COMB-MCM: Computing-on-Memory-Boundary NN Processor with Bipolar Bitwise Sparsity Optimization for Scalable Multi-Chiplet-Module Edge Machine Learning

Zhu, Haozhe; Jiao, Bo; Zhang, Jinshan; Jia, Xinru; Wang, Yunzhengmao; Guan, Tianchan; Wang, Shengcheng; Niu, Dimin; Zheng, Hongzhong; Chen, Chixiao; Wang, Mingyu; Zhang, Lihua; Zeng, Xiaoyang; Liu, Qi; Xie, Yuan; Liu, Ming

doi:10.1109/isscc42614.2022.9731657

Cited by 18 publications

(2 citation statements)

References 1 publication

Supporting

Mentioning

Contrasting

Order By: Relevance

“…For example [38] embedded CIM macros into the standard memory space of a CPU. However, the further development of system-level integration starts treating CIM macros as high density processing cores at scale instead of memory banks [39,40,41,42]. Meanwhile, the emergence of digital CIM and existence of near-memory computing (NMC) completely blur the boundary between CIM macros and traditional digital architectures.…”

Section: Micro-architecture-level Cim Macrosmentioning

confidence: 99%

A Survey of Computing-in-Memory Processor: From Circuit to Application

Sun,

Yue,

et al. 2024

IEEE Open J. Solid-State Circuits Soc.

View full text Add to dashboard Cite

The computing-in-memory (CIM) technique is emerging with the evolvement of big data and artificial intelligence (AI) application. The manuscript presents a systematic review of existing CIM works in a bottom-up view from circuit to application. Various types of CIM circuits based on different volatile/non-volatile devices are introduced. The micro CIM architectures are illustrated to support multibit precision computation. After that, several types of processor-level CIM chips are analyzed to reveal the system architecture design considerations. The corresponding CIM tool chains and applications beyond AI applications are also introduced. From circuit to application levels, this manuscript analyzes the design trade-offs, remained challenges and possible future design trends at different design hierarchies of CIM processors.INDEX TERMS Computing-in-memory (CIM), CIM processor, Circuit-level CIM cells, CIM micro architecture, CIM tool chains, CIM applications

show abstract

Section: Micro-architecture-level Cim Macrosmentioning

confidence: 99%

A Survey of Computing-in-Memory Processor: From Circuit to Application

Sun,

Yue,

et al. 2024

IEEE Open J. Solid-State Circuits Soc.

View full text Add to dashboard Cite

show abstract

“…Nevertheless, the embedded digital circuits in the memory array tend to degrade the storage density. In addition, digital CIM works have been widely discussed in SRAM [ 6 , 7 , 8 , 9 , 10 ]. It is still rarely discussed in NVM, such as NOR Flash memory.…”

Section: Introductionmentioning

confidence: 99%

Simulation of a Fully Digital Computing-in-Memory for Non-Volatile Memory for Artificial Intelligence Edge Applications

Feng

Zhou

et al. 2023

Micromachines

View full text Add to dashboard Cite

In recent years, digital computing in memory (CIM) has been an efficient and high-performance solution in artificial intelligence (AI) edge inference. Nevertheless, digital CIM based on non-volatile memory (NVM) is less discussed for the sophisticated intrinsic physical and electrical behavior of non-volatile devices. In this paper, we propose a fully digital non-volatile CIM (DNV-CIM) macro with compressed coding look-up table (LUT) multiplier (CCLUTM) using the 40 nm technology, which is highly compatible with the standard commodity NOR Flash memory. We also provide a continuous accumulation scheme for machine learning applications. When applied to a modified ResNet18 network trained under the CIFAR-10 dataset, the simulations indicate that the proposed CCLUTM-based DNV-CIM can achieve a peak energy efficiency of 75.18 TOPS/W with 4-bit multiplication and accumulation (MAC) operations.

show abstract

Design memristor-based computing-in-memory for AI accelerators considering the interplay between devices, circuits, and system

An,

Wang,

et al. 2023

Sci. China Inf. Sci.

View full text Add to dashboard Cite

COMB-MCM: Computing-on-Memory-Boundary NN Processor with Bipolar Bitwise Sparsity Optimization for Scalable Multi-Chiplet-Module Edge Machine Learning

Cited by 18 publications

References 1 publication

A Survey of Computing-in-Memory Processor: From Circuit to Application

A Survey of Computing-in-Memory Processor: From Circuit to Application

Simulation of a Fully Digital Computing-in-Memory for Non-Volatile Memory for Artificial Intelligence Edge Applications

Design memristor-based computing-in-memory for AI accelerators considering the interplay between devices, circuits, and system

Contact Info

Product

Resources

About