TIMAQ: A Time-Domain Computing-in-Memory-Based Processor Using Predictable Decomposed Convolution for Arbitrary Quantized DNNs

Jianxun Yang; Yuyao Kong; Zhao ZhangZhuangzhi LiuJing ZhouYiqi WangYonggang LiuChenfu GuoTe HuCongcong LiLeibo LiuJin ZhangShaojun WeiJun YangShouyi Yin

首页> 外文期刊>IEEE Journal of Solid-State Circuits >TIMAQ: A Time-Domain Computing-in-Memory-Based Processor Using Predictable Decomposed Convolution for Arbitrary Quantized DNNs

【24h】

TIMAQ: A Time-Domain Computing-in-Memory-Based Processor Using Predictable Decomposed Convolution for Arbitrary Quantized DNNs

机译：TIMAQ: A Time-Domain Computing-in-Memory-Based Processor Using Predictable Decomposed Convolution for Arbitrary Quantized DNNs

获取原文

获取原文并翻译 | 示例

掌桥外文数据库（机构版） >>

开具论文收录证明 >>

文献代查 >>

页面导航

摘要
著录项
相关主题

摘要

Energy-efficient processors are crucial for accelerating deep neural networks (DNNs) on edge devices with limited battery capacity. To reduce energy consumption, time-domain computing-in-memory (TD-CIM) is a splendid architecture, which consumes low computation and memory access energy due to low toggle rate of time-based signals and less data movements, respectively. When deploying DNNs in TD-CIMs, quantization is required, which has two types: uniform quantization (UQ) and nonuniform quantization (NUQ). To reach the same accuracy for one DNN, NUQ achieves smaller model size than UQ. Due to varying weight distributions across layers, mixed-precision quantization can further reduce model size, without degrading accuracy. However, previous TD-CIMs are inefficient for mixed-precision NUQ-DNNs due to their adopted bit-serial convolution increasing computation amount significantly. To address that, we propose a unique-weight convolution to accelerate mixed-precision NUQ-DNNs by a special kernel decomposition, reducing computation count remarkably. Based on that, we design a TD-CIM-based processor, TIMAQ, with three architectural techniques: 1) bit-cross-flipping-based kernel decomposer to reduce memory accesses and operations of decomposing kernels; 2) dual-mode-complementary predictor to remove redundant computations; and 3) activation-weight-adaptive pulse quantizer to decrease pulse quantization energy and error. Fabricated in 28-nm CMOS technology and tested on 1–8-b NUQ-DNNs, TIMAQ achieves 2.4–152.7-TOPS/W peak energy efficiency.

著录项

来源
《IEEE Journal of Solid-State Circuits》 |2021年第10期|3021-3038|共18页
作者
Jianxun Yang; Yuyao Kong; Zhao ZhangZhuangzhi LiuJing ZhouYiqi WangYonggang LiuChenfu GuoTe HuCongcong LiLeibo LiuJin ZhangShaojun WeiJun YangShouyi Yin;
展开▼
作者单位

School of Integrated Circuits, Tsinghua University, Beijing, China;

School of Electronic Science and Engineering, Southeast University, Nanjing, China;

Ingenic Semiconductor Company Ltd., Beijing, China;

展开▼
收录信息美国《科学引文索引》(SCI);美国《工程索引》(EI);
原文格式 PDF
正文语种英语
中图分类基本电子电路;
关键词
Indexes; Kernel; Convolution; Quantization (signal); Time-domain analysis; Memory management; Delays;

TIMAQ: A Time-Domain Computing-in-Memory-Based Processor Using Predictable Decomposed Convolution for Arbitrary Quantized DNNs

摘要

著录项

相关主题

期刊订阅