[연구] 이진규 교수, IEEE RTSS에 10년 연속 논문 게재
- 소프트웨어융합대학
- 조회수1252
- 2021-12-13
이진규 교수는 지난 12월 7일~12월 10일에 열린 IEEE RTSS (Real-Time Systems Symposium)에 아래의 논문을 교신저자로 게재 하였습니다.
IEEE RTSS는 매년 30여편의 논문만이 발표되는 실시간 시스템 분야 Top 1 국제학술대회이며, 이진규 교수는 이로써 2012년부터 2021년 까지 10년 연속 IEEE RTSS에 논문을 게재하는 업적을 이루었습니다. (10년간 주저자 10편, 공저자 5편)
http://2021.rtss.org/
[논문정보]
- LaLaRAND: Flexible Layer-by-Layer CPU/GPU Scheduling for Real-Time DNN Task
- Woosung Kang, Kilho Lee, Jinkyu Lee, Insik Shin and Hoon Sung Chwa
- Deep neural networks (DNNs) have shown remarkable success in various machine-learning (ML) tasks useful for many safety-critical, real-time embedded systems. The foremost design goal for enabling DNN execution on real-time embedded systems is to provide worst-case timing guarantees with limited computing resources. Yet, the state-of-the-art ML frameworks hardly leverage heterogeneous computing resources (i.e., CPU, GPU) to improve the schedulability of real-time DNN tasks due to several factors, which include a coarse-grained resource allocation model (one-resource-per-task), the asymmetric nature of DNN execution on CPU and GPU, and lack of schedulabilityaware CPU/GPU allocation scheme. This paper presents, to the best of our knowledge, the first study of addressing the above three major barriers and examining their cooperative effect on schedulability improvement. In this paper, we propose LaLaRAND, a real-time layer-level DNN scheduling framework, that enables flexible CPU/GPU scheduling of individual DNN layers by tightly coupling CPU-friendly quantization with fine-grained CPU/GPU allocation schemes (one-resource-per-layer) while mitigating accuracy loss without compromising timing guarantees. We have implemented and evaluated LaLaRAND on top of the state-of-theart ML framework to demonstrate its effectiveness in making more DNN task sets schedulable by 56% and 80% over an existing approach and a baseline (vanilla PyTorch), respectively, with only up to -0.4% of performance (inference accuracy) difference.