the problem of partitioning irregular graphs for parallel computations on homogeneous systems has been extensively studied. However these solutions fail when the target system architecture exhibits heterogeneity in re...
this paper reports on our experiments on using the Infineon TriCore as a building block for a multimedia processor. the experiments aim to obtain a highperformance processor using two strategies: integrating multimed...
详细信息
ISBN:
(纸本)0769517919
this paper reports on our experiments on using the Infineon TriCore as a building block for a multimedia processor. the experiments aim to obtain a highperformance processor using two strategies: integrating multimedia units into the TriCore CPU and constructing the TriCore in multiprocessor configuration. the design and implementation of the multimedia units for video, audio, and text compressions are discussed. Two hardware architectures for IMA ADPCM audio compression multimedia unit were designed: direct architecture and sequential architecture. the multimedia unit for text compression is based on a modification from another design;our design uses a more efficient timing operation and has a better hardware utilization than the original design. Two algorithms for parallel motion-estimation were implemented on the multiple TriCore system. the results show that the TriCore is a good building block for a multiprocessor system.
this paper introduces a new user-level communication protocol designed to provide high-performance data transfers across high-bandwidth, high-delay, networks. the protocol incorporates the most important enhancements ...
Summary form only given. this paper focuses on theoretical and practical aspects of the high-performance multikey sorting problem on computer clusters, with particular emphasis on the Alpha Maci Cluster, a world-class...
详细信息
Summary form only given. this paper focuses on theoretical and practical aspects of the high-performance multikey sorting problem on computer clusters, with particular emphasis on the Alpha Maci Cluster, a world-class high-performance supercomputerthat has many processors interconnected by a wide range of high-speed network connections. Even though the focus of this paper is on multikey sorting problems, developing new data structures and techniques for designing high-performance algorithms on computer clusters are of boththeoretical and practical interest. We investigate strategies for developing, implementing, and refining high-performance algorithms for sorting multi-dimensional data on computer clusters. In addition, maximizing the performance of such distributed memory machines requires efficient data structures coupled with good load balancing.
Although parallelization of algebraic fractal computations has been done in the past, the issue of efficient parallel computation has not been fully addressed in the literature. the objective of this paper is to exami...
详细信息
Although parallelization of algebraic fractal computations has been done in the past, the issue of efficient parallel computation has not been fully addressed in the literature. the objective of this paper is to examine the computational characteristics of algebraic fractal computations and determine an efficient scheme for parallel computation.
the following topics were dealt with: multiple processor architectures; networks and grids; non-numerical algorithms including sorting and graph algorithms; computation models; numerical parallel algorithms; schedulin...
the following topics were dealt with: multiple processor architectures; networks and grids; non-numerical algorithms including sorting and graph algorithms; computation models; numerical parallel algorithms; scheduling and performance evaluation including compiling, thread migration and meta computing; and highperformancecomputing applications including computational chemistry, command and control, and finance.
Emerging trends in heterogeneous distributed metacomputing and in Web Services technologies exhibit several commonalities that each domain can exploit. In this paper, we present an architectural model and design issue...
this paper presents a case study in developing an application class specific high-level interface for shared memory parallel programming. the application class we focus on is data mining. Withthe availability of larg...
暂无评论