.description.journalRegisteredClass | kci | - |
Items in ScholarWorks are protected by copyright, with all rights reserved, unless otherwise indicated.
Soongsil University Library 369 Sangdo-Ro, Dongjak-Gu, Seoul, Korea (06978)02-820-0733
COPYRIGHT ⓒ SOONGSIL UNIVERSITY, ALL RIGHTS RESERVED.
Certain data included herein are derived from the © Web of Science of Clarivate Analytics. All rights reserved.
You may not copy or re-distribute this material in whole or in part without the prior written consent of Clarivate Analytics.