检索结果-南通市图书馆

Analysis and modeling of resistive switching mechanisms oriented to resistive random-access memory

Chinese Physics B 2013年第3期22卷 522-527页

作者：黄达吴俊杰唐玉华 State Key Laboratory of High Performance Computing National University of Defense Technology School of Computer National University of Defense Technology

With the progress of the semiconductor industry,the resistive random-access memory（RAM） has drawn increasing *** discovery of the memristor has brought much attention to this *** has focused on the resistive switching characteristics of different materials and the analysis of resistive switching *** discuss the resistive switching mechanisms of different materials in this paper and analyze the differences of those mechanisms from the view point of circuitry to establish their respective circuit ***,simulations are *** give the prospect of using different materials in resistive RAM on account of their resistive switching mechanisms,which are applied to explain their resistive switchings.

关键词： resistive random-access memory resistive switching mechanism circuit model

来源：

维普期刊数据库

同方期刊数据库评论

在线全文

学校读者我要写书评

暂无评论

Architecture-circuit-technology co-optimization for resistive random access memory-based computation-in-memory chips

引用

Science China(Information Sciences) 2023年第10期66卷 152-161页

作者： Yuyi LIU Bin GAO Jianshi TANG Huaqiang WU He QIAN School of Integrated Circuits Beijing National Research Center for Information Science and Technology(BNRist) Tsinghua University

Computation-in-memory(CIM) chips offer an energy-efficient approach to artificial intelligence computing workloads. resistive random-access memory(RRAM)-based CIM chips have proven to be a promising solution for overcoming the von Neumann bottleneck. In this paper, we review our recent studies on the architecture-circuit-technology co-optimization of scalable CIM chips and related hardware demonstrations. To further minimize data movements between memory and computing units, architecture optimization methods have been introduced. Then, we propose a device-architecture-algorithm co-design simulator to provide guidelines for designing CIM systems. A physics-based compact RRAM model and an array-level analog computing model were embedded in the simulator. In addition, a CIM compiler was proposed to optimize the on-chip dataflow. Finally, research perspectives are proposed for future development.

关键词： resistive random-access memory computation-in-memory compact model device-architecture-algorithm co-design compiler

来源：

同方期刊数据库评论

在线全文

同方期刊数据库

学校读者我要写书评

暂无评论

CMN: a co-designed neural architecture search for efficient computing-in-memory-based mixture-of-experts

引用

Science China(Information Sciences) 2024年第10期67卷 100-114页

作者： Shihao HAN Sishuo LIU Shucheng DU Mingzi LI Zijian YE Xiaoxin XU Yi LI Zhongrui WANG Dashan SHANG Department of Electrical and Electronic Engineering The University of Hong Kong access-AI Chip Center for Emerging Smart Systems InnoHK Centers Key Lab of Fabrication Technologies for Integrated Circuits Chinese Academy of Sciences Laboratory of Microelectronics Devices and Integrated Technology Institute of MicroelectronicsChinese Academy of Sciences University of Chinese Academy of Sciences School of Microelectronics Southern University of Science and Technology

Artificial intelligence(AI) has experienced substantial advancements recently, notably with the advent of large-scale language models(LLMs) employing mixture-of-experts(MoE) techniques, exhibiting human-like cognitive skills. As a promising hardware solution for edge MoE implementations, the computing-in-memory(CIM) architecture collocates memory and computing within a single device, significantly reducing the data movement and the associated energy consumption. However, due to diverse edge application scenarios and constraints, determining the optimal network structures for MoE, such as the expert's location,quantity, and dimension on CIM systems remains elusive. To this end, we introduce a software-hardware co-designed neural architecture search(NAS) framework, CIM-based MoE NAS(CMN), focusing on identifying a high-performing MoE structure under specific hardware constraints. The results of the NYUD-v2 dataset segmentation on the RRAM(SRAM) CIM system reveal that CMN can discover optimized MoE configurations under energy, latency, and performance constraints, achieving 29.67×(43.10×) energy savings,175.44×(109.89×) speedup, and 12.24× smaller model size compared to the baseline MoE-enabled Visual Transformer, respectively. This co-design opens up an avenue toward high-performance MoE deployments in edge CIM systems.

关键词： mixture-of-experts computing-in-memory neural architecture search resistive random-access memory static random-access memory

来源：

同方期刊数据库

同方期刊数据库评论

在线全文

学校读者我要写书评

暂无评论

Organic-2D composite material-based RRAM with high reliability for mimicking synaptic behavior

引用

Journal of Materiomics 2024年第2期10卷 440-447页

作者： Tangyou Sun Fantao Yu Xiaosheng Tang Haiou Li Fabi Zhang Zhimou Xu Qing Liao Zhiqiang Yu Xingpeng Liu Peihua Wangyang Hezhang Li Ying Peng Guangxi Key Laboratory of Precision Navigation Technology and Application Guilin University of Electronic TechnologyGuilin541004China College of Optoelectronic Engineering Chongqing UniversityChongqing400044China School of Optical and Electronic Information Huazhong University of Science and TechnologyWuhan430074China School of Electrical and Information Engineering Guangxi University of Science and TechnologyLiuzhou545006China Department of Precision Instrument Tsinghua UniversityBeijing100084China

The field of artificial intelligence and neural computing has been rapidly expanding due to the imple-mentation of resistive random-access memory(RRAM)based artificial ***,the low flexibility of conventional RRAM materials hinders their ability to mimic synaptic behavior *** overcome such limitation,organic-2D composites with high mechanical properties are proposed as the active layer of ***,we enhance the reliability of the device by ZrO_(2)insertion layer,resulting in stable synaptic *** Ag/PVA:h-BN/ZrO_(2)/ITO devices show stable bipolar resistive switching behavior with an ON/OFF ratio of over 5×10^(2),a~2400 cycles endurance and a long retention time(>6×10^(3)s),which are essential for the development of high-performance *** also study the possible synaptic mechanism and dynamic plasticity of the memory device,observing the transition from short-term potentiation(STP)to long-term potentiation(LTP)under the effect of continuous voltage ***,the device exhibits both long-term depression(LTD)and paired-pulse facili-tation(PPF)properties,which have significant implications for the design of organic-2D composite material RRAMs that aim to mimic biological synapses,representing promising avenues for the devel-opment of advanced neuromorphic computing systems.

关键词： resistive random-access memory Synaptic mimicking Organic-2D composite material Neuromorphic computing

来源：

维普期刊数据库评论

在线全文

维普期刊数据库

学校读者我要写书评

暂无评论

PRAP-PIM:A weight pattern reusing aware pruning method for ReRAM-based PIM DNN accelerators

引用

High-Confidence Computing 2023年第2期3卷 50-59页

作者： Zhaoyan Shen Jinhao Wu Xikun Jiang Yuhao Zhang Lei Ju Zhiping Jia School of Computer Science and Technology Shandong UniversityQindao 266237China School of Cyber Science and Technology Shandong UniversityQindao 266237China Department of Computer Science and Technology Tsinghua UniversityBeijing 100084China

resistive random-access memory(ReRAM)based Processing-in-memory(PIM)frameworks are proposed to accelerate the working process of DNN models by eliminating the data movement between the computing and memory *** further mitigate the space and energy consumption,DNN model weight sparsity and weight pattern repetition are exploited to optimize these ReRAM-based ***,most of these works only focus on one aspect of this software/hardware codesign framework and optimize them individually,which makes the design far from *** this paper,we propose PRAP-PIM,which jointly exploits the weight sparsity and weight pattern repetition by using a weight pattern reusing aware pruning *** relaxing the weight pattern reusing precondition,we propose a similarity-based weight pattern reusing method that can achieve a higher weight pattern reusing *** results show that PRAP-PIM achieves 1.64×performance improvement and 1.51×energy efficiency improvement in popular deep learning benchmarks,compared with the state-of-the-art ReRAM-based DNN accelerators.

关键词： resistive random-access memory Processing-in-memory Deep neural network Model compression

来源：

维普期刊数据库评论

在线全文

维普期刊数据库

学校读者我要写书评

暂无评论

欢迎您,

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

在线全文

在线全文

在线全文

在线全文

在线全文

请选择保存的检索档案：

请选择收藏分类：

通借通还

欢迎您,

建议与咨询 留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

在线全文

在线全文

在线全文

在线全文

在线全文

请选择保存的检索档案： 新增检索档案 确定 取消

请选择收藏分类： 新增自定义分类 确定 取消

通借通还

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

请选择保存的检索档案：

请选择收藏分类：