Reinventing explicit parallel programming for improved engineering of high performance computing software

Bangalore, Purushotham; Skjellum, Anthony; Gray, Jeff; Bryant, Barrett R.

doi:10.1049/ic:20040421

Cited by 6 publications

(4 citation statements)

References 18 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The # model was proposed to meet the demands of the high performance computing community for integrating parallel programming techniques with modern large scale software engineering techniques [26]. In order to demonstrate the # model ability to work with separation of encapsulated and cross-cutting concerns, three common examples from parallel programming practice were used.…”

Section: Discussionmentioning

confidence: 99%

“…In the second phase, research efforts have been coordinated in order to standardize parallel programming languages and libraries, yielding the development of PVM (Parallel Virtual Machine) [16], MPI (Message Passing Interface) [20], HPF (High Performance Fortran) [14], PETSc [2], ScaLAPACK, etc. The expansion in scale of applications of parallel computing, due to cluster and grid computing, has made the HPC community to search for parallel programming models and languages that reconcile generality (G), high level of abstraction (A), portability (P) and efficiency (E), allowing to apply advanced software engineering concepts into the program development process [26]. The GAPE requirements have not been reconciled by current programming techniques despite the efforts that began fifteen years ago with CRPC [13,4].…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

The # model

Carvalho¹,

Lins

2005

Proceedings of the 2005 ACM Symposium on Applied Computing

View full text Add to dashboard Cite

The computer science community has been looking for parallel languages and models with a higher level of abstraction and modularity, without performance penalties, that could be used in conjunction with advanced software engineering techniques, and that are suitable to work with large-scale programs. This paper discusses how the # parallel programming model addresses the issues of modularity and abstraction of parallel programs using the techniques of separation of concerns.

show abstract

Section: Discussionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

The # model

Carvalho¹,

Lins

2005

Proceedings of the 2005 ACM Symposium on Applied Computing

View full text Add to dashboard Cite

show abstract

“…As a specific application area, we believe that there is a strong potential for impact if aspects can be used to improve the modularization of template libraries tailored for high-performance scientific computing. Such applications rely heavily on parametric polymorphism to specialize mathematical operations on vectors, arrays, and matrices [16], [18]. This paper presents the initial design and implementation of a template-aware aspect language.…”

Section: Motivationmentioning

confidence: 99%

A Program Transformation Technique to Support AOP within C++ Template.

Roychoudhury¹,

Gray²,

Zhang³

et al. 2010

JOT

View full text Add to dashboard Cite

Aspect-oriented programming (AOP) provides assistance in modularizing concerns that crosscut the boundaries of system decomposition. Aspects have the potential to interact with many different kinds of language constructs in order to modularize crosscutting concerns. Although several aspect languages have demonstrated advantages in applying aspects to traditional modularization boundaries (e.g., object-oriented hierarchies), additional language concepts such as parametric polymorphism can also benefit from aspects. Many popular programming languages support parametric polymorphism (e.g., C++ templates), but the combination of aspects and generics is a topic in need of further investigation. The paper enumerates the general challenges of uniting aspects with C++ templates. It also underlines the need for new language constructs to extend AOP support to C++ templates and provides an initial solution to realize this goal.

show abstract

“…The software development process using explicit parallelization leaves little scope for code reuse because it involves ad-hoc design decisions [4]. In case a programmer wants to set-up communication between the processors, there are multiple options available, each with specific trade-offs (e.g., synchronous/asynchronous, point-topoint/one-sided/collective) but there are no well-established rules or design patterns to select one option over the other.…”

Section: Introductionmentioning

confidence: 99%

Raising the level of abstraction for developing message passing applications

2010

View full text Add to dashboard Cite

Message Passing Interface (MPI) is the most popular standard for writing portable and scalable parallel applications for distributed memory architectures. Writing efficient parallel applications using MPI is a complex task, mainly due to the extra burden on programmers to explicitly handle all the complexities of message-passing (viz., inter-process communication, data distribution, load-balancing, and synchronization). The main goal of our research is to raise the level of abstraction of explicit parallelization using MPI such that the effort involved in developing parallel applications is significantly reduced in terms of the reduction in the amount of code written manually while avoiding intrusive changes to existing sequential programs. In this research, generative programming tools and techniques are combined with a domainspecific language, Hi-PaL (High-Level Parallelization Language), for automating the process of generating and inserting the required code for parallelization into the existing sequential applications. The results show that the performance of the generated applications is comparable to the manually written versions of the applications, while requiring no explicit changes to the existing sequential code.

show abstract

Reinventing explicit parallel programming for improved engineering of high performance computing software

Abstract: Abstract

Cited by 6 publications

References 18 publications

The # model

The # model

A Program Transformation Technique to Support AOP within C++ Template.

Raising the level of abstraction for developing message passing applications

Contact Info

Product

Resources

About