Abstract not found

We review the main research issues and problems in the area of general-purpose, massively parallel MIMD computers. It will be shown that the successful exploitation of this class of systems depends on the ability to define and experiment with design methodologies taking into account several aspects: abstract machine architecture, physical machine architecture, parallelizing and restructuring tools, and machine independent parallel languages. We report on our research experience in this area, and in particular our work in the Pisa Parallel Processing Project. 1 Introduction In spite of the increasing interest in parallel processing during the last decade, the widespread use of this technology is still thwarted by a lack of highly parallel computers that are general-purpose. As for sequential computing, the formation of a large community of users requires easy and high-level programmability, software modularity and portability. To achieve widespread acceptance of parallel computers, sev...

Structured parallel programming is one of the possible solutions to exploit Programmability, Portability and Performance in the parallel programming world. The power of this approach stands in the possibility to build an optimizing template-- based compiler using low time complexity algorithms. In order to optimize the code, this compiler needs formulas that describe the performance of language constructs over the target architecture. We propose a set of parameters able to describe current parallel systems and build deterministic analytical models for basic forms of parallelism. The analytical model describes construct performance in a parametric way.This can be done by knowing that the compiler exploits a template--based support and giving template implementors guidelines to follow to make actual implementation perform as predicted. ACM--CR Subject Classification: Keyword and phrases: Skeletons, performance modeling, parallel languages, template-- based compilers Machine...

Structured parallel programming is one of the possible solutions to exploit Programmability, Portability and Performance in the parallel programming world. Programming using high level parallel constructs permits the programmer to focus on the development of the parallel algorithms rather than on their low level implementation. The power of this approach stands in the possibility of modeling the performance of the high level constructs and building optimizing template-based compilers. These compilers use formulas describing the performance of language constructs over the target architectures and tune their parametric implementation to achieve high performance figures. In this paper, we describe a set of applications developed using such a programming methodology and show the results obtained. ACM--CR Subject Classification: Keyword and phrases: parallel programming skeletons languages Design and evaluation of parallel applications using a structured parallel language David...

With the term distributed system we mean a system with no centralized, shared memory. This definition is very broad: many real life systems, like computer networks, or distributed memory multiprocessors, falling into it. In spite of this generality, it is possible to produce abstract models capturing the essential features shared by all distributed systems, ignoring the unimportant physical details. In this project we address two specific problems, namely packet scheduling and clock synchronization, formulated on general models of distributed systems. This allows to apply the obtained results in a wide variety of situations. Keywords: distributed systems, message and process scheduling, synchronization, algorithms, complexity 1 State of the Arts and Trends The dramatic growth in number of powerful, easy-to-use, portable, and affordable computers, combined with globally accessible communication networks, has resulted in a large and growing user community which demands for sop...

This paper presents a parallel programming methodology that ensures easy programming, efficiency, and portability of programs to different machines belonging to the class of the general-purpose, distributed memory, MIMD architectures. The methodology is based on the definition of a new, high-level, explicitly parallel language, called P^3L, and of a set of static tools that automatically adapt the program features for each target architecture. P^3 L does not require programmers to specify process activations, the actual parallelism degree, scheduling, or interprocess communications, i.e. all those features that need to be adjusted to harness each specific target machine. Parallelism is, on the other hand, expressed in a structured and qualitative way, by hierarchical composition of a restricted set of language constructs, corresponding to those forms of parallelism that are frequently encountered in parallel applications, and that can be efficiently implemented. The eff...

The syntax, semantics and pragmatics of the P³L programming language are described here. The version of the language discussed is the P³L version 1.0. Some features have been left out that will appear in future versions of the language. The P³L described here uses C++ has "host language". However, many of the arguments and topics discussed in the report apply independently of the host language chosen for the implementation.

This paper presents a parallel programming methodology that ensures easy programming, e ciency, and portability of programs to di erent machines belonging to the class of the generalpurpose, distributed memory, MIMD architectures. The methodology is based on the de nition of a new, high-level, explicitly parallel language, called P 3 L, and of a set of static tools that automatically adapt the program features for each target architecture. P 3 L does not require programmers to specify process activations, the actual parallelism degree, scheduling, or interprocess communications, i.e. all those features that need to be adjusted to harness each speci c target machine. Parallelism is, on the other hand, expressed in a structured and qualitative way, by hierarchical composition of a restricted set of language constructs, corresponding to those forms of parallelism that are frequently encountered in parallel applications, and that can e ciently be implemented. The e cient portability ofP 3 L applications is guaranteed by the compiler along with the novel structure of the support. The compiler automatically adapts the program features for each speci c architecture, accessing the costs (in terms of performance) of the low-level mechanisms exported by thearchitecture itself. In our methodology, these costs, along with other features of