상세정보
학술대회
A Novel Shared Memory Framework for Distributed Deep Learning in High-Performance Computing Architecture
- 저자
- 안신영, 김중헌, 강성원
- 발행일
- 201805
- 출처
- International Conference on Software Engineering (ICSE) 2018, pp.191-192
- 협약과제
- 18HS2700, 대규모 딥러닝 고속 처리를 위한 HPC 시스템 개발, 최완
- 초록
- This paper proposes a novel virtual shared memory framework, Soft Memory Box (SMB), which directly shares the memory of remote nodes among distributed processes to improve communication performance/speed via deep learning parameter sharing.
- KSP 제안 키워드
- Communication performance, Distributed processes, High Performance Computing, Remote Node, Virtual Shared Memory, deep learning(DL), parameter sharing