作者:
BOOTH, SPhysics Department
The University of Edinburgh The King's Buildings Edinburgh EH9 3JZ Scotland
The structure of the i860 is reviewed briefly. The code generation strategies used by compilers and assembly language programmers are compared. The problems associated with generating efficient code for the i860 are h...
The structure of the i860 is reviewed briefly. The code generation strategies used by compilers and assembly language programmers are compared. The problems associated with generating efficient code for the i860 are highlighted and a case study from QCD is used to demonstrate these points. Timing results for the case study are presented.
Although software tools already have a place on serial and vector computers they are becoming increasingly important for parallel computing. Message passing libraries, parallel operating systems and high level paralle...
Although software tools already have a place on serial and vector computers they are becoming increasingly important for parallel computing. Message passing libraries, parallel operating systems and high level parallel languages are the basic software tools necessary to implement a parallel processing program. These tools up to now have been specific to each parallel computer system and a short survey will be given. The aim of another class of software tools for parallelcomputers is to help in writing or rewriting application programs. Because automatic parallelization tools are not very successful, an interactive component has to be incorporated. We will concentrate here on the discussion of SPEFY, a parallel program development facility.
Quantum chemistry has become an essential tool in many areas of chemical research;however, quantum chemistry is not yet playing a role in many exciting new chemical disciplines, such as medicinal chemistry and materia...
详细信息
Quantum chemistry has become an essential tool in many areas of chemical research;however, quantum chemistry is not yet playing a role in many exciting new chemical disciplines, such as medicinal chemistry and materials science, where the size of the chemical systems has been too large to study using ab initio chemical methods. The development of massivelyparallel supercomputers offers the potential to predict properties relevant to a variety of problems in these burgeoning new fields. The goal of this project is to develop a set of parallelized ''production codes'' for initially a relatively limited set of methods. As a key part of this project we are experimenting with the use of modern programming languages and methodologies to make these programs both portable and reusable. This paper describes the development of a massivelyparallel direct SCF program, MPSCF. For systems over a few hundred basic functions, MPSCF running on 256 nCUBE processors performs nearly as well as Gaussian 90 running on a single processor Cray Y-MP. On the next generation of parallelcomputers, such as the Intel Touchstone Delta, MPSCF should allow the SCF calculations on chemical systems too large for vector supercomputers.
Magnetic monopole currents are believed to have a strong impact on the Monte Carlo dynamics of compact quantum electrodynamics near the phase transition. We have indications that updating algorithms have to form and t...
详细信息
Magnetic monopole currents are believed to have a strong impact on the Monte Carlo dynamics of compact quantum electrodynamics near the phase transition. We have indications that updating algorithms have to form and to break up monopole current clusters in order to force the tunneling of the system between the coexisting phases, the confined phase and the Coulomb phase. This might be responsible for long lived metastabilities in the time series. We present parallel numerical algorithms which allow the fast identification of clusters of monopole loops, the determination of their topological properties as well as their visualization in the course of a Monte Carlo simulation. The algorithms are implemented on the Connection Machine CM-5. We compare the data parallelprogramming style with a message passing approach. Complexity and performance are discussed.
The increased computational power of massivelyparallelcomputers and high bandwidths low latency computer networks will make a wide range of previously unpractical problems feasible. This will inevitably result in th...
详细信息
ISBN:
(纸本)0818643404
The increased computational power of massivelyparallelcomputers and high bandwidths low latency computer networks will make a wide range of previously unpractical problems feasible. This will inevitably result in the need to develop parallel software whose complexity far exceeds that of parallel programs being developed today. These programs will combine task and data parallelism within a single application. In this workshop, we will discuss multi-paradigm parallel programs and programming languages to support their development.
For a long time efficient use of massivelyparallelcomputers has been hindered by dependencies introduced in software through low-level implementation practice. This paper presents a programming environment and langu...
详细信息
The article deals with rationale and concepts of a programming model for massive parallelism. We mention the basic properties of massivelyparallel applications and develop a programming model for data parallelism on ...
详细信息
The article deals with rationale and concepts of a programming model for massive parallelism. We mention the basic properties of massivelyparallel applications and develop a programming model for data parallelism on distributed-memory computers. Its key features are a suitable combination of homogeneity and heterogeneity aspects, a unified representation of data point configuration and interconnection schemes by explicit virtual data topologies, and various synchronization schemes and nondeterminisms. The outline of the linguistic representation and the abstract executional model are given.< >
parallelprogramming is a difficult task involving many complex issues such as resource allocation, and process coordination. We propose a solution to this problem based on the use of a repertoire of parallel algorith...
详细信息
parallelprogramming is a difficult task involving many complex issues such as resource allocation, and process coordination. We propose a solution to this problem based on the use of a repertoire of parallel algorithmic forms, known as skeletons. The use of skeletons enables the meaning of a parallel program to be separated from its behaviour. Central to this methodology is the use of transformations and performance models. Transformations provide portability and implementation choices, whilst performance models guide the choices by providing predictions of execution time. We describe the methodology and investigate the use and construction of performance models by studying an example.< >
Data parallelism is a powerful approach to parallel computation, particularly when it is used with complex data types. Categorical data types are extensions of abstract data types that structure computations in a way ...
详细信息
Data parallelism is a powerful approach to parallel computation, particularly when it is used with complex data types. Categorical data types are extensions of abstract data types that structure computations in a way that is useful for parallel implementation. In particular, they decompose the search for good algorithms on a data type into subproblems, all homomorphisms can be implemented by a single recursive, and often parallel, schema, and they are equipped with an equational system that can be used for software development by transformation.< >
暂无评论