曙光3000超级服务器设计的关键问题研究

曙光3000超级服务器是基于SMP机群体系结构的通用计算机系统，具有可扩展性，可用性，可管理性和高可用性的技术特点，该文着重介绍曙光3000系统设计中的若干关键问题，包括与SMP机群体系结构相关的可扩展性问题，系统软件中重要的可用性设计、底层通信的多种应用才机群管理系统的跨平台支持设计，另外还论述了超级服务器设计中存在的问题和作者的看法。     

曙光机群资源管理的设计与实现

机群资源管理负责机群系统内计算所需资源的管理，该文介绍机群资源管理的主要功能和重要的研究问题，着重介绍曙光3000机群系统的资源管理的设计与实现，包括分区管理，任务管理和通信管理功能及其对高可用性，可扩展性，动态特性的支持；并给出相应的性能分析结果，此外，还介绍了相关系统的资源管理和对未来该领域研究的方法。     

基于机群操作系统的并行调试器

并行调试工具的设计，是并行计算环境工具研究开发中的一个突出难点。介绍了一个在曙光3000上实现的并行调试器DCDB3．0。该调试器是未来曙光4000机群操作系统的一部分，是曙光3000上的第1个可运行版本，采用典型的客户／服务器模式。客户端的用户界面可将冗繁的调试信息与操作可视化。客户端可以远离提供服务的大型机，其远程通信依赖的是机群操作系统中的DRPC和任务管理，前者提供远程方法调用，后者使得客户端能够在服务器上启动相应的任务。DCDB3．0的服务器端负责处理调试任务和同客户端进行信息交互。DCDB3．0的功能具有可扩放性，使得可以在此平台上研究一些高级并行调试技术的实现。改进了已有的方式，实现了重放技术，并计划进一步添加其他高级并行调试技术。     

计算机辅助药物分子对接并行演化设计

对分子对接理论作了简单介绍,建立了一个基于柔性配体分子与刚性受体分子对接的数学规划模型,将分子对接中的构象优化搜索转化为求解约束极小化问题的过程,并采用带有空间收缩的多种群并行遗传算法进行求解．在分布式存储的并行机曙光3000上模拟计算表明,该设计具有很高的并行加速比,在保证分子对接的准确性和有效性的前提下,大大提高了分子构象搜索的速度．     

曙光1000大规模并行计算机系统软件的设计

曙光１０００是基于消息的大规模并行计算机系统。本文介绍曙光１０００系统软件的设计目标，重要的微核心，用户空间服务。客户／服务模型和标准用户界面的设计思想，实现技术路线，以及组织结构，并详细阐述了系统诊断与监控，系统管理，并行程序开发环境与工具和调试器的设计。     

基于高速通信协议的COSMOS机群文件系统性能研究

作为曙光3000超级服务器的重要组成部分,COSMOS机群文件系统对机群文件系统协议,结构及性能优化等问题进行全面深入的探讨,首先描述了基于曙光3000机群高速通协议BCL－3的COSMOS文件系统的实现,然后引入并发带宽利用率,描述了通信与I／O对机群文件系统性能影响程序,最后介绍了有关性能实验并对实验结果作出解释。     

基于分区租借的大规模作业管理系统

PWS(partitioned workload system)是引入分区化技术以及资源租借技术的作业管理系统(JMS)。分区技术加强了机群系统中资源的可管理性，通过资源租借技术提高了系统资源的利用率。优化的通信方式让PWS能够支持上千节点的大规模机群系统。PWS已经在曙光4000A超级计算机以及相关高性能计算领域中得到了应用。该文介绍了PWS的设计与实现。     

基于GPS的工务巡道作业监控系统的研究

提出一种基于GPS的巡道工巡道作业监控系统来解决铁路工务部门保证巡道作业质量的一种新思路，并对系统手持机的设计要求作了较充分的介绍，对上位机的软件应具备的功能作了阐述。现场实验数据表明，该系统的技术路线可行，有利于加强巡道作业的质量监控和责任追究。     

“曙光1000”并行机讲座 第一讲 “曙光1000”和中国成都高性能计算中心

经国家科委批准，中国成都高性能计算中心于１９９７年１１月１１日在西南交通大学举行隆重的建成挂牌仪式，国务院信息办、国家科委、国家教委、铁道部、电子部、中科院和省、市领导到会祝贺。省、市新闻单位对此作了及时的报道。为了充分利用中心的并行计算机资源，为全国，特别是我国西部地区的用户服务，本刊特开辟高性能计算机讲座，介绍“曙光１０００”的系统结构、性能和应用。讲座共三讲，本期刊出第一讲，除简介“曙光１０００”外，着重介绍如何申请国家高性能计算基金来使用“曙光１０００”为科研服务。下一讲的题目是：“曙光１０００”并行计算机编程环境。     

TC1700嫁出千套 曙光服务器举办产业成果大会

4月4日,伴随成都理工大学王院长与曙光总裁历军先生的握手,第1000套曙光TC1700高性能机群系统正式交付用户使用。曙光TC1700是曙光公司高端计算产品线中的一款机群产品,是863重大成果曙光3000超级服务器技术下移的成果,TC1700是首款跨操作系统平台和硬     

New classification of medical staff clinical services for optimal reconstruction of job workflow in a surgical ward: Application of spectrum analysis and sequence relational analysis

In order to optimize the job workflow of medical staff, clinical job workflow was investigated from the viewpoint of its periodicity and the strength of causal association among jobs. Time-motion study for the staff at a surgical ward was carried out. To detect the periodicity of the occurrence of each job element, its frequency histogram was determined, and the discrete Fourier transformation was applied. For the analysis on the strength of the relationship among the job-sequence, the sequence relational analysis was developed, which was the expansion of the relation analysis to the sequence process. The job elements were classified into five incident patterns based on the periodicity of each element and into three patterns based on the association with other job elements. Based on time-motion study data, job workflow patterns of medical staff’ were clarified based on the incident pattern of the job elements and the association with other job elements.     

网格计算中任务调度研究综述

阐述了网格计算领域任务调度的特点和目标;综述了现有的任务调度技术和算法,包括启发性智能任务调度,基于Agent的任务调度,基于Petri网的任务调度,基于成本的任务调度等算法,以及任务调度的负载均衡问题,最后给出任务调度的研究展望。     

New classification of medical staff clinical services for optimal reconstruction of job workflow in a surgical ward: Application of spectrum analysis and sequence relational analysis

In order to optimize the job workflow of medical staff, clinical job workflow was investigated from the viewpoint of its periodicity and the strength of causal association among jobs. Time-motion study for the staff at a surgical ward was carried out. To detect the periodicity of the occurrence of each job element, its frequency histogram was determined, and the discrete Fourier transformation was applied. For the analysis on the strength of the relationship among the job-sequence, the sequence relational analysis was developed, which was the expansion of the relation analysis to the sequence process. The job elements were classified into five incident patterns based on the periodicity of each element and into three patterns based on the association with other job elements. Based on time-motion study data, job workflow patterns of medical staff’ were clarified based on the incident pattern of the job elements and the association with other job elements.     

An efficient job scheduling algorithm for mixed turnaround and deadline applications

Computer job scheduling is often performed with little understanding of the formal properties of the jobs being scheduled. One reason for this is that optimal solutions for job scheduling on computers are difficult to obtain if the job stream has mixed objectives, i.e., it consists of some jobs whose turnaround time has to be minimized and others whose deadlines must be met. A practical algorithm for scheduling mixed job streams on monoprogrammed computers, with potential application to a multiprogramming environment, is presented. The algorithm takes into account variable cost rates for each job. Experimental results illustrate the efficiency of the algorithm in terms of both its proximity to optimal solutions and its low computational complexity.     

面向司法专题分析的大数据计算资源调度方法

针对司法专题分析过程中面临的交互式分析类数据处理执行效率低的问题,提出了一种基于任务类型的计算资源调度方法,为任务类型建立计算资源配额管理机制。在类型配额内具备抢占式优先调度权,在类型配额外可以借用其他任务类型的空闲资源。实验与分析表明,该方法能够在兼顾普通大数据处理任务执行效率的前提下显著提升交互式分析类任务的执行效率。     

基于Hopfield神经网络的作业车间生产调度方法

该文提出了基于Hopfield神经网络的作业车间生产调度的新方法.文中给出了作业车 间生产调度问题(JSP)的约束条件及其换位矩阵表示,提出了新的包括所有约束条件的计算能 量函数表达式,得到相应的作业车间调度问题的Hopfield神经网络结构与权值解析表达式,并 提出相应的Hopfield神经网络作业车间调度方法.为了避免Hopfield神经网络容易收敛到局部 极小,从而产生非法调度解的缺点,将模拟退火算法应用于Hopfield神经网络求解,使Hopfield 神经网络收敛到计算能量函数的最小值0,从而保证神经网络输出是一个可行调度方案.该文 改进了已有文献中提出的作业调度问题的Hopfield神经网络方法,与已有算法相比,能够保证 神经网络稳态输出为可行的作业车间调度方案.     

A metaheuristic for the fixed job scheduling problem under spread time constraints

In a scheduling problem, a job is said to be fixed when its due date corresponds exactly to its release date plus its processing time. This paper addresses the fixed job scheduling problem where processors are subject to spread time constraints, i.e., the amount of time spent between the starting time of the first job on a processor and the completion time of the last job on the same processor should not exceed a given duration. Existing exact approaches are tested on medium size instances. As large instances are clearly intractable with these approaches, a greedy heuristic and a grouping genetic algorithm are proposed. Computational results show the effectiveness of the proposed heuristics.     

航班过站地面服务的优化调度算法

航班过站服务流程是定位型和零工型的混合流程，其调度问题是一个有时间窗和作业调整时间的多目标多设备并行作业动态排序问题．在分析其区别于一般制造业生产作业排序特点的基础上，给出一个考虑了不同设备加工能力的新的启发式算法——设备能力差分配法．通过对服务作业分类，将多目标优化问题转化为服务类作业排序最优化问题．算例分析显示，能力差分配算法在减少航班延误数量、时间上以及平衡设备生产能力上均优于现有的先到先服务和最小负荷调度算法．     

Computational algorithms for networks of queues with rejection blocking

Summary Open, closed and mixed queueing networks with reversible routing, multiple job classes and rejection blocking are investigated. In rejection blocking networks blocking event occurs when upon completion of its service of a particular station's server, a job attempts to proceed to its next station. If, at that moment, its destination station is full, the job is rejected. The job goes back to the server of the source station and immediately receives a new service. This is repeated until the next station releases a job and a place becomes available. In the model jobs may change their class membership and general service time distributions depending on the job class are allowed. Two station types are considered: Either the scheduling discipline is symmetric, in which case the service time distributions are allowed to be general and dependent on the job class or the service time distributions at a station are all identical exponential distributions, in which case more general scheduling disciplines are allowed. An exact product form solution for equilibrium state probabilities is presented. Using the exact product form solution of the equilibrium state distribution, algorithms for computation of performance measures, such as mean number of jobs and throughputs, are derived. The complexity of the algorithms is discussed.