检索结果-南通市图书馆

A user-space file system for on-demand legacy desktop software

在线全文

同方期刊数据库

学校读者我要写书评

暂无评论

science china(Information sciences) 2011年第6期54卷 1142-1150页

作者： ZHANG YouHui 1,2 , SU GeLin 1,2 & zheng weimin 1,2 1 department of computer science and technology, tsinghua university, beijing 100084, china 2 tsinghua National Laboratory for Information science and technology, tsinghua university, beijing 100084, china 1. department of computer science and technology Tsinghua University Beijing 100084 China2. Tsinghua National Laboratory for Information Science and Technology Tsinghua University Beijing 100084 China

Some user-level virtualization technologies have been used to convert legacy software (like the existing Windows desktop applications) into the on-demand software without any modification. To give the client a friendly and compatible method to access on-demand legacy software across the Internet, this paper presents a clientend file system for this usage case. It is a Windows user-space file system based on the cloud storage (where the on-demand software is stored), which converts the local file system accesses into remote visits. Quite a few optimizations are adopted and adjusted to suit the file-access-pattern of on-demand software, including the local cache, metadata/data/software prefetch and content-addressable storage (CAS), to decrease the number of remote visits and/or to overlap IO (input/output) operations with software execution. Detailed access-pattern analyses are also presented. This file system has been implemented and tests show that it is practical for much daily-used software-a local cache of limited size can provide up to 80% hit ratio and the corresponding overhead of running-time is about 37%. Owing to this method, on any compatible and networked computer, a user can use his/her personalized software conveniently although it does not exist on the local host.

关键词： on-demand software software as a service user-space file system cloud storage

CUDA-Zero:a framework for porting shared memory GPU applications to multi-GPUs

在线全文

同方期刊数据库

学校读者我要写书评

暂无评论

science china(Information sciences) 2012年第3期55卷 663-676页

作者： CHEN DeHao , CHEN WenGuang & zheng weimin department of computer science and technology, tsinghua university, beijing 100084, china 1. department of computer science and technology Tsinghua University Beijing 100084 China

As the prevalence of general purpose computations on GPU, shared memory programming models were proposed to ease the pain of GPU programming. However, with the demanding needs of more intensive workloads, it’s desirable to port GPU programs to more scalable distributed memory environment, such as multi-GPUs. To achieve this, programs need to be re-written with mixed programming models (e.g. CUDA and message passing). Programmers not only need to work carefully on workload distribution, but also on scheduling mechanisms to ensure the efficiency of the execution. In this paper, we studied the possibilities of automating the process of parallelization to multi-GPUs. Starting from a GPU program written in shared memory model, our framework analyzes the access patterns of arrays in kernel functions to derive the data partition schemes. To acquire the access pattern, we proposed a 3-tiers approach: static analysis, profile based analysis and user annotation. Experiments show that most access patterns can be derived correctly by the first two tiers, which means that zero efforts are needed to port an existing application to distributed memory environment. We use our framework to parallelize several applications, and show that for certain kinds of applications, CUDA-Zero can achieve efficient parallelization in multi-GPU environment.

关键词： CUDA parallelization data access pattern multi-GPU

Predicting the behavior of large scale P2P systems by parallel discrete event simulation

在线全文

同方期刊数据库

学校读者我要写书评

暂无评论

science china(Information sciences) 2010年第6期53卷 1109-1121页

作者： zheng weimin1,YU HongLiang1,SHI GuangYu2 & CHEN Jian2 1department of computer science and technology,tsinghua university,beijing 100084,china 2Huawei Technologies Co.,Ltd,Shenzhen 518129,china 1. department of computer science and technology Tsinghua University Beijing 100084 China2. Huawei Technologies Co. Ltd Shenzhen 518129 China

P2P systems are becoming the dominator of *** systems are typically composed of thousands to millions of physical computers,which make it difficult to predict their behaviors without a large scale distributed system *** paper is an attempt to predict the behavior of large scale P2P systems by building a novel parallel simulator:AegeanSim,which provides parallel discrete event simulation of such systems on high performance server *** abstract the execution of P2P applications with a specific event model,and parallel the simulation of events in a cluster,thus expanding the simulation scale and boosting the simulation process dramatically.A 1-stage synchronization method is proposed to improve the *** event callback interface is designed to implement their application logic easily while keeping the simulator *** use AegeanSim to predict the behavior of a typical P2P system:*** comparing the simulation behavior of BT with that of related BT studies and verifying its efficiency,scalability and accuracy,we make predictions about the behavior of BT system assuming they are assaulted by different man-made system *** reasonable results are found:(1) Tracker isolation can hardly work because of its short attacking time window.(2) Limiting the bandwidth of BT may be an efficient way to control it.

关键词： P2P simulation parallel

Optimizing write operation on replica in data grid

在线全文

同方期刊数据库

学校读者我要写书评

暂无评论

science china(Information sciences) 2011年第1期54卷 1-11页

作者： XU PengZhi,WU YongWei,HUANG XiaoMeng, YANG GuangWen & zheng weimin department of computer science and technology,tsinghua university,beijing 100084,china 1. department of computer science and technology Tsinghua University Beijing 100084 China

Replication technique can improve access performance and data availability of data grid systems. So far,most related work on data replication in data grid systems focuses on read-only *** write operations have been widely used in practice,its optimization problem has not been suciently *** this paper,we show that write requests from users can be processed by a nearest site,regardless of the location of latest written *** policy of write nearby instead of write to the latest replica can improve the performance of write operations,while inconsistency is caused between multiple ***,we propose two metadata management algorithms to make sure that users can retrieve up-to-date datasets at any *** extend OptorSim to support the implementation and evaluation of our *** results of simulation show that our scheme achieves performance gains up to 3.3 compared with WriteMaster scheme.

关键词： replication data grid simulation access latency

A　Nonprecedence－based　Query　Schedule　Algorithm　on　Heterogeneous　Multiprocessor　System

在线全文

同方期刊数据库

学校读者我要写书评

暂无评论

Journal of Systems Engineering and Electronics 1995年第3期6卷 63-69页

作者： He Jun bang Dingxing & zheng weimin(department of computer science & technology,tsinghua university, beijing, 100084, china) department of computer science & technology Tsinghua University Beijing 100084 China

There has been a number of algorithms designed to handle intra-query or inter-query scheduling in multiprocessor-based parallel database system. They all have the assumption that the processorsare identical. But in some cases, such as shared nothing environment, this assumption would not be fullyvalid. In this paper we devise and evaluate a scheduling algorithm ELLIST to handle nonprecedence-basedheterogeneous malleable scheduling problem. It uses LLIST-NM as subroutine that handles nonmalleablescheduling without precedence and assuming the processors are identical. Even though the problem we considered is NP-hard in the strong sense, the schedule generated by our algorithm is seen experimentally toachieve results that are close to optimum when there are enough tasks to be scheduled.

关键词： Heterogeneous malleable scheduling Makespan Affinity

维普期刊数据库

在线全文

学校读者我要写书评

暂无评论

基于特征子空间的三维模型相关反馈检索方法

系统仿真学报 2008年第18期20卷 4918-4922页

作者：李立群覃征冷彪清华大学软件学院北京100084 清华大学计算机科学与技术系北京100084

从三维模型中提取出的高维特征向量一般分布在原始空间的某个非线性低维子空间中,而测地距度量算法可以有效地甄别这种非线性分布。提出一种在相关反馈中,通过更新测地距离来捕捉三维模型特征点的分布特性的检索算法。为了解决相关反馈... 详细信息

从三维模型中提取出的高维特征向量一般分布在原始空间的某个非线性低维子空间中,而测地距度量算法可以有效地甄别这种非线性分布。提出一种在相关反馈中,通过更新测地距离来捕捉三维模型特征点的分布特性的检索算法。为了解决相关反馈后期效果提升缓慢的缺陷,作者提出了模型潜力值理论来辅助使用了测地距核函数的支持向量机学习器改善相关反馈机制。

关键词：相关反馈测地距流形学习模型潜力

维普期刊数据库

Right-and-left visual field stimulation:A frequency and space mixed coding method for SSVEP based brain-computer interface

在线全文

学校读者我要写书评

暂无评论

science china(Information sciences) 2011年第12期54卷 2512-2518页

作者： YAN zheng, GAO XiaoRong & GAO ShangKai department of Biomedical Engineering, School of Medicine, tsinghua university, beijing 100084, china 1. department of Biomedical Engineering School of Medicine Tsinghua University Beijing 100084 China

On the condition of using limited frequencies, fewer targets could be presented in brain-computer interface (BCI) based on steady-state visual evoked potentials (SSVEP). This paper proposes a novel coding method for SSVEP that, through a combination of frequency and spatial information, could increase the number of targets. Each target was composed of two flickers placed in the right-and-left visual fields. Given the role of the optic chiasm in the vision pathway, the two frequency components could be projected to contralateral occipital regions. Canonical correlation analysis (CCA) was utilized to detect the frequency components from the right or left visual cortex. Asymmetric index as a supplementary feature was also computed. Linear discriminant analysis (LDA) was used for target recognition. The attractive feature of this method is that it would substantially increase the number of targets. If the number of frequency was N, the method could present N times more targets than conventional SSVEP-based BCIs.

关键词： brian-computer interface (BCI) frequency-space mixed coding steady-state visual evoked potential (SSVEP)

A low power and small area all digital delay-locked loop based on ring oscillator architecture

在线全文

同方期刊数据库

学校读者我要写书评

暂无评论

science china(Information sciences) 2012年第2期55卷 453-460页

作者： zheng JiaPeng1,LI Wei2,LU XueQing2,CHENG YuHua1 & WANG YangYuan1 1department of Microelectronics,School of Electronics Engineering and computer science,Peking university,beijing 100871,china 2Semiconductor Manufacturing International Corp,Shanghai 201203,china 1. department of Microelectronics School of Electronics Engineering and Computer Science Peking University Beijing 100871 China2. Semiconductor Manufacturing International Corp Shanghai 201203 China

A 133-500 MHz,5.2 mW @500 MHz,0.021 mm2 all digital delay-locked loop(ADDLL)is *** power and area reduction of the proposed ADDLL are achieved by implementing a high frequency ring oscillator(ROSC)to count the reference clocks such that the one-clock cycle delay chain and the phase detector in a conventional Master block are no longer *** proposed ADDLL has better immunity to PVT(process,voltage,and temperature)than most conventional DLLs,which do not update the control word signals after the locking process,since the control signals for slave delay line are updated in every 256 reference *** in 0.13 um CMOS process,the measured RMS jitter is 10.83 ps at 500 MHz while the RMS jitter of the input signal is 9.97 ps.

关键词： all digital delay locked loop(DLL) phase locked loop(PLL) ring oscillator

A new method for simulating human emotions

在线全文

同方期刊数据库

学校读者我要写书评

暂无评论

Journal of university of science and technology beijing 2003年第2期10卷 72-74页

作者： weimin Xue, Zhiliang Wang, Wenhong Xia, and Wenbo MengInformation Engineering School, university of science and technology beijing, beijing 100083, china Information Engineering School University of Science and Technology Beijing Beijing 100083 China

How to make machines express emotions would be instrumental in establishing acompletely new paradigm for man machine interaction. A new method for simulating and assessingartificial psychology has been developed for the research of the emotion robot. The human psychologyactivity is regarded as a Markov process. An emotion space and psychology model is constructedbased on Markov process. The conception of emotion entropy is presented to assess the artificialemotion complexity. The simulating results play up to human psychology activity. This model can alsobe applied to consumer-friendly human-computer interfaces, and interactive video etc.

关键词： emotion robot psychology entropy Markov model artificial psychology

维普期刊数据库