BROWSE
Detail
Journal Article
AB9: A Neural Processor for Inference Acceleration
- Authors
- Yong Cheol Peter Cho, Jaehoon Chung, Jeongmin Yang, Chun-Gi Lyuh, HyunMi Kim, Chan Kim, Je-seok Ham, Minseok Choi, Kyoungseon Shin, Jinho Han, Youngsu Kwon
- Issue Date
- 2020-08
- Citation
- ETRI Journal, v.42, no.4, pp.491-504
- ISSN
- 1225-6463
- Publisher
- 한국전자통신연구원 (ETRI)
- Language
- English
- Type
- Journal Article
- Abstract
- We present AB9, a neural processor for inference acceleration. AB9 consists of a systolic tensor core (STC) neural network accelerator designed to accelerate artificial intelligence applications by exploiting the data reuse and parallelism characteristics inherent in neural networks while providing fast access to large on-chip memory. Complementing the hardware is an intuitive and user-friendly development environment that includes a simulator and an implementation flow that provides a high degree of programmability with a short development time. Along with a 40-TFLOP STC that includes 32k arithmetic units and over 36혻MB of on-chip SRAM, our baseline implementation of AB9 consists of a 1-GHz quad-core setup with other various industry-standard peripheral intellectual properties. The acceleration performance and power efficiency were evaluated using YOLOv2, and the results show that AB9 has superior performance and power efficiency to that of a general-purpose graphics processing unit implementation. AB9 has been taped out in the TSMC 28-nm process with a chip size of 17혻×혻23 mm2. Delivery is expected later this year.
- KSP Keywords
- 28 nm, Acceleration performance, Development environment, Development time, General-purpose graphics processing unit(GPGPU), High degree, Network Accelerator, Neural processor, On-chip memory, Power Efficiency, User-friendly
This work is distributed under the term of Korea Open Government License (KOGL)
(Type 4: : Type 1 + Commercial Use Prohibition+Change Prohibition)
(Type 4: : Type 1 + Commercial Use Prohibition+Change Prohibition)
