检索结果-内蒙古大学图书馆

SSRN 2024年

作者： Zou, Xiaofeng Peng, Yuanxi Li, Tuo Kong, Lingjun Zhang, Lu Institute for Quantum Information State Key Laboratory of High Performance Computing College of Computer National University of Defense Technology Changsha41003 China Cloud Computing Equipment Industry Innovation Center Jinan250101 China

The Kyber algorithm is one of the standardization of Post Quantum Cryptography project proposed by NIST to withstand quantum attacks. In order to increase throughput and reduce the execution time limited by the high computational complexity of Kyber algorithm, a RISC-V based processor Seesaw is designed to accelerate Kyber algorithm. The 32 specialized extension instructions are firstly designed to enhance the parallel computing ability of the processor and accelerate all processes of the Kyber algorithm by analyzing its characteristic in depth. Then, by carefully designing hardware such as vector registers and algorithm execution units on the RISC-V processor, support of microarchitecture for extension instructions were achieved. Seesaw supports 4096 bits vector calculation through its vector registers and execution unit to meet high-throughput requirements, and is implemented on the FPGA. Besides, we modify the compiler simultaneously to adapt to the instruction extension and execution of Seesaw. Experimental results show that the processor achieves a speed-up of 432× and 18864× for hash and NTT compared to that without extension instructions, respectively, and a speed-up of 5.6× for execution of Kyber algorithm compared with the advanced hardware design. © 2024, The Authors. All rights reserved.

关键词： Vectors

来源：评论

学校读者我要写书评

暂无评论

Understand Data Preprocessing for Effective End-to-End Training of Deep Neural Networks

arXiv

引用

arXiv 2023年

作者： Gong, Ping Ma, Yuxin Li, Cheng Ma, Xiaosong Noh, Sam H. University of Science and Technology of China China Qatar Computing Research Institute Qatar Virginia Tech United States Anhui Province Key Laboratory of High Performance Computing China

In this paper, we primarily focus on understanding the data preprocessing pipeline for DNN Training in the public cloud. First, we run experiments to test the performance implications of the two major data preprocessing methods using either raw data or record files. The preliminary results show that data preprocessing is a clear bottleneck, even with the most efficient software and hardware configuration enabled by NVIDIA DALI, a high-optimized data preprocessing library. Second, we identify the potential causes, exercise a variety of optimization methods, and present their pros and cons. We hope this work will shed light on the new co-design of "data storage, loading pipeline" and "training framework" and flexible resource configurations between them so that the resources can be fully exploited and performance can be maximized. © 2023, CC BY.

关键词： Pipelines

来源：评论

学校读者我要写书评

暂无评论

Memory Access Optimization of high-Order CFD Stencil Computations on GPU 1

引用

21st International Conference on Parallel and Distributed computing, Applications, and Technologies, PDCAT 2020

作者： Wang, Shengxiang Li, Zhuoqian Che, Yonggang Institute for Quantum Information and State Key Laboratory of High Performance Computing College of Computer National University of Defense Technology Changsha China

ISBN: (数字)9783030692445

ISBN: (纸本)9783030692438

Stencils computations are a class of computations commonly found in scientific and engineering applications. They have relatively lower arithmetic intensity. Therefore, their performance is greatly affected by memory access. This paper studies the issue of memory access optimization for the key stencil computations of a high-order CFD program on the NVidia GPU. Two methods are used to optimize the performance. First, we use registers to cache the data used by the stencil computations in the kernel. We use the CUDA warp shuffle functions to exchange data between neighboring grid points, and adjust the thread computation granularity to increase the data reuse. Second, we use the shared memory to buffer the grid data used by the stencil computations in the kernel, and utilize loop tiling to reduce redundant accesses to the global memory. performance evaluation is done on an NVidia Tesla K80 GPU. The results show that compared to the original implementation that only uses the global memory, the optimized implementation that utilizes the registers achieves a maximum speedup of 2.59 and 2.79 relatively for 15M and 60M grids, and the optimized implementation that utilizes the shared memory achieves a maximum speedup of 3.51 and 3.36 relatively for 15M and 60M grids. © 2021, Springer Nature Switzerland AG.

关键词： Graphics processing unit

来源：评论

学校读者我要写书评

暂无评论

high Order Asymptotic Preserving Finite Difference Weno Schemes with Constrained Transport for Mhd Equations in All Sonic Mach Numbers

SSRN

引用

SSRN 2023年

作者： Chen, Wei Wu, Kailiang Xiong, Tao School of Mathematical Sciences Xiamen University Fujian Provincial Key Laboratory of Mathematical Modeling and High-Performance Scientific Computing Fujian Xiamen361005 China Guangdong Shenzhen518055 China

In this paper, a high-order semi-implicit (SI) asymptotic preserving (AP) and divergencefree finite difference weighted essentially nonoscillatory (WENO) scheme is proposed for magnetohydrodynamic (MHD) equations. We consider the sonic Mach number Ε ranging from 0 to O(1). high-order accuracy in time is obtained by SI implicit-explicit Runge–Kutta (IMEX-RK) time discretization. high-order accuracy in space is achieved by finite difference WENO schemes with characteristic-wise reconstructions. A constrained transport method is applied to maintain a discrete divergence-free condition. We formally prove that the scheme is AP. Asymptotic accuracy (AA) in the incompressible MHD limit is obtained if the implicit part of the SI IMEX-RK scheme is stiffly accurate. Numerical experiments are provided to validate the AP, AA, and divergence-free properties of our proposed approach. Besides, the scheme can well capture discontinuities such as shocks in an essentially nonoscillatory fashion in the compressible regime, while it is also a good incompressible solver with uniform large-time step conditions in the low sonic Mach limit. © 2023, The Authors. All rights reserved.

关键词： Mach number

来源：评论

学校读者我要写书评

暂无评论

Wave-particle duality relation with a quantum N-path beamsplitter

引用

Chinese Physics B 2021年第5期30卷 220-228页

作者： Dong-Yang Wang Jun-Jie Wu Yi-Zhi Wang Yong Liu An-Qi Huang Chun-Lin Yu Xue-Jun Yang Institute for Quantum Information&State Key Laboratory of High Performance Computing College of Computer Science and TechnologyNational University of Defense TechnologyChangsha 410073China College of Information and Communication National University of Defense TechnologyXi'an 710006China China Greatwall Quantum Laboratory China Greatwall Technology Group CO.LTD.Changsha 410073China

The wave-particle duality relation derived by Englert sets an upper bound of the extractable information from wave and particle properties in a two-path ***,previous studies demonstrated that the introduction of a quantum beamsplitter in the interferometer could break the limitation of this upper bound,due to interference between wave and particle *** the other line,a lot of efforts have been made to generalize this relation from the two-path setup to the N-path ***,it is an interesting question that whether a quantum N-path beamsplitter can break the limitation as *** paper systemically studies the model of a quantum N-path beamsplitter,and finds that the generalized wave-particle duality relation between interference visibility and path distinguishability is also broken in certain *** further study the maximal extractable information's reliance on the interference between wave and particle properties,and derive a quantitative *** then propose an experimental methodology to verify the break of the *** work reflects the effect of quantum superposition on wave-particle duality,and exhibits a new aspect of the relation between visibility and path distinguishability in N-path ***,it implies the observer's influence on wave-particle duality.

关键词： wave-particle duality interference visibility path distinguishability quantum N-path beamsplitter

来源：评论

学校读者我要写书评

暂无评论

Emerging theory and phenomena in thermal conduction: A selective review

引用

Science China(Physics,Mechanics & Astronomy) 2022年第11期65卷 2-17页

作者： Jie Chen Jia He Dongkai Pan Xiaotian Wang Nuo Yang Jiaojiao Zhu Shengyuan A.Yang Gang Zhang Center for Phononics and Thermal Energy Science China-EU Joint Lab for NanophononicsMOE Key Laboratory of Advanced Microstructured MaterialsSchool of Physics Science and EngineeringTongji UniversityShanghai 200092China School of Energy and Power Engineering Huazhong University of Science and Technology(HUST)Wuhan 430074China School of Physical Science and Technology Southwest UniversityChongqing 400715China Research Laboratory for Quantum Materials Singapore University of Technology and DesignSingapore 487372Singapore Institute of High Performance Computing Agency for ScienceTechnology and Research(A*STAR)Singapore 138632Singapore

Recently, emerging phonon phenomena have been discovered and rapidly developed, which have become an active hot research topic. In this review article, we present state-of-the-art advances in several fascinating phonon transport phenomena. First, we summarize the recent progress on the wave nature of phonons, including phonon coherence and its effects on thermal conductivity and the topological properties of phonons. Then, we discuss the particle nature of phonons, including the weak coupling of phonons and the high-order phonon anharmonicity. Finally, we present the summary and a brief outlook. This review presents the advanced understanding of some emerging phonon phenomena in solid materials, which provides new opportunities for further advancement in a wide variety of applications.

关键词： phonon transport coherence topological phonon high-order phonon interaction

来源：评论

学校读者我要写书评

暂无评论

Radio Modulation Recognition Based on Attention-based Convolutional Neural Network and Time-frequency Analysis 2

Radio Modulation Recognition Based on Attention-based Convol...

引用

2nd International Seminar on Artificial Intelligence, Networking and Information technology, AINIT 2021

作者： Li, Xinlong Huang, Da Yang, Wenjing Li, Xueqiong Institute for Quantum Information and State Key Laboratory of High Performance Computing College of Computer Science and Technology National University of Defense Technology Changsha China

ISBN: (纸本)9781665412964

Radio modulation recognition is the key link of modern electronic warfare. This paper applies the idea of deep learning to radio modulation recognition. Since the modulation type is the most important information about the subtle features of the radio signal, we use short time Fourier transform (STFT) to convert the radio signal into a time-frequency distribution map to form a joint distribution of frequency and time. The attention-based convolutional neural network is used to classify and recognize the modulation types of time-frequency distribution maps. The simulation results show that time-frequency analysis can improve the accuracy of recognition, and adding attention can effectively improve the performance of the network. © 2021 IEEE.

关键词： Convolutional neural networks

来源：评论

学校读者我要写书评

暂无评论

Synergistic enhancement of mechanical and tribological properties in WC-reinforced 316L stainless steel matrix composite fabricated by laser powder bed fusion

引用

Journal of Materials Research and technology 2025年 37卷 617-628页

作者： Jianye Liu Zhicong Zhao Dongyun Rao Yang Yang Lijuan Zheng Chengyong Wang Pu Deng Liuhui Niu Zhenghua Huang Guangdong Provincial Engineering and Technology Research Center for Metal 3D Printing Guangdong Hanbang 3D Technology Co. Ltd. Zhongshan 528427 China State Key Laboratory for High Performance Tools School of Electromechanical Engineering Guangdong University of Technology Guangzhou 510006 China

Austenitic 316L stainless steel, renowned for its excellent toughness and corrosion resistance, is widely used in aerospace, precision molds, and automotive industries. However, its limited strength at both room and high temperatures, coupled with poor tribological properties, restricts broader applications. To address these limitations, this study has fabricated a novel stainless steel matrix composite, 3 %WC/316L, with excellent strength-toughness, heat-resistance, and tribological properties, leveraging the inherent advantages of laser powder bed fusion technology through synergistic strengthening mechanisms: fine grain strengthening, particle reinforcement, and precipitation hardening. Following aging treatment at 800 °C for 2 h, the composite comprises a columnar grain structure with uniformly dispersed partially-melted spherical WC particles (∼1 vol%). A well-bonded diffusion zone with a width of 1∼2 μm forms between the unmelted WC cores and the matrix. Crucially, abundant submicron-scale M 23 C 6 carbides (100–400 nm) precipitate along grain boundaries and within the columnar grains adjacent to the particles. The composite demonstrates superior mechanical properties: at room temperature, the ultimate tensile strength (UTS), yield strength (YS), elongation to fracture (EL), and impact toughness (α kv ) can reach 802 MPa, 554 MPa, 30.5 %, and 55.5 J/cm 2 , respectively; at 400 °C, UTS, YS, and EL still remain at 616 MPa, 370 MPa, and 20.8 %, respectively. Furthermore, the tribological properties are significantly enhanced: at room-temperature, hardness reaches 18.9 HRC, with the average coefficient of friction (COF) and volumetric wear rate ( W r ) reduced to 0.6212 and 50.3 × 10 −6 mm 3 /(N·m), respectively; at 400 °C, COF and W r further decrease to 0.5291 and 43.6 × 10 −6 mm 3 /(N·m), respectively.

关键词： Particle-reinforced stainless steel matrix composite Laser powder bed fusion 316L stainless steel Spherical WC reinforcement Tensile mechanical and tribological properties

来源：评论

学校读者我要写书评

暂无评论

LightTraffic: On Optimizing CPU-GPU Data Traffic for Efficient Large-scale Random Walks

LightTraffic: On Optimizing CPU-GPU Data Traffic for Efficie...

引用

International Conference on Data Engineering

作者： Yipeng Xing Yongkun Li Zhiqiang Wang Yinlong Xu John C. S. Lui University of Science and Technology of China Anhui Province Key Laboratory of High Performance Computing Department of Computer Science and Engineering The Chinese University of Hong Kong

As a fundamental tool for graph analysis, random walk receives extensive attention in both industry and academia. For computing massive random walks, recent works show that GPUs provide a good option to accelerate the performance. However, due to the limited memory space of modern GPUs, it is infeasible to have both the graph data and walk index fully reside in GPU memory when running large-scale random walks. Thus, it necessitates an out-of-GPU-memory design, but this inevitably induces large amounts of CPU-GPU data transmission traffic and thus hinders the overall performance. In this paper, we develop LightTraffic, which optimizes the data transmission between CPU and GPU memory under the constraint of GPU memory capacity with various system designs, including a memory-efficient scheme for partition-based management and multiple scheduling techniques. LightTraffic is a fully out-of-GPU-memory design, so it supports running large-scale random walks on GPUs. Experiments on our prototype show that LightTraffic outperforms various state-of-the-art CPU-based in-memory systems which also support large-scale random walks. For example, compared to the CPU-based systems FlashMob and ThunderRW, which are highly optimized for random walks, LightTraffic achieves 1.7−5.0× and 1.4 − 12.8× performance speedup, respectively. It also achieves up to an order of magnitude speedup when compared to the GPU-based system Subway which also supports large-scale random walks with an out-of-GPU-memory design for graph data.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Interface Bonding Characteristics and Thermal Damage Behavior of Brazed Diamond with Nb-Doped Ni-Cr Filler Alloys

SSRN

引用

SSRN 2024年

作者： Xiao, Ang Zhang, Jian Pang, Xiaotong Zhang, Mingjun Sun, Junhao Wang, Kaiming Zhang, Ao Liu, Junyi Hunan Provincial Key Laboratory of Intelligent Manufacturing Technology for High-performance Mechanical Equipment Changsha University of Science and Technology Changsha410114 China SJTU Yibin Innovation Center for Advanced Materials Yibin644002 China

Ni-Cr alloy is one of the important filler for brazed diamond tools because of its high hardness, high wear resistance and good wettability to diamond. However, the high brazing temperature and catalytic action of Ni element contained in alloy tend to cause the thermal damage of diamond and interface failure of brazed samples. In this work, the noval Ni-Cr filler alloys doped with different contents of Nb element are prepared, and the microstructures, mechanical and wear resistance properties of brazed layers used these alloys are investigated. On this basis, the interfacial bonding characteristics and thermal damage behaviors of diamond brazed with Nb-doped Ni-Cr filler alloys are further studied. The results show that the doping of Nb element improves the hardness and wear resistance of Ni-Cr filler layer, and the filler alloys exhibit good wettability and climbing effect to diamond. The refined Cr3C2 and Cr7C3 carbides are formed on the surface of diamond, which effectively enhances the interface bonding strength between filler alloy and diamond. Meanwhile, the doped Nb element consumes some Ni and generates NbNi compound, which weakens, to some extent, the graphitization catalysis and chemical erosion of Ni to diamond. When the doping amounts of Nb are 1-1.2 wt%, the brazed diamond samples present excellent interfacial bonding strength and mechanical properties. © 2024, The Authors. All rights reserved.

关键词： Diamonds

来源：评论

学校读者我要写书评

暂无评论

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

请选择保存的检索档案：

请选择收藏分类：

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

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

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：