DC Field | Value | Language |
---|---|---|
dc.contributor.author | PARKNAEBEOM | - |
dc.contributor.author | 김율화 | - |
dc.contributor.author | 안대현 | - |
dc.contributor.author | 김태수 | - |
dc.contributor.author | KIM, JAE JOON | - |
dc.date.accessioned | 2021-06-01T08:03:42Z | - |
dc.date.available | 2021-06-01T08:03:42Z | - |
dc.date.created | 2021-03-11 | - |
dc.date.issued | 2020-08-11 | - |
dc.identifier.uri | https://oasis.postech.ac.kr/handle/2014.oak/106068 | - |
dc.description.abstract | In Long Short-Term Memory (LSTM) neural network models, a weight matrix tends to be repeatedly loaded from DRAM if the size of on-chip storage of the processor is not large enough to store the entire matrix. To alleviate heavy overhead of DRAM access for weight loading in LSTM computations, we propose a weight reuse scheme which utilizes the weight sharing characteristics in two adjacent time-step computations. Experimental results show that the proposed weight reuse scheme reduces the energy consumption by 28.4-57.3% and increases the overall throughput by 110.8% compared to the conventional schemes. | - |
dc.language | English | - |
dc.publisher | Association for Computing Machinery | - |
dc.relation.isPartOf | 2020 ACM/IEEE International Symposium on Low Power Electronics and Design, ISLPED 2020 | - |
dc.relation.isPartOf | ACM International Conference Proceeding Series | - |
dc.title | Time-step interleaved weight reuse for LSTM neural network computing | - |
dc.type | Conference | - |
dc.type.rims | CONF | - |
dc.identifier.bibliographicCitation | 2020 ACM/IEEE International Symposium on Low Power Electronics and Design, ISLPED 2020 | - |
dc.citation.conferenceDate | 2020-08-10 | - |
dc.citation.conferencePlace | US | - |
dc.citation.title | 2020 ACM/IEEE International Symposium on Low Power Electronics and Design, ISLPED 2020 | - |
dc.contributor.affiliatedAuthor | PARKNAEBEOM | - |
dc.contributor.affiliatedAuthor | 김율화 | - |
dc.contributor.affiliatedAuthor | 안대현 | - |
dc.contributor.affiliatedAuthor | 김태수 | - |
dc.contributor.affiliatedAuthor | KIM, JAE JOON | - |
dc.identifier.scopusid | 2-s2.0-85098269636 | - |
dc.description.journalClass | 1 | - |
dc.description.journalClass | 1 | - |
Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.
library@postech.ac.kr Tel: 054-279-2548
Copyrights © by 2017 Pohang University of Science ad Technology All right reserved.