1-D PE 어레이로 컨볼루션 연산을 수행하는 저전력 DCNN 가속기
Power-Efficient DCNN Accelerator Mapping Convolutional Operation with 1-D PE Array
2
조회
0
다운로드
0
인용
0
참고문헌
키워드
초록
In this paper, we propose a novel method of performing convolutional operations on a 2-D Processing Element(PE) array. The conventional method [1] of mapping the convolutional operation using the 2-D PE array lacks flexibility and provides low utilization of PEs. However, by mapping a convolutional operation from a 2-D PE array to a 1-D PE array, the proposed method can increase the number and utilization of active PEs. Consequently, the throughput of the proposed Deep Convolutional Neural Network(DCNN) accelerator can be increased significantly. Furthermore, the power consumption for the transmission of weights between PEs can be saved. Based on the simulation results, the performance of the proposed method provides approximately 4.55%, 13.7%, and 2.27% throughput gains for each of the convolutional layers of AlexNet, VGG16, and ResNet50 using the DCNN accelerator with a (weights size) x (output data size) 2-D PE array compared to the conventional method. Additionally the proposed method provides approximately 63.21%, 52.46%, and 39.23% power savings.
Abstract
In this paper, we propose a novel method of performing convolutional operations on a 2-D Processing Element(PE) array. The conventional method [1] of mapping the convolutional operation using the 2-D PE array lacks flexibility and provides low utilization of PEs. However, by mapping a convolutional operation from a 2-D PE array to a 1-D PE array, the proposed method can increase the number and utilization of active PEs. Consequently, the throughput of the proposed Deep Convolutional Neural Network(DCNN) accelerator can be increased significantly. Furthermore, the power consumption for the transmission of weights between PEs can be saved. Based on the simulation results, the performance of the proposed method provides approximately 4.55%, 13.7%, and 2.27% throughput gains for each of the convolutional layers of AlexNet, VGG16, and ResNet50 using the DCNN accelerator with a (weights size) x (output data size) 2-D PE array compared to the conventional method. Additionally the proposed method provides approximately 63.21%, 52.46%, and 39.23% power savings.
참고문헌 (13)
- Tripathi, Milan, "Analysis of convolutional neural network based image classification techniques", Journal of Innovative Image Processing (JIIP), 3(2), pp.100~117 (2021)
- 김진호, "딥러닝 신경망을 이용한 문자 및 단어 단위의 영문 차량 번호판 인식", (사)디지털산업정보학회 논문지, 16(4), pp.19~28 (2020), (사)디지털산업정보학회
- 김민재, "CNN 기반의 IEEE 802.11 WLAN 프레임 포맷 검출", (사)디지털산업정보학회 논문지, 16(2), pp.27~33 (2020), (사)디지털산업정보학회
- Krizhevsky, "Imagenet classification with deep convolutional neural networks", Advances in neural information processing systems, pp.1097~1105 (2012)
- Simonyan, Karen, "Very deep convolutional networks for large-scale image recognition", arXiv preprint arXiv:1409.1556, ~ (2014)
- He, Kaiming, "Deep residual learning for image recognition", Proceedings of the IEEE conference on computer vision and pattern recognition, pp.770~778 (2016)
- Chen, Y.-H, "A Spatial Architecture for Energy-Efficient Dataflow for Convolutional Neural Networks", ACM SIGARCH computer architecture news, 44(3), pp.367~379 (2016)
- "NVDLA Open Source Project" (2017)[URL]
- N. P. Jouppi, "In-datacenter performance analysis of a tensor processing unit", International Symposium on Computer Architecture (ISCA) (2017)
- A. Parashar, "SCNN: An accelerator for compressed-sparse convolutional neural networks", International Symposium on Computer Architecture (ISCA) (2017)
- Y. Chen, "DaDianNao: A machine-learning supercomputer", International Symposium on Microarchitecture (MICRO) (2014)
- T. Chen, "Architectural Support for Programming Languages and Operating Systems (ASPLOS)" (2014)
- V. Sze, "Synthesis Lectures on Computer Architecture", pp.44~46 (2020), Morgan & Claypool Publishers