2019 International Conference On Computer Aided Design

The Premier Conference Devoted to Technical Innovations in Electronic Design Automation

November 4-7, 2019The Westin Westminster Westminster, CO

MP Associates, Inc.
MONDAY November 04, 1:45pm - 3:45pm | Westminster IV
EVENT TYPE: REGULAR SESSION
SESSION 2A
Accelerating Neural Networks
Moderator:
Hai Helen Li - Duke Univ.
This session includes four papers on designing neural network accelerators. The first paper proposes a modular DNN accelerator generator with configurable dataflow and an NN-to-architecture mapper to generate energy-efficient designs. The second paper presents an accelerator for the instance normalization layer in generative neural networks. The third paper designs a CNN accelerator framework that can reduce off-chip memory accesses for super-resolution tasks with large feature maps. The last paper develops an FPGA-based accelerator for auto-regressive convolutional neural networks to accelerate audio synthesis.

2A.1MAGNet: A Modular Accelerator Generator for Neural Networks
 Speaker: Rangharajan Venkatesan - NVIDIA Corp.
 Authors: Rangharajan Venkatesan - NVIDIA Corp.
Yakun Sophia Shao - NVIDIA Corp.
Miaorong Wang - Massachusetts Institute of Technology
Jason Clemons - NVIDIA Corp.
Steve Dai - NVIDIA Corp.
Matthew Fojtik - NVIDIA Corp.
Ben Keller - NVIDIA Corp.
Alicia Klinefelter - NVIDIA Corp.
Nathaniel Pinckney - NVIDIA Corp.
Priyanka Raina - Stanford Univ.
Yanqing Zhang - NVIDIA Corp.
Brian Zimmer - NVIDIA Corp.
William J. Dally - NVIDIA Corp.
Joel Emer - NVIDIA Corp.
Stephen W. Keckler - NVIDIA Corp.
Brucek Khailany - NVIDIA Corp.
2A.2ACG-Engine: An Inference Accelerator for Content Generative Neural Networks
 Speaker: Haobo Xu - Institute of Computing Technology, Chinese Academy of Sciences & Univ. of Chinese Academy of Sciences
 Authors: Haobo Xu - Institute of Computing Technology, Chinese Academy of Sciences & Univ. of Chinese Academy of Sciences
Ying Wang - Institute of Computing Technology, Chinese Academy of Sciences
Yujie Wang - Institute of Computing Technology, Chinese Academy of Sciences
Jiajun Li - Institute of Computing Technology, Chinese Academy of Sciences & Univ. of Chinese Academy of Sciences
Bosheng Liu - Institute of Computing Technology, Chinese Academy of Sciences & Univ. of Chinese Academy of Sciences
Yinhe Han - Institute of Computing Technology, Chinese Academy of Sciences
2A.3eSRCNN: A Framework for Optimizing Super-Resolution Tasks on Diverse Embedded CNN Accelerators
 Speaker: Youngbeom Jung - Korea Advanced Institute of Science and Technology
 Authors: Youngbeom Jung - Korea Advanced Institute of Science and Technology
Yeongjae Choi - Korea Advanced Institute of Science and Technology
Jaehyeong Sim - Korea Advanced Institute of Science and Technology
Lee-Sup Kim - Korea Advanced Institute of Science and Technology
2A.4FastWave: Accelerating Autoregressive Convolutional Neural Networks on FPGA
 Speaker: Shehzeen S. Hussain - Univ. of California, San Diego
 Authors: Shehzeen S. Hussain - Univ. of California, San Diego
Mojan Javaheripi - Univ. of California, San Diego
Paarth Neekhara - Univ. of California, San Diego
Ryan Kastner - Univ. of California, San Diego
Farinaz Koushanfar - Univ. of California, San Diego