A prototyping approach to evaluating the performance of parallel anddistributedsoftware architectures is proposed. The approach utilizes conventional object-oriented modelling to describe the system requirements. Al...
详细信息
A prototyping approach to evaluating the performance of parallel anddistributedsoftware architectures is proposed. The approach utilizes conventional object-oriented modelling to describe the system requirements. Alternative architectural approaches are then derived and expressed in the PARSE process graph notation, and performance prototyping and formal validation tools allow various aspects of the proposed solution to be rapidly explored. This paper describes work-in-progress on the HL language and prototyping environment which enables designers to express architectural structures in a simple language, and provides an execution environment which allows the potential performance of the architecture to be investigated. The paper gives an overview of the language and tools, and illustrates its application with an example.
In this paper, we discuss the problem of developing performance-oriented software and the need for methodologies. We then present the EDPEPPS (Environment for Design and Performance Evaluation of Portable parallel Sof...
详细信息
In this paper, we discuss the problem of developing performance-oriented software and the need for methodologies. We then present the EDPEPPS (Environment for Design and Performance Evaluation of Portable parallelsoftware) approach to the problem of designing and evaluating high-performance (parallel) applications. The EDPEPPS toolset is based on a rapid prototyping philosophy, where the designer synthesizes a model of the intended software which may be simulated, and the performance is subsequently analyzed using visualization. The toolset combines a graphical design tool, a simulation facility, and a visualization tool. The same design is used to produce a code suitable for simulation and real execution.
Dynamic reconfiguration is the ability to modify a parallel or distributed system while it is running. We adopt the framework developed by Jeff Kramer and colleagues at the system architecture level: changes must occu...
详细信息
Dynamic reconfiguration is the ability to modify a parallel or distributed system while it is running. We adopt the framework developed by Jeff Kramer and colleagues at the system architecture level: changes must occur in a consistent state, which is brought about by `freezing' some system components. The goal is to reduce system disruption, i.e., to minimize. 1. the part of the system to be `frozen' and 2. the time taken by reconfiguration operations. Towards the first goal we take a connection based approach instead of a component based one. To reduce time, we refine the reconfiguration algorithm by executing changes in parallel as much as possible. Our model also handles hierarchic systems.
In order to be truly portable, a program must be tolerant of a wide range of development and execution environments, and a parallel program is just one which must be tolerant of a very wide range. First, the term `tol...
详细信息
In order to be truly portable, a program must be tolerant of a wide range of development and execution environments, and a parallel program is just one which must be tolerant of a very wide range. First, the term `tolerant programming' is defined. Then, a formal model called F-Nets is described in which parallel algorithms are expressed as folded partial-orderings of operations, and this is argued to provide a suitable framework for building tolerant programs. Finally, software Cabling (SC), a very-high-level graphical programming language, demonstrates how many of the features normally expected from today's computer languages (e.g. data abstraction and data parallelism) can be obtained within the F-Net paradigm.
Performance comparison between image compression methods is an activity difficult to realize because of the few standard environments for validation. The objective of this paper is to present OPENPRESS as a single ope...
详细信息
Performance comparison between image compression methods is an activity difficult to realize because of the few standard environments for validation. The objective of this paper is to present OPENPRESS as a single open platform for validation allowing the integration of most of the current and future image compression methods, to facilitate their comparison and validation on a common and coherent set of reference images. Another goal of OPENPRESS is to ensure a broad distribution of these compression methods to the potential users. For that purpose, the proposed environment involves advanced computer science techniques like distributed computing, platform-independent softwareengineering, client/server decomposition, etc.
Techniques for automatic program recognition, at the algorithmic level, could be of high interest for the area of Program parallelization, because the selection of suitable parallelization strategies is driven by algo...
详细信息
Techniques for automatic program recognition, at the algorithmic level, could be of high interest for the area of Program parallelization, because the selection of suitable parallelization strategies is driven by algorithmic features of the code. In this paper a technique for the specification and automatic recognition of algorithmic concepts is presented. Its flexibility and expressivity power to specify the hierarchy, the constraints and the relationships among concepts allow it to deal with recognition of algorithmic concepts within optimized code, irregular computations, and in presence of code sharing, delocalization, implementation variations and other problems related to program recognition in the context of the imperative languages typically used for scientific computation.
CSP distributed programming model offers formal bases that allow the development of high quality and secure distributedsoftware for multicomputers. However, the model imposes restrictions that make it complicated to ...
详细信息
CSP distributed programming model offers formal bases that allow the development of high quality and secure distributedsoftware for multicomputers. However, the model imposes restrictions that make it complicated to reuse software components developed with languages that follow the model. This problem is mainly due to each component in the application interacting with each other according to a static communication channel structure. This fact complicates the reuse of a component in a different context from the original. In this work, a methodology for designing distributedsoftware is presented. This methodology allows to build distributed applications in a structured manner by layering instances of highly parameterized components. Also, a rotation to describe such components supporting the presented methodology is proposed. This methodology, together with the aforementioned notation, can be easily implemented in any programming language that follows the CSP model.
There has been an increasing emphasis on formality in software system specification in the last few years. A number of standards bodies are recommending the use of formal notations for specifying softwaresystems. Par...
详细信息
There has been an increasing emphasis on formality in software system specification in the last few years. A number of standards bodies are recommending the use of formal notations for specifying softwaresystems. parallel anddistributedsystems have their own complex features such as: the concurrent interactions between various system components;the reactive nature of the systems;various message passing schemes between system components. Object-Z is an extension to the Z language specifically to facilitate specification in an object-oriented style. Because parallel anddistributedsystems are typically complex systems, the extra structuring afforded by the various Object-Z modelling constructs (i.e. the class, object containment constructs, and various composite operation expressions) enables the various hierarchical relationships and the communication between system components to be succinctly specified. Object-Z history invariants allow system temporal properties to be specified as well. In this paper, the use of Object-Z in the specification of parallel anddistributedsystems is demonstrated by presenting a case study based on a multi-lift system. To enhance the understandability of the formal model, OMT notation is used to grasp the static structure of the system, and a finite state machine diagram is used to highlight the system behaviour.
A number of MP interfaces are available today but one of the most popular is the PVM software package. PVM permits the user to configure his own virtual computer by hooking together a heterogeneous collection of UNIX ...
详细信息
A number of MP interfaces are available today but one of the most popular is the PVM software package. PVM permits the user to configure his own virtual computer by hooking together a heterogeneous collection of UNIX based machines on which the user has a valid login and which are accessible over some network. To provide high-level graphical support for PVM based program development, an integrated programming environment (GRADE) is being developed. GRADE currently provides tools to construct, execute, debug, monitor and visualize message-passing based parallel programs. GRADE offers the programmer the same graphical user interface during the whole development process and high-level graphical programming abstraction mechanisms to construct parallel applications.
Monitoring and debugging parallel programs is a difficult activity. There are many situations where the traditional `stop the world, I want to get off' approach to debugging is simply unsuitable. Frequently, nonin...
详细信息
Monitoring and debugging parallel programs is a difficult activity. There are many situations where the traditional `stop the world, I want to get off' approach to debugging is simply unsuitable. Frequently, nonintrusive monitoring of the program execution is more productive in locating sources of error and also in monitoring `correct' programs for such purposes as performance measurement and tuning. This paper presents a number of space- and time-efficient tools and techniques to support nonintrusive, non-stop monitoring and debugging of parallel programs running on a shared-memory multiprocessor. The techniques include the use of spy tasks, circular history buffers, vectors of use bits, and data structure audits. Particular emphasis is placed on issues that pertain to parallel computing, such as dealing with concurrent execution, shared memory and data caches.
暂无评论