Colonnade: A Reconfigurable SRAM-Based Digital Bit-Serial Compute-In-Memory Macro for Processing Neural Networks

计算机科学 计算机硬件 静态随机存取存储器 宏 架空(工程) 块(置换群论) 栏(排版) 并行计算 嵌入式系统 数学 程序设计语言 电信 几何学 帧(网络) 操作系统
作者
Hyunjoon Kim,Taegeun Yoo,Tony Tae-Hyoung Kim,Bongjin Kim
出处
期刊:IEEE Journal of Solid-state Circuits [Institute of Electrical and Electronics Engineers]
卷期号:56 (7): 2221-2233 被引量:77
标识
DOI:10.1109/jssc.2021.3061508
摘要

This article (Colonnade) presents a fully digital bit-serial compute-in-memory (CIM) macro. The digital CIM macro is designed for processing neural networks with reconfigurable 1-16 bit input and weight precisions based on bit-serial computing architecture and a novel all-digital bitcell structure. A column of bitcells forms a column MAC and used for computing a multiply-and-accumulate (MAC) operation. The column MACs placed in a row work as a single neuron and computes a dot-product, which is an essential building block of neural network accelerators. Several key features differentiate the proposed Colonnade architecture from the existing analog and digital implementations. First, its full-digital circuit implementation is free from process variation, noise susceptibility, and data-conversion overhead that are prevalent in prior analog CIM macros. A bitwise MAC operation in a bitcell is performed in the digital domain using a custom-designed XNOR gate and a full-adder. Second, the proposed CIM macro is fully reconfigurable in both weight and input precision from 1 to 16 bit. So far, most of the analog macros were used for processing quantized neural networks with very low input/weight precisions, mainly due to a memory density issue. Recent digital accelerators have implemented reconfigurable precisions, but they are inferior in energy efficiency due to significant off-chip memory access. We present a regular digital bitcell array that is readily reconfigured to a 1-16 bit weight-stationary bit-serial CIM macro. The macro computes parallel dot-product operations between the weights stored in memory and inputs that are serialized from LSB to MSB. Finally, the bit-serial computing scheme significantly reduces the area overhead while sacrificing latency due to bit-by-bit operation cycles. Based on the benefits of digital CIM, reconfigurability, and bit-serial computing architecture, the Colonnade can achieve both high performance and energy efficiency (i.e., both benefits of prior analog and digital accelerators) for processing neural networks. A test-chip with 128 ×128 SRAM-based bitcells for digital bit-serial computing is implemented using 65-nm technology and tested with 1-16 bit weight/input precisions. The measured energy efficiency is 117.3 TOPS/W at 1 bit and 2.06 TOPS/W at 16 bit.

科研通智能强力驱动
Strongly Powered by AbleSci AI
科研通是完全免费的文献互助平台,具备全网最快的应助速度,最高的求助完成率。 对每一个文献求助,科研通都将尽心尽力,给求助人一个满意的交代。
实时播报
修勾想要干饭咧完成签到,获得积分10
刚刚
粗犷的海冬完成签到,获得积分10
刚刚
满三江完成签到,获得积分10
刚刚
orixero的应助被背书强采纳,获得10
1秒前
天真小甜瓜完成签到,获得积分10
1秒前
Elixir完成签到,获得积分10
1秒前
小田完成签到,获得积分10
1秒前
1秒前
温暖的若之完成签到,获得积分10
1秒前
1秒前
杨帆完成签到,获得积分10
1秒前
zain完成签到,获得积分10
2秒前
Lan_Hao完成签到,获得积分20
2秒前
skyhappy发布了新的文献求助10
2秒前
2秒前
自由的白开水完成签到,获得积分10
2秒前
3秒前
弥里发布了新的文献求助10
3秒前
xxx完成签到,获得积分10
3秒前
七里香发布了新的文献求助10
3秒前
yangyangyang完成签到,获得积分10
3秒前
Shuaiyang完成签到 ,获得积分10
4秒前
4秒前
4秒前
C13完成签到,获得积分10
4秒前
wind发布了新的文献求助20
5秒前
5秒前
酷波er的应助被slx采纳,获得10
5秒前
羞涩的文轩完成签到,获得积分10
6秒前
6秒前
Eleanor完成签到,获得积分10
6秒前
青阳完成签到,获得积分0
6秒前
7秒前
万能图书馆的应助被dangniuma采纳,获得10
7秒前
7秒前
7秒前
完美世界的应助被吴小根采纳,获得10
7秒前
啊姚爱学习完成签到,获得积分10
7秒前
SCI的芷蝶发布了新的文献求助10
8秒前
Glileo完成签到,获得积分10
8秒前
高分求助中
(应助此贴封号)【重要!!请各用户(尤其是新用户)详细阅读】【科研通的精品贴汇总】 10000
拉伸高分子熔体中缠结行为的模拟和小角中子散射研究 1700
The World Atlas of Wine 9th edition 1000
奈特基础组织学:配套组织病理学(第4版)(Netter's Essential Histology With Correlated Histopathology) 800
John McDowells Bildungstheorie 800
八旗制度の研究 600
plantTFDB 508
热门求助领域 (近24小时)
化学 材料科学 医学 生物 纳米技术 计算机科学 工程类 有机化学 化学工程 生物化学 物理 复合材料 内科学 催化作用 光电子学 人工智能 细胞生物学 无机化学 物理化学 数学
热门帖子
关注 科研通微信公众号,转发送积分 8167280
求助须知:如何正确求助?哪些是违规求助? 9560076
关于积分的说明 21754886
捐赠科研通 7758257
什么是DOI,文献DOI怎么找? 3429309
关于科研通互助平台的介绍 2553653
邀请新用户注册赠送积分活动 2474136