日前，由我中心长聘教轨副教授冷静文（通讯作者）指导的上海交通大学电子信息与电气工程学院2020级博士研究生郭聪，作为第一作者发表题为“Balancing Efficiency and Flexibility for DNN Acceleration via Temporal GPU-Systolic Array Integration”的学术论文，该论文被Design Automation Conference 2020录用。

Balancing Efficiency and Flexibility for DNN Acceleration via Temporal GPU-Systolic Array Integration

The research interest in specialized hardware accel erators for deep neural networks (DNN) spiked recently owing to their superior performance and effificiency. However, today’s DNN accelerators primarily focus on accelerating specifific “kernels” such as convolution and matrix multiplication, which are vital but only part of an end-to-end DNN-enabled application. Meaningful speedups over the entire application often require supporting computations that are, while massively parallel, ill-suited to DNN accelerators. Integrating a general-purpose processor such as a CPU or a GPU incurs signifificant data movement overhead and leads to resource under-utilization on the DNN accelerators.

We propose Simultaneous Multi-mode Architecture (SMA), a novel architecture design and execution model that offers general-purpose programmability on DNN accelerators in order to accelerate end-to-end applications. The key to SMA is the temporal integration of the systolic execution model with the GPU-like SIMD execution model. The SMA exploits the common components shared between the systolic-array accelerator and the GPU, and provides lightweight reconfifiguration capability to switch between the two modes in-situ. The SMA achieves up to 63% performance improvement while consuming 23% less energy than the baseline Volta architecture with TensorCore.

近日，中心长聘教轨副教授张拳石为通讯作者、程煦为第一作者，共同发表的题为“Explaining Knowledge Distillation by Quantifying the Knowledge”的学术论文被CVPR 2020收录。

国际计算机视觉与模式识别会议(CVPR)是IEEE一年一度的学术性会议，会议的主要内容是计算机视觉与模式识别技术。根据官方统计数据，今年投稿6656篇，最终录用1470篇，录用率仅22%！

近日，中心助理教授丁家昕在物联网研究领域取得积极进展，题为“Differentially Private Range Counting in Planar Graphs for Spatial Sensing”的研究成果，被INFOCOM 2020录用。

INFOCOM（IEEE International Conference on Computer Communications）是计算机网络领域三大顶级国际会议之一（CCF认定为A类会议），具有极高的影响力。本次INFOCOM 2020将于2020年4月在中国北京召开，会议投稿总数为1354篇，最终录用268篇，录用率仅为19.8%。