Towards a Uniied Parallel Architecture Class
نویسنده
چکیده
Traditionally, parallel programming paradigms are associated with only a single class of parallel architecture. This has resulted in a lack of portability of software between parallel architectures, also a lack of portability of programmers' skills. This report describes the author's work in unifying the basic parallel architecture classes with the aim of producing a single, uniied parallel architecture class. Such a uniied class would be capable of supporting all the parallel programming paradigms traditionally associated with the individual parallel architecture classes uniied to form the single parallel architecture class. The results of this uniication process demonstrate that all parallel architecture classes are capable of eeciently simulating the computation (control methods and memory access) of all other parallel architecture classes, communication can not be so simulated eeciently due to certain architecture classes being non-scalable in terms of bandwidth.
منابع مشابه
Further Towards a Uni cation of
Traditionally, parallel programming paradigms are associated with only a single class of parallel architecture. No single model of parallel computation has been able to span the range of parallel architecture classes. This has resulted in a lack of portability of software between parallel architectures, also a lack of portability of programmers' skills. This report describes the latest developm...
متن کاملScalable Parallel Computing: A Grand Unified Theory and its Practical Development
The Bulk Synchronous Parallel (BSP) model provides a uniied framework for the design and programming of general purpose parallel computing systems. In this paper we describe the BSP model and discuss some of the developments in architecture, algorithms and programming languages which are currently being pursued as part of this new, uniied approach to scalable parallel computing.
متن کاملUniied Formulation of a Wide Class of Scalable Interconnection Networks Based on Recursive Graphs
In this paper, we propose two new families of scalable interconnection networks called recursive replacement networks (RRN) and recursively linked networks (RLN) for large-scale parallel and distributed processing. RRN and RLN are formulated hierarchically using the notion of recursive graph (RG) and are characterized mainly by their inter-cluster graphs. RGs not only generate a wide class of h...
متن کاملSystematic synthesis of parallel architectures for the computation of higher order cumulants
Fine granularity parallel architectures for the eecient estimation of Higher Order Statistics are systematically derived in this paper. A uniied methodology for constructing Locally Recursive Algorithms and Space-Time linear mapping operators that lead to highly pipelined architectures consisting of multiple, tightly coupled array stages is discussed rst. Then a farm of processors is synthesize...
متن کاملTowards Mixed Computation / Communication in
This paper presents an overlapping technique of communications by computations based on pipelined communications. This allows to improve the execution time of most parallel numerical algorithms. Some simple examples are developed namely, global combine, matrix-vector product and bi-dimensional Fast Fourier Transform, to illustrate the eeciency of this technique. Moreover, we propose an uniied f...
متن کاملذخیره در منابع من
با ذخیره ی این منبع در منابع من، دسترسی به آن را برای استفاده های بعدی آسان تر کنید
عنوان ژورنال:
دوره شماره
صفحات -
تاریخ انتشار 1994