多核多线程下java设计模式性能提升

现代大型系统中多核多线程下的应用越来越多,java语言发展至今对于并行机制有了很大改善,架构中的设计模式从性能的角度也在发生着变化,文章着重介绍几个典型设计模式进行探讨,以便在系统重构或开发中得到性能改进。以及JDK在设计模式上改进的期待。     

基于单点登录面向服务架构的系统整合技术研究

为实现面向未来的无缝系统集成的目标,提出了面向服务架构采用单点登录技术和Web服务技术进行系统整合的观点;阐述了数据整合、应用整合方法,针对历史遗留系统的集成问题,给出了基于数据同步的松耦合的解决办法;并将这些设计应用到大型金融系统中,进行了实践验证;实践表明面向服务架构采用单点登录技术和Web服务技术进行系统整合切实可行,是系统集成的发展方向.     

遗留系统并行架构的设计与实现

面向服务的体系架构(SOA)为遗留系统的再工程提供了解决方案,使得遗留系统可以支持分布式应用环境,但是由于技术的陈旧和架构的局限性,无法支持多线程、并行处理以及内存泄露等问题依旧在部分遗留系统中存在,极大地限制了它们的应用。为了解决这几个问题,通过深入分析研究Windows 通信基础(WCF)的通信机制,提出了一种并行架构,对WCF的基本架构进行了改造,即在默认的体系架构中添加一层服务控制器,用来在客户端和服务端之间传递消息和选择服务,很好地解决了这几个问题,并在某大型金融软件中得到了应用。     

基于混合编程模型的支持向量机训练并行化

支持向量机(support vector machine,SVM)是一种广泛应用于统计分类以及回归分析的监督学习方法.基于内点法(interior point method,IPM)的SVM训练具有空间占用小、送代趋近快等优点,但随着训练数据集规模的增大,仍面临处理速度与存储空间所带来的双重挑战.针对此问题,提出利用CPU-GPU异构系统进行大规模SVM训练的混合并行机制.首先利用计算统一设备架构(compute unified device architecture,CUDA)对基于内点法的SVM训练算法的计算密集部分并行化,并改进算法使其适合利用cuBLAS线性代数库加以实现,提高训练速度;然后利用消息传递接口(message passing interface,MPI)在集群系统上实现CUDA加速后算法的分布并行化,利用分布存储有效地增加所处理数据集规模并减少训练时间;进而基于Fermi架构支持的页锁定内存技术,打破了GPU设备存储容量不足对数据集规模的限制.结果表明,利用消息传递接口(MPI)和CUDA混合编程模型以及页锁定内存数据存储策略,能够在CPU-GPU异构系统上实现大规模数据集的高效并行SVM训练,提升其在大数据处理领域的计算性能和应用能力.     

基于机群架构的并行数据库中间件系统改进研究

基于机群架构的高性能计算机已经被应用到很多领域,如大气预测、油藏模拟、CFD仿真模拟和Web服务等.随着数据量的爆炸式增长,传统的集中式的数据库系统已经难以满足各种应用的需求,基于机群式架构的并行数据库系统为增强海量数据的存储以及处理能力提供了一种途径.对以前实现的一个基于机群架构的并行数据库中间件系统ParaMidSQL进行了改进.通过对并行选择、并行排序、并行连接等关键操作的测试,给出对并行数据库中间件系统改进后的性能分析.     

基于三级并行的遥感业务化处理系统研究

分析了遥感业务化处理系统的现实需求,针对当前遥感业务化处理系统存在的不足,结合遥感业务化处理系统和遥感影像并行处理算法两方面研究的成果,提出了基于三级并行的遥感业务化处理系统设计方案,将数据并行、功能并行、任务并行融合到遥感业务化系统中,并在环境与灾害监测预报小卫星星座系统工程中得到应用,为环境减灾提供了快速的遥感处理支持。介绍了系统的体系结构,工作流程,重点研究了三级并行的实现方式,并借助产品实例对其性能进行了测试。结果表明,基于三级并行的遥感业务化处理系统设计方案能够有效的提高遥感业务化处理的效率。     

质变的飞跃专业卡王AMD FirePro V9800面世

在连续发布五款Evergreen DXl1新架构的FirePro 3D专业图形工作站显卡之后,AMD再次推出了一款定位于超高端市场的旗舰级产品：FirePro V9800,面向那些需要极致性能、处理大型视觉化和复杂数据集的专业客户。     

基于GPU的并行化Apriori算法的设计与实现

大数据和高度并行的计算架构的时代已经来临,如何让传统的串行数据挖掘方法在当下获得更高的效率是一个值得探讨的问题。根据现代GPU大规模并行运算架构的特点(单结构多数据),对传统的串行Apriori算法进行并行化处理。使用最新的CUDA技术完成对传统串行Apriori算法中的支持度统计、候选集生成这两个计算的并行化实现,讨论了多种实现方法的差异,并提出改进方案。实验表明:改进后的并行算法使支持度统计在10000条事务的条件下效率提高16%,候选集生成在10000条事务的条件下效率提高25%。     

一种高性能I／O方法

并存文伴系统是解决I／O瓶颈问题的重要途径。研究表明，科学应用中跨越式的文件访问模式与现存并行文件系统访问这些数据的方法的结合，对于大型数据集的访问其I／O性能是难以接受的。为了提高并行文件系统中对不连续数据的I／O性能，创建了一种新型高性能I／O方法：用户自定义文件视图结合合并I／O请求。并且在WPFS并行文件系统中实现了该方法。研究和实验结果表明，该方法具有增强科学应用性能的潜力。     

基于CUDA的SVM算法并行化研究

SVM算法在统计分类以及回归分析中得到了广泛的应用。而随着物联网的迅速发展,SVM算法在各种应用中往往需要解决大量数据的快速处理问题。在SVM算法并行化研究中,首先对SVM算法进行分析研究,提出了基于CUDA的SVM算法并行化方案;其次,进一步研究海量数据的处理,提出海量数据处理的并行化方案;最后,通过实验分析对比了并行化算法的性能。     

Legacy code and parallel computing: updating and parallelizing a numerical model

In this paper, we present several important details in the process of legacy code parallelization, mostly related to the problem of maintaining numerical output of a legacy code while obtaining a balanced workload for parallel processing. Since we maintained the non-uniform mesh imposed by the original finite element code, we have to develop a specially designed data distribution among processors so that data restrictions are met in the finite element method. In particular, we introduce a data distribution method that is initially used in shared memory parallel processing and obtain better performance than the previous parallel program version. Besides, this method can be extended to other parallel platforms such as distributed memory parallel computers. We present results including several problems related to performance profiling on different (development and production) parallel platforms. The use of new and old parallel computing architectures leads to different behavior of the same code, which in all cases provides better performance in multiprocessor hardware.

     

A Parallelization Approach for Hard Real-Time Systems and Its Application on Two Industrial Programs

Applications in industry often have grown and improved over many years. Since their performance demands increase, they also need to benefit from the availability of multi-core processors. However, a reimplementation from scratch and even a restructuring of these industrial applications is very expensive, often due to high certification efforts. Therefore, a strategy for a systematic parallelization of legacy code is needed. We present a parallelization approach for hard real-time systems, which ensures a high reusage of legacy code and preserves timing analysability. To show its applicability, we apply it on the core algorithm of an avionics application as well as on the control program of a large construction machine. We create models of the legacy programs showing the potential of parallelism, optimize them and change the source codes accordingly. The parallelized applications are placed on a predictable multi-core processor with up to 18 cores. For evaluation, we compare the worst case execution times and their speedups. Furthermore, we analyse limitations coming up at the parallelization process.     

Scalable Parallel and Distributed Expert Database Systems with Predictive Load Balancing

Much prior work in AI on various attempts to speed up rule-based systems by parallel processing has been reported. Unfortunately, many of these results indicate that there is limited parallelism to be found when rules are applied to relatively small amounts of data. Thus, one can predict that much greater parallelism can be extracted when rules are applied to large amounts of data. However, traditional compile-time parallelization strategies as developed for main-memory based systems do not scale to large databases. We propose a scalable strategy for the efficient parallel implementation of rule-based systems operating upon large databases. We concentrate on load balancing techniques in a synchronous model of rule execution, where the variance in runtime of the distributed sites is minimized per cycle of rule processing, thus increasing utilization and speedup. We demonstrate that static load balancing techniques are insufficient, and thus low overhead dynamic load balancing is the key to successful scaling. We present a form of dynamic load balancing that is based upon predicting future system loads, rather than conventional demand-driven approaches that monitor current system state. We analyze a number of possible predictive dynamic load balancing protocols by isoefficiency analysis to guide the design of a parallel database rule processing system.     

遥感影像的高性能并行处理技术研究

随着空间遥感技术和对地观测技术的不断发展,光学、热红外和微波等不同技术手段可以获取同一地区的多种遥感影像数据（多时相、多光谱、多传感器、多平台和多分辨率等）,每天获取的遥感数据量越来越大。同时,大量的遥感应用需要快速地对这些遥感数据进行处理与分析,提供辅助决策信息。因此,如果不能及时进行数据处理,这些数据就会失去时效性,甚至失去数据本身的价值。高性能计算与并行处理技术,加速了遥感影像数据处理与信息提取的进度,如大规模多处理系统、网格与云计算技术、通用图形处理器（GPGPU）等。文中综述了高性能计算、并行处理及云计算技术应用于遥感领域的最新进展,给出了一些研究与应用范例,并提出了当前高性能遥感影像处理所面临的一些挑战。     

A Framework for Method-Specific Knowledge Compilation from Databases

Generality and scale are important but difficult issues in knowledge engineering. At the root of the difficulty lie two challenging issues: how to accumulate huge volumes of knowledge and how to support heterogeneous knowledge and processing. One approach to the first issue is to reuse legacy knowledge systems, integrate knowledge systems with legacy databases, and enable sharing of the databases by multiple knowledge systems. We present an architecture called HIPED for realizing this approach. HIPED converts the second issue above into a new form: how to convert data accessed from a legacy database into a form appropriate to the processing method used in a legacy knowledge system. One approach to this reformed issue is to use method-specific compilation of data into knowledge. We describe an experiment in which a legacy knowledge system called INTERACTIVE KRITIK is integrated with an ORACLE database. The experiment indicates the computational feasibility of method-specific data-to-knowledge compilation.     

Extensible Parallel Query Processing for Exploratory Geoscientific Data Mining

Exploratory data mining and analysis requires a computing environment which provides facilities for the user-friendly expression and rapid execution of scientific queries. In this paper, we address research issues in the parallelization of scientific queries containing complex user-defined operations. In a parallel query execution environment, parallelizing a query execution plan involves determining how input data streams to evaluators implementing logical operations can be divided to be processed by clones of the same evaluator in parallel. We introduced the concept of relevance window that characterizes data lineage and data partitioning opportunities available for an user-defined evaluator. In addition, we developed a query parallelization framework by extending relational parallel query optimization algorithms to allow the parallelization characteristics of user-defined evaluators to guide the process of query parallelization in an extensible query processing environment. We demonstrated the utility of our system by performing experiments mining cyclonic activity, blocking events, and the upward wave-energy propagation features from several observational and model simulation datasets.     

基于消息传递接口的大规模生物网络比对并行化算法

为有效降低生物网络比对算法的时间复杂度,提出一种基于可扩展的蛋白质相互作用网络比对(SPINAL)算法的消息传递接口(MPI)并行化实现方法。该方法将MPI并行化思想运用在SPINAL算法中,在多核环境中采用并行排序代替算法原本的排序方式,并结合负载均衡策略合理分配任务。实验结果表明,与未使用并行排序以及负载均衡策略相比,该方法在处理大规模生物网络比对时能有效地缩短计算时间,提高运算效率,对于不同组比对数据都有较为稳定的优化保障,具有良好的可扩展性。     

Distributed volume visualization: A step towards integrated data analysis and image synthesis

Integrated on-the-fly data analysis and image synthesis is one of the most dominant challenges on modern volume visualization tools. Since volume visualization algorithms are them self computationally complex and memory intensive there is hardly a chance to efficiently integrate data analysis tools on standard single processor architectures. The development and spreading of multiprocessor systems with large scale memory allow for the direct analysis and modification of the data during the rendering process. But to efficiently integrate both tasks similar parallelization strategies and a unique data layout must be chosen. Furthermore, since more and more systems with hardware assisted processing and visualization options can be accessed across high performance networks, distributed visualization tools should benefit from these options whenever possible. In the following paper a prototyped parallel visualization environment will be exemplified in which data analysis and image synthesis are simultaneously performed on demand. The objective is to outline basic parallelization and distribution aspects which enable flexible integration of different tasks rather than to present a concrete implementation. In this context the development and design of a unique algorithmic framework for integrated and/or distributed data analysis and image synthesis dominates the work hereafter.     

浅谈超大型Oracle数据库的基础设计和优化设计

Oracle数据库因其在处理大型数据时的优越性,而在商业领域应用的越来越广泛,成为目前比较流行的数据库平台之一,拥有越来越多的用户,已成为许多大型商业应用系统的后台数据库系统。但是在实际应用过程中,因为对Oracle的不熟悉或者了解不够,而出现一些问题,影响Oracle的使用,也体现不出Oracle在处理海量数据时的优越性。因此,从Oracle的基础设计和优化设计两个方面入手,将Oracle的优越性一一体现出来,并对容易忽视的问题尤其重点说明。     

Parallelism and Scalability in an Image Processing Application

The recent trends in processor architecture show that parallel processing is moving into new areas of computing in the form of many-core desktop processors and multi-processor system-on-chips. This means that parallel processing is required in application areas that traditionally have not used parallel programs. This paper investigates parallelism and scalability of an embedded image processing application. The major challenges faced when parallelizing the application were to extract enough parallelism from the application and to reduce load imbalance. The application has limited immediately available parallelism and further extraction of parallelism is limited by small data sets and a relatively high parallelization overhead. Load balance is difficult to obtain due to the limited parallelism and made worse by non-uniform memory latency. Three parallel OpenMP implementations of the application are discussed and evaluated. We show that with some modifications relative speedups in excess of 9 on a 16 CPU system can be reached.