It becomes essential to use Batch jobs, run through Batcon, to process a significant amount of repetitive work. Experience has shown that much effort is still required to investigate jobs which fail and have to be cor...
详细信息
It becomes essential to use Batch jobs, run through Batcon, to process a significant amount of repetitive work. Experience has shown that much effort is still required to investigate jobs which fail and have to be corrected and rerun. While most error situations can be anticipated and allowed for in a control file, effective handling is limited because Batcon allows sensing and action upon either an error or noerror condition. A small change to Batcon is proposed which allows differentiation between errors and much improved analysis of error conditions.
A technique is described for using a parallel array computer to perform searching operations in a hash indexed table. Some evaluations of the efficiency of this technique are given.
A technique is described for using a parallel array computer to perform searching operations in a hash indexed table. Some evaluations of the efficiency of this technique are given.
The paper presents a case study on the design of communication protocols using formalized design patterns. On principle, the approach adopts the well-known design patterns concept for reusing design solutions. However...
详细信息
ISBN:
(纸本)0818683775
The paper presents a case study on the design of communication protocols using formalized design patterns. On principle, the approach adopts the well-known design patterns concept for reusing design solutions. However a formal description technique (FDT) is applied as design language and therefore knowledge about pattern instantiation or its impact on the embedding context can be precisely specified. We consider formalization to be a prerequisite for increased correctness of resulting products and tool-supported pattern application. The FDT of our choice is the specification and description language SDL. Thus pattern application follows strict rules ill terms of the SDL syntax. Additionally, assumptions and effects stated in terms of the SDL semantics help to ensure the intended working of an SDL-pattern,within the embedding context. An SDL-pattern based re-engineering of the Internet Stream Protocol ST2+ is presented. In particular it is demonstrated how SDL-pattern based design can boost the developer's confidence in the resulting product.
We present the concept of the SOMA workflow developed at the Finnish IT Center for Science CSC. The SOMA workflow unites multiplatform UNIX/LINUX computing resources and third-party software for calculating molecular ...
详细信息
We present the concept of the SOMA workflow developed at the Finnish IT Center for Science CSC. The SOMA workflow unites multiplatform UNIX/LINUX computing resources and third-party software for calculating molecular structure and properties. The presented workflow components consist of the computing program XML descriptions, the core workflow program Grope, the toolkit for parsing program input and output, and the extranet interface. The program Grape and the developed XML descriptions of scientific programs allow researchers to link molecular modeling software into highly sophisticated computational workflows. SOMA collects the calculated data produced by the workflow and stores the computed information in the Chemical Markup Language (CML) format. The extranet interface is used for user authentication, building of the program interfaces and the workflows, and for sorting, filtering, and visualizing the results.
This paper introduces a methodology to model the execution time of several computation and communication routines developed in the frame of the CrossGrid project. The purpose of the methodology is to provide performan...
详细信息
ISBN:
(纸本)3540269185
This paper introduces a methodology to model the execution time of several computation and communication routines developed in the frame of the CrossGrid project. The purpose of the methodology is to provide performance information about some selected computational kernels when they are executed in a grid. The models are based on analytical expressions obtained from exhaustive monitorized measurements. Even though the kernels that are considered in this work include both applications dependent and general purpose, the methodology can be applied to any kind of kernel in which the most relevant part in terms of execution time is due to computations and/or communications. We focused on MPI-based communications. In addition, an interactive Graphical User Interface was developed to summarize and show the information provided by the models from different views.
The paper studies the problems of definition, modification and control of input-output systems (including automatic emergency takeover). The organization and the control of distributed input-output systems are present...
详细信息
The paper studies the problems of definition, modification and control of input-output systems (including automatic emergency takeover). The organization and the control of distributed input-output systems are presented followed by the aspect of language definition.
General task scheduling for uniprocessor systems is an NP-hard problem. Many special cases, however, have simple optimal dynamic scheduling policies. Multiprocessor systems have NP-hard scheduling problems even for si...
详细信息
General task scheduling for uniprocessor systems is an NP-hard problem. Many special cases, however, have simple optimal dynamic scheduling policies. Multiprocessor systems have NP-hard scheduling problems even for simple special cases. This paper presents necessary conditions for optimal multiprocessor which produces only schedules meeting the necessary conditions. Experiments evaluating the algorithm are discussed. Further research directions leading to distributed implementations of multiprocessor scheduling policies are presented.
A job model which represents overlapping tasks within individual jobs is constructed and then extended so as to represent a system model. The system model is further refined so as to include two different CPU scheduli...
详细信息
A job model which represents overlapping tasks within individual jobs is constructed and then extended so as to represent a system model. The system model is further refined so as to include two different CPU scheduling algorithms; nonpreemptive and processor-sharing. Having a high degree of overlap within a job appears to be equivalent to the addition of another job in the system when nonpreemptive algorithms are used. When a processor-sharing algorithm is used, a high degree of overlap within a job is approximately equivalent to doubling the number of jobs in the system.
The Dartmouth Time Sharing System (DTSS) had a mean time between failures of 41 hours for the 1972 calendar year. During this period, the system was available for 6047 hours, and over 4. 3 million jobs were run for 13...
详细信息
The Dartmouth Time Sharing System (DTSS) had a mean time between failures of 41 hours for the 1972 calendar year. During this period, the system was available for 6047 hours, and over 4. 3 million jobs were run for 13,000 users at 50 colleges and high schools in the U. S. and Canada. The purpose of this paper is to describe some of the techniques used to achieve a high degree of service reliability in a system undergoing continual development and improvement. The first part of the paper provides a brief description of DTSS and defines the concepts of ″jobs″ and ″failures″ as used in this environment. Subsequent parts of the paper describe the software techniques and management strategy used to achieve system reliability objectives.
The Finite Element Machine is a special purpose computer designed from 1024 microprocessors communicating by way of explicit word-serial channels. The algorithms for structural analysis using the finite element method...
详细信息
The Finite Element Machine is a special purpose computer designed from 1024 microprocessors communicating by way of explicit word-serial channels. The algorithms for structural analysis using the finite element method which have been proposed for this machine require both maximum and summation over a set of numbers stored one per processor. Three distinct types of sum and maximum algorithm making different use of the communications paths have been formulated and analyzed with respect to execution time. The results give some insight into the best way to use communications paths in a multiprocessor. The study demonstrated the need for a special hardware mechanism to support the global sum and maximum operations in this machine. The design of this hardware unit is also discussed.
暂无评论