Audio and Image Cross-Modal Intelligence via a 10TOPS/W 22nm SoC with Back-Propagation and Dynamic Power Gating - Citegraph

Paper Info

Title
Audio and Image Cross-Modal Intelligence via a 10TOPS/W 22nm SoC with Back-Propagation and Dynamic Power Gating

Abstract
We present an ultra-low-power multimedia signal processor (MMSP) SoC that integrates a versatile deep neural network (DNN) engine with audio and image signal processing accelerators for cross-modal IoT intelligence. The proposed MMSP features 2MB MRAM to store all DNN weights on-chip with an energy-efficient dataflow using an MRAM-cache and dynamic power gating. The SoC achieves up to 3-10 TOPS/W peak energy efficiency and consumes only 0.25-3.84 mW. Being the first to demonstrate CNN, GAN, and back-propagation (BP) on a single accelerator SoC for cross-modal fusion, it outperforms state-of-the-art DNN processors by 1.4 - 4.5× in energy efficiency.

Year	DOI	Venue
2022	10.1109/VLSITechnologyandCir46769.2022.9830226	2022 IEEE Symposium on VLSI Technology and Circuits (VLSI Technology and Circuits)
DocType	ISSN	ISBN
Conference	0743-1562	978-1-6654-9773-2
Citations	PageRank	References
0	0.34	0
Authors
17

Authors (17 rows)

Cited by (0 rows)

References (0 rows)

Name	Order	Citations	PageRank
Zichen Fan	1	0	0.34
Hyochan An	2	0	0.34
Qirui Zhang	3	0	0.34
Boxun Xu	4	0	0.34
Li Xu	5	0	0.34
Chien-Wei Tseng	6	0	0.34
Yimai Peng	7	0	0.34
Ang Cao	8	0	0.34
Bowen Liu	9	0	0.34
Changwoo Lee	10	0	0.34
Zhehong Wang	11	0	0.34
Fanghao Liu	12	0	0.34
Guanru Wang	13	0	0.34
Shenghao Jiang	14	0	0.34
Hun-Seok Kim	15	0	1.01
David Blaauw	16	0	0.68
Dennis Sylvester	17	0	0.34

1