期刊文献+
共找到15,846篇文章
< 1 2 250 >
每页显示 20 50 100
Cluster synchronization of master-slave complex networks via adaptive feedback pinning control
1
作者 LIU Ziping GONG Siyi 《上海师范大学学报(自然科学版中英文)》 2025年第4期389-400,共12页
This paper investigates the problem of cluster synchronization of master-slave complex net-works with time-varying delay via linear and adaptive feedback pinning controls.We need not non-delayed and delayed coupling m... This paper investigates the problem of cluster synchronization of master-slave complex net-works with time-varying delay via linear and adaptive feedback pinning controls.We need not non-delayed and delayed coupling matrices to be symmetric or irreducible.We have the advantages of using adaptive control method to reduce control gain and pinning control technology to reduce cost.By con-structing Lyapunov function,some sufficient synchronization criteria are established.Finally,numerical examples are employed to illustrate the effectiveness of the proposed approach. 展开更多
关键词 cluster synchronization TIME-VARYING master-slave complex networks DELAYED adaptive feedback control pinning control
在线阅读 下载PDF
关键核心技术国产替代的创新模式研究——基于CPU技术头部企业的双案例分析 被引量:5
2
作者 王砚羽 卢婷 刘汝芳 《科学学研究》 北大核心 2025年第4期712-722,750,共12页
关键核心技术是国之重器。本研究以龙芯中科和天津飞腾两家CPU技术头部企业为案例,探讨关键核心技术国产替代的创新模式。得出如下结论:(1)识别出关键核心技术国产替代的两种路径,龙芯中科采用“技术引进+学习导向的消化吸收”路径,最... 关键核心技术是国之重器。本研究以龙芯中科和天津飞腾两家CPU技术头部企业为案例,探讨关键核心技术国产替代的创新模式。得出如下结论:(1)识别出关键核心技术国产替代的两种路径,龙芯中科采用“技术引进+学习导向的消化吸收”路径,最终实现了核心技术自主可控;而天津飞腾采用“技术引进+应用导向的消化吸收”路径,存在卡脖子风险。(2)不同的国产替代创新模式塑造了不同的企业技术能力和生态能力。本研究为理解技术限制背景下的技术主权提供了新的视角,为政策制定者和行业利益相关者提供管理启示。 展开更多
关键词 关键核心技术 国产替代 创新模式 案例研究 cpu芯片
原文传递
Master-Slave机器手协调控制选择行列式算法
3
作者 杨慧 《中国民航学院学报》 2003年第A02期210-213,共4页
提出了一种基于选择行列式的Master-Slave机器手的协调控制算法。通过选择行列式的导入,实现了对Master机器手和Slave机器手各个变量的协调控制。本方法应用于装有视觉传感器的机器手的追踪控制,经过实验证明了基于选择行列式的算法能... 提出了一种基于选择行列式的Master-Slave机器手的协调控制算法。通过选择行列式的导入,实现了对Master机器手和Slave机器手各个变量的协调控制。本方法应用于装有视觉传感器的机器手的追踪控制,经过实验证明了基于选择行列式的算法能够实现人类的某些技能赋予机器手,在支援机器手的同时还能够通过有效地协调遥控,完成机器手的自主控制。 展开更多
关键词 选择行列式 master-slave机器手 协调控制 视觉传感器
在线阅读 下载PDF
基于CPU-GPU的超音速流场N-S方程数值模拟
4
作者 卢志伟 张皓茹 +3 位作者 刘锡尧 王亚东 张卓凯 张君安 《中国机械工程》 北大核心 2025年第9期1942-1950,共9页
为深入分析超音速流场的特性并提高数值计算效率,设计了一种高效的加速算法。该算法充分利用中央处理器-图形处理器(CPU-GPU)异构并行模式,通过异步流方式实现数据传输及处理,显著加速了超音速流场数值模拟的计算过程。结果表明:GPU并... 为深入分析超音速流场的特性并提高数值计算效率,设计了一种高效的加速算法。该算法充分利用中央处理器-图形处理器(CPU-GPU)异构并行模式,通过异步流方式实现数据传输及处理,显著加速了超音速流场数值模拟的计算过程。结果表明:GPU并行计算速度明显高于CPU串行计算速度,其加速比随流场网格规模的增大而明显提高。GPU并行计算可以有效提高超音速流场的计算速度,为超音速飞行器的设计、优化、性能评估及其研发提供一种强有力的并行计算方法。 展开更多
关键词 超音速流场 中央处理器-图形处理器 异构计算 有限差分
在线阅读 下载PDF
基于CPU-FPGA的SoC实验系统设计
5
作者 王丽杰 钱俊宏 +4 位作者 何俊峰 王蕊 贺媛 刘凤敏 张彤 《吉林大学学报(信息科学版)》 2025年第3期518-523,共6页
针对现有微电子与集成电路专业课程大多以理论为主,缺少仿真实验,FPGA(Field Progra mmable Gate Array)实操类实验项目严重不足的问题,设计了一套基于CPU(Central Processing Unit)-FPGA的SoC(System on Chip)实验系统。利用ModelSim... 针对现有微电子与集成电路专业课程大多以理论为主,缺少仿真实验,FPGA(Field Progra mmable Gate Array)实操类实验项目严重不足的问题,设计了一套基于CPU(Central Processing Unit)-FPGA的SoC(System on Chip)实验系统。利用ModelSim等仿真工具,以FPGA为开发平台实现CPU系统功能。以RISC-V(Reduced Instruction Set Computer)精简指令集为该CPU的指令集,以模块化为设计思想,从微处理器的局部到总体设计5级流水线CPU。系统融合了软硬件开发,能激发学生的学习兴趣。搭建的实验平台逐步实现CPU的配置与指令集至整个CPU的架构、编程、仿真、下载与调试,使学生对FPGA实现集成电路系统设计有深入理解,有助于专业理论课程的学习。通过将OBE(Outcomes-Based Education)教学理论应用于集成电路EDA(Electronic Design Automation)课程的仿真实验结果表明,这种设计方法与内容适用于产学研相结合,并能提高学生创新创业能力。 展开更多
关键词 中央处理器 现场可编程门阵列 实验系统 流水线技术
在线阅读 下载PDF
A Master-Slave Blockchain Paradigm and Application in Digital Rights Management 被引量:5
6
作者 Zhaofeng Ma Weihua Huang +2 位作者 Wei Bi Hongmin Gao Zhen Wang 《China Communications》 SCIE CSCD 2018年第8期174-188,共15页
Upon flaws of current blockchain platforms of heavyweight, large capacity of ledger, and time-consuming of synchronization of data, in this paper, we proposed a new paradigm of master-slave blockchain scheme(MSB) for ... Upon flaws of current blockchain platforms of heavyweight, large capacity of ledger, and time-consuming of synchronization of data, in this paper, we proposed a new paradigm of master-slave blockchain scheme(MSB) for pervasive computing that suitable for general PC, mobile device such as smart phones or PADs to participants in the working of mining and verification, in which we separated traditional blockchain model in 2 layer defined as master node layer and a series of slavery agents layer, then we proposed 2 approaches for partially computing model(PCM) and non-computing of model(NCM) in the MSB blockchain, Finally large amounts of simulations manifest the proposed master-slave blockchain scheme is feasible, extendible and suitable for pervasive computing especially in the 5 G generation environment, and can apply in the DRM-related applications. 展开更多
关键词 master-slave blockchain proof ofequivalent work proof of contribution pervasive computing DRM application
在线阅读 下载PDF
Achievement of chaotic synchronization trajectories of master-slave manipulators with feedback control strategy 被引量:3
7
作者 Qingkai Han Lina Hao Hao Zhang Bangchun Wen 《Acta Mechanica Sinica》 SCIE EI CAS CSCD 2010年第3期433-439,共7页
This paper addresses a master-slave synchro- nization strategy for complex dynamic systems based on feedback control. This strategy is applied to 3-DOF pla- nar manipulators in order to obtain synchronization in such ... This paper addresses a master-slave synchro- nization strategy for complex dynamic systems based on feedback control. This strategy is applied to 3-DOF pla- nar manipulators in order to obtain synchronization in such complicated as chaotic motions of end-effectors. A chaotic curve is selected from Duffing equation as the trajectory of master end-effector and a piecewise approximation method is proposed to accurately represent this chaotic trajectory of end-effectors. The dynamical equations of master-slave manipulators with synchronization controller are derived, and the Lyapunov stability theory is used to determine the stability of this controlled synchronization system. In numer- ical experiments, the synchronous motions of end-effectors as well as three joint angles and torques of master-slave manipulators are studied under the control of the proposed synchronization strategy. It is found that the positive gain matrix affects the implementation of synchronization con- trol strategy. This synchronization control strategy proves the synchronization's feasibility and controllability for com- plicated motions generated by master-slave manipulators. 展开更多
关键词 master-slave manipulators - Chaoticsynchronization trajectory - Feedback control
在线阅读 下载PDF
Exponential networked synchronization of master-slave chaotic systems with timevarying communication topologies 被引量:1
8
作者 杨东升 刘振伟 +1 位作者 赵琰 刘兆冰 《Chinese Physics B》 SCIE EI CAS CSCD 2012年第4期151-158,共8页
The networked synchronization problem of a class of master-slave chaotic systems with time-varying communication topologies is investigated in this paper. Based on algebraic graph theory and matrix theory, a simple li... The networked synchronization problem of a class of master-slave chaotic systems with time-varying communication topologies is investigated in this paper. Based on algebraic graph theory and matrix theory, a simple linear state feedback controller is designed to synchronize the master chaotic system and the slave chaotic systems with a time- varying communication topology connection. The exponential stability of the closed-loop networked synchronization error system is guaranteed by applying Lyapunov stability theory. The derived novel criteria are in the form of linear matrix inequalities (LMIs), which are easy to examine and tremendously reduce the computation burden from the feedback matrices. This paper provides an alternative networked secure communication scheme which can be extended conveniently. An illustrative example is given to demonstrate the effectiveness of the proposed networked synchronization method. 展开更多
关键词 exponential networked synchronization master-slave chaotic systems algebraic graph theory communication topology
原文传递
基于多核CPU的雷达导引头数字化实时仿真研究 被引量:1
9
作者 苏灏杨 夏伟杰 +1 位作者 吴雪 王宇 《遥测遥控》 2025年第2期92-99,共8页
雷达导引头仿真系统在导引头精确打击目标的过程中起着重要的作用。随着仿真系统的复杂度不断提升以及数据处理需求的日益增长,传统的串行计算仿真技术已难以满足雷达导引头数字仿真系统对实时性的严格要求。针对现有雷达导引头仿真过... 雷达导引头仿真系统在导引头精确打击目标的过程中起着重要的作用。随着仿真系统的复杂度不断提升以及数据处理需求的日益增长,传统的串行计算仿真技术已难以满足雷达导引头数字仿真系统对实时性的严格要求。针对现有雷达导引头仿真过程中耗时过长的问题,本文提出了一种全流程的数字化实时仿真方法。首先,将传统全流程仿真架构的核心部分——接收控制系统指令、接收回波仿真数据、SAR(Synthetic aperture Radar,合成孔径雷达)成像处理、成像结果上传与界面动态更新进行流水并行化。其次,利用OpenMP(开放式多处理)多核并行模型,对SAR成像算法主要步骤进行多核CPU(Central Processing Unit,中央处理器)并行处理。然后,引入高性能数学计算库FFTW3(西方最快傅里叶变换第3版)快速实现成像算法的傅里叶变换,加快SAR成像算法处理速度。最后仿真结果表明:该全流程的设计方法相较于传统的串行仿真,加速比达到100倍左右,同时加速前后的SAR图像相似度接近于1。在处理精度和效果一致的前提下,该方法能够完成雷达导引头系统的全流程实时仿真,具有较好的工程应用前景。 展开更多
关键词 雷达导引头 SAR仿真 多核cpu 并行计算 实时仿真
在线阅读 下载PDF
A master-slave generalized predictive synchronization control for preheating process of multi-cavity hot runner system 被引量:1
10
作者 Hongyi Qu Shengyong Mo +3 位作者 Ke Yao Zhao-Xia Huang Zhihao Xu Furong Gao 《Chinese Journal of Chemical Engineering》 SCIE EI CAS CSCD 2023年第10期270-280,共11页
As a key component of injection molding,multi-cavity hot runner(MCHR)system faces the crucial problem of polymer melt filling imbalance among the cavities.The thermal imbalance in the system has been considered as the... As a key component of injection molding,multi-cavity hot runner(MCHR)system faces the crucial problem of polymer melt filling imbalance among the cavities.The thermal imbalance in the system has been considered as the leading cause.Hence,the solution may rest with the synchronization of those heating processes in MCHR system.This paper proposes a’Master-Slave’generalized predictive synchronization control(MS-GPSC)method with’Mr.Slowest’strategy for preheating stage of MCHR system.The core of the proposed method is choosing the heating process with slowest dynamics as the’Master’to track the setpoint,while the other heating processes are treated as‘Slaves’tracking the output of’Master’.This proposed method is shown to have the good ability of temperature synchronization.The corresponding analysis is conducted on parameters tuning and stability,simulations and experiments show the strategy is effective. 展开更多
关键词 Process control Thermodynamics process Model-predictive control Multi-cavity hot runner system master-slave synchronization Mr.Slowest
在线阅读 下载PDF
Research on Master-slave Telerobot Force Telepresence Technology
11
作者 蔡鹤皋 陈卫东 +1 位作者 贾国武 赵杰 《Journal of Harbin Institute of Technology(New Series)》 EI CAS 1997年第4期69-72,共4页
A 3 DOF master-slave telerobot system is established for study on force telepresence technology. A force feedback and position control scheme is adopted in the bilateral force response control system,and force fidelit... A 3 DOF master-slave telerobot system is established for study on force telepresence technology. A force feedback and position control scheme is adopted in the bilateral force response control system,and force fidelity and controllability experiments demonstrate feasibility of the con-trol system. 展开更多
关键词 master-slave TELEROBOT FORCE TELEPRESENCE BILATERAL FORCE response control MANOEUVRABILITY
在线阅读 下载PDF
Jet sweeping angle control by fluidic oscillators with master-slave designs
12
作者 Ziyan LI Kaiwen ZHOU +1 位作者 Yingzheng LIU Xin WEN 《Chinese Journal of Aeronautics》 SCIE EI CAS CSCD 2021年第5期145-162,共18页
The fluidic oscillator is an instrument that can continuously generate a spatially sweeping jet entirely based on its internal geometry without any moving parts.However,the traditional fluidic oscillator has an inhere... The fluidic oscillator is an instrument that can continuously generate a spatially sweeping jet entirely based on its internal geometry without any moving parts.However,the traditional fluidic oscillator has an inherent limitation,that is,the spreading angle cannot be controlled independently,rather by the jet volume flow rate and internal geometry.Accordingly,two types of fluidic oscillators based on the master-slave design are developed in current study to decouple this correlation.In both designs,the master layer inherits the similar oscillation mechanisms of a sweeping jet,and the slave layer resembles a steady jet channel.The difference between the two designs is that Design A has a short diverging exit in the slave layer,but Design B adds a long interaction chamber in the exit channel to intensify flow instability.The external flow fields and governing oscillation properties of these two designs are experimentally explored with time-resolved Particle Image Velocimetry(PIV),while the internal flow dynamics and driving oscillation mechanisms are numerically investigated.By fixing the total volume flow rate,the jet spreading angle of Design A can be increased smoothly from 0°to above 100°by increasing the proportion of master layer’s flow rate from 0 to 100%.For Design B,the control authority of the master layer is significantly enhanced by adding the interaction chamber in the slave layer.In addition,the added chamber causes notable jet oscillation even when the master layer has none input. 展开更多
关键词 Adjustable spreading angle CFD Fluidic devices master-slave fluidic oscillator TR-PIV
原文传递
利用CPU和GPU混合并行方法快速构建海洋扰动重力梯度基准图
13
作者 黄炎 李姗姗 +3 位作者 吕明昊 范雕 谭勖立 冯进凯 《武汉大学学报(信息科学版)》 北大核心 2025年第3期515-527,共13页
高精度、高分辨率的海洋扰动重力梯度基准图是将水下扰动重力梯度辅助惯性导航付诸于实践的关键技术之一,可依据边值问题理论,基于移去-恢复技术计算求得。基于传统串行算法存在计算效率低、耗时长等问题,为满足大范围乃至全球海洋扰动... 高精度、高分辨率的海洋扰动重力梯度基准图是将水下扰动重力梯度辅助惯性导航付诸于实践的关键技术之一,可依据边值问题理论,基于移去-恢复技术计算求得。基于传统串行算法存在计算效率低、耗时长等问题,为满足大范围乃至全球海洋扰动重力梯度基准图的快速构建需求,利用中央处理器(central processing unit,CPU)与图形处理器(graph processing unit,GPU)混合并行(CPU+GPU)编程平台,设计了一种高效的混合并行计算方案。首先,利用数组收缩膨胀方法,有效解决了CPU进行勒让德函数递推计算过程中内存读写冲突问题;然后,引入Hilbert空间填充曲线,将二维格网重力异常数据降维成为具有高度聚簇性的一维数组,通过其在GPU显存中的快速索引,实现了Stokes积分的高效计算。在CPU型号为Intel Xeon(R)Gold 6130、GPU型号为Tesla V100的计算机上进行实验,利用该混合并行方案计算6°×6°范围、1′分辨率的全张量海洋扰动重力梯度基准图仅需15.84 s,而传统串行方法需要35 min以上;计算2°×2°范围、30″分辨率基准图仅需22.7 s,串行方法则需要52 min;计算2°×2°范围、10″分辨率基准图需要26.7 min,串行方法则需要两天以上。在保证并行计算绝对误差小于等于1×10^(-6) E、T_(xx)+T_(yy)+T_(zz)绝对值的最大值小于0.1 E、均方根误差小于0.01 E的前提下,实现了全球全张量海洋扰动重力梯度基准图模型的快速构建。 展开更多
关键词 扰动重力梯度 边值问题 cpu GPU 混合并行
原文传递
Adaptive H_∞ Synchronization of Master-slave Systems with Mixed Time-varying Delays and Nonlinear Perturbations: An LMI Approach
14
作者 Hamid Reza Karimi 《International Journal of Automation and computing》 EI 2011年第4期381-390,共10页
This paper proposes an adaptive synchronization problem for the master and slave structure of linear systems with nonlinear perturbations and mixed time-varying delays comprising different discrete and distributed tim... This paper proposes an adaptive synchronization problem for the master and slave structure of linear systems with nonlinear perturbations and mixed time-varying delays comprising different discrete and distributed time delays. Using an appropriate Lyapunov-Krasovskii functional, some delay-dependent sufficient conditions and an adaptation law including the master-slave parame- ters are established for designing a delayed synchronization law in terms of linear matrix inequalities(LMIs). The time-varying controller guarantees the H ∞ synchronization of the two coupled master and slave systems regardless of their initial states. Particularly, it is shown that the synchronization speed can be controlled by adjusting the updated gain of the synchronization signal. Two numerical examples are given to demonstrate the effectiveness of the method. 展开更多
关键词 Adaptive synchronization master-slave systems DELAY H performance nonlinear perturbations.
在线阅读 下载PDF
Tremor suppression in master-slave robot remote welding system
15
作者 陈洪堂 李海超 +1 位作者 高洪明 吴林 《China Welding》 EI CAS 2011年第2期63-66,共4页
Welding seam tracking precision was decreased due to human hand tremor during the master-slave welding teleoperation. To solve this problem, a master-slave robot remote welding system was built, the system consisted o... Welding seam tracking precision was decreased due to human hand tremor during the master-slave welding teleoperation. To solve this problem, a master-slave robot remote welding system was built, the system consisted of a master manipulator with six degree of freedom ( DOF ) , an industrial computer control system and a slave Motoman HP3 J robot, and human hand tremor and digital filtering were discussed. An optimal digital filter was designed to clean human tremor signal for improving the welding seam tracking precision. The experimental results show that the digital filter suppresses the operator' s tremor signal. 展开更多
关键词 remote welding master-slave teleoperation hand tremor digital filtering
在线阅读 下载PDF
Hierarchical hybrid control network design based on LON and master-slave RS-422/485 protocol
16
作者 彭可 陈际达 陈岚 《Journal of Central South University of Technology》 2002年第3期202-207,共6页
Aiming at the weaknesses of LON bus, combining the coexistence of fieldbus and DCS (Distribu ted Control Systems) in control networks, the authors introduce a hierarchical hybrid control network design based on LON an... Aiming at the weaknesses of LON bus, combining the coexistence of fieldbus and DCS (Distribu ted Control Systems) in control networks, the authors introduce a hierarchical hybrid control network design based on LON and master slave RS 422/485 protocol. This design adopts LON as the trunk, master slave RS 422/485 control networks are connected to LON as special subnets by dedicated gateways. It is an implementation method for isomerous control network integration. Data management is ranked according to real time requirements for different network data. The core components, such as control network nodes, router and gateway, are detailed in the paper. The design utilizes both communication advantage of LonWorks technology and the more powerful control ability of universal MCUs or PLCs, thus it greatly increases system response speed and performance cost ratio. 展开更多
关键词 LON fieldbus master-slave RS-422/485 PROTOCOL HIERARCHICAL hybrid control NETWORKS router gateway NETWORKS integration
在线阅读 下载PDF
Distributed Generation Islanding Effect on Distribution Networks and End User Loads Using the Master-Slave Islanding Method
17
作者 Lambros Ekonomou George P. Fotis +1 位作者 Vasiliki Vita Valeri Mladenov 《Journal of Power and Energy Engineering》 2016年第10期1-24,共24页
This study aims to address the feasibility of planned islanding operation and to investigate the effect of unplanned islanding using the master-slave islanding method for controlling the distributed generation units d... This study aims to address the feasibility of planned islanding operation and to investigate the effect of unplanned islanding using the master-slave islanding method for controlling the distributed generation units during grid-connected and islanding operation. Neplan desktop power simulation tool was used for the modelling and simulation of a realistic MV network with four different distributed generation technologies (diesel, gas, hydro and wind) along with their excitation and governor control systems, while an exponential model was used to represent the loads in the network. The dynamic and steady state behavior of the four distributed generation technologies were investigated during grid-connected operation and two transition modes to the islanding situation, planned and unplanned. The obtained results that validated through various case studies have shown that a suitable planned islanding transition could provide support to critical loads at the event of electricity utility outages. 展开更多
关键词 Distributed Generation Distribution Networks ISLANDING master-slave Islanding Method Neplan Simulation Tool
在线阅读 下载PDF
基于强化学习的负载感知CPU资源分配和管理方法
18
作者 许荣飞 苏志远 +1 位作者 麻付强 吴保锡 《计算机技术与发展》 2025年第10期81-88,共8页
随着CPU核的数量增多,合理分配CPU核对于降低系统功耗具有重要意义,如何根据系统运行时的负载情况进行精准的CPU资源分配和管理是一个关键的问题。现在处理器设计提供了很多对功耗优化的机制(比如动态电压频率调整DVFS),但是要让这些机... 随着CPU核的数量增多,合理分配CPU核对于降低系统功耗具有重要意义,如何根据系统运行时的负载情况进行精准的CPU资源分配和管理是一个关键的问题。现在处理器设计提供了很多对功耗优化的机制(比如动态电压频率调整DVFS),但是要让这些机制发挥作用,只有芯片的支持是不够的,还需要软硬协同设计。当前缺乏基于软件来最大化利用这些硬件机制的手段。近年来,机器学习在各个领域展现出巨大的潜力,很多基于机器学习的研究工作应运而生。其中,强化学习具有较强的自适应性,适用于动态感知系统环境并进行资源管理。因此,该文提出了一种基于强化学习的负载感知CPU资源分配和管理方法——RLWAM。该方法提出基于最小原则根据系统中运行时的任务负载进行CPU资源分配和管理,基于强化学习提出了面向上述场景的Q-Learning算法,包括面向任务和系统的状态建模方式、面向绑核、调频和资源整合的动作空间和激励函数,从而帮助系统进一步降低功耗。最后,通过在真实平台上从单类型任务上的绑核调频和多类型任务上的资源整合两个场景对该方法进行实验验证,结果表明该方法具有显著的有效性和可扩展性。 展开更多
关键词 负载感知 强化学习 多核系统 cpu资源分配管理 绑核调频
在线阅读 下载PDF
微处理器性能分析与优化:基于SPEC CPU2017的对比研究
19
作者 徐晗 郭振江 肖俊华 《高技术通讯》 北大核心 2025年第3期241-249,共9页
运行标准测试程序是进行微处理器设计空间探索的基本手段。横向对比分析国内外主流处理器在标准测试程序场景下的各项性能指标,有助于识别国产处理器的性能瓶颈,为进一步的性能优化指明方向。本文基于SPEC CPU2017对3款微处理器进行同... 运行标准测试程序是进行微处理器设计空间探索的基本手段。横向对比分析国内外主流处理器在标准测试程序场景下的各项性能指标,有助于识别国产处理器的性能瓶颈,为进一步的性能优化指明方向。本文基于SPEC CPU2017对3款微处理器进行同频性能测试和对比分析,分别是龙芯LA464架构的3A5000微处理器、AMD ZEN1架构的R3-1200以及Intel Skylake架构的i3-9100f。根据测试结果,3A5000定点性能与R3-1200基本相同,比i3-9100f低10%左右;3A5000浮点性能相当于另2款微处理器的70%左右。本文从动态指令数和每周期指令数(instruction per cycle,IPC)2个角度对微处理器进行比较和分析。基于SPEC CPU2017的结果显示,3A5000定点动态指令数和浮点动态指令数分别比另2款微处理器多约10%和25%。在3A5000上使用激进的自动向量化编译优化策略、优化立即数乘法编译效率等手段可以将其性能提升10%左右。3A5000的定点IPC比另外2款微处理器高4%左右,浮点IPC低8%左右。发射宽度、执行单元数量、功能和延迟等微结构参数接近是3款微处理器IPC差别较小的主要原因。 展开更多
关键词 SPEC cpu2017 性能分析 龙芯3A5000 向量化 体系结构
在线阅读 下载PDF
基于云平台CPU与GPU协同处理的光学卫星遥感影像正射融合方法
20
作者 于潇 张一 +2 位作者 吕丽红 张强 王得成 《空间科学学报》 北大核心 2025年第5期1416-1424,共9页
系统探讨了基于国产云平台调度下自主可控CPU和GPU协同处理的光学卫星遥感影像正射融合方法执行效率问题,通过数据流配置、中间数据存储访问优化等手段进一步提高了该方法执行效率.在云平台调度下,使用飞腾S2500和英伟达A100对高分二号... 系统探讨了基于国产云平台调度下自主可控CPU和GPU协同处理的光学卫星遥感影像正射融合方法执行效率问题,通过数据流配置、中间数据存储访问优化等手段进一步提高了该方法执行效率.在云平台调度下,使用飞腾S2500和英伟达A100对高分二号卫星多光谱影像进行正射融合的试验,结果表明,该方法可很大程度提高光学卫星遥感影像正射融合效率,与传统X86架构CPU与GPU协同的正射融合算法相比,加速比为14.3倍以上,数据处理时间压缩至8.4 s内,其中GPU运算耗时仅1 s,可满足并优化大数据量的光学卫星遥感影像快速正射融合的要求. 展开更多
关键词 正射融合 国产云平台 cpu和GPU协同处理 数据流配置 存储访问优化
在线阅读 下载PDF
上一页 1 2 250 下一页 到第
使用帮助 返回顶部