Maya is a platform for investigating the impact of different memory coherence protocols on parallel architectures. We present the implementations of several weak memory protocols, together with some new primitives ded...
详细信息
Maya is a platform for investigating the impact of different memory coherence protocols on parallel architectures. We present the implementations of several weak memory protocols, together with some new primitives dedicated to weak memories using Maya. The results of some user applications are summarized and the impact of weak memories on the efficiency of these parallel programs is discussed.
On the basis of analyzing the evolution and drawbacks of current network fault diagnosis methods, a novel network data link troubleshooting system (NDTS) based on fuzzy neural network is proposed. NDTS tightly combine...
详细信息
The importance of multiuser detection for CDMA-based ad hoc network is addressed in this paper. Conventionally, the terminal in CDMA-based ad hoc network uses matched filter to receive packets, so the performance (e.g...
详细信息
This paper addresses the query performance issue for Relational OLAP (ROLAP) datacubes. We present a distributed multi-dimensional ROLAP indexing scheme which is practical to implement, requires only a small communica...
详细信息
ISBN:
(纸本)0769519199
This paper addresses the query performance issue for Relational OLAP (ROLAP) datacubes. We present a distributed multi-dimensional ROLAP indexing scheme which is practical to implement, requires only a small communication volume, and is fully adapted to distributed disks. Our solution is efficient for spatial searches in high dimensions and scalable in terms of data sizes, dimensions, and number of processors. Our method is also incrementally maintainable. Using "surrogate" group-bys, it allows for the efficient processing of arbitrary OLAP queries on partial cubes, where not all of the group-bys have been materialized. Our experiments show that the ROLAP advantage of better scalability, in comparison to MOLAP can be maintained while providing, at the same time, a fast and flexible index for OLAP queries.
We consider a cascade model of N different processors performing a distributedparallel simulation. The main goal of the study is to show that the long-time dynamics of the system have a cluster behaviour. To attack t...
详细信息
ISBN:
(纸本)3540294988
We consider a cascade model of N different processors performing a distributedparallel simulation. The main goal of the study is to show that the long-time dynamics of the system have a cluster behaviour. To attack this problem we combine two methods: stochastic comparison and Foster-Lyapunov functions.
applications that both access and generate large data sets increasingly draw our attention in high energy physics, astronomy, genomics and other disciplines. The Data Grids, like Gfarm, seek to harness geographically ...
详细信息
The performance of unstructured mesh applications presents a number of complexities and subtleties that do not arise for dense structured meshes. From a programming point of view, the handling of unstructured meshes h...
详细信息
The performance of unstructured mesh applications presents a number of complexities and subtleties that do not arise for dense structured meshes. From a programming point of view, the handling of unstructured meshes has an increased complexity in order to manage the necessary data structures and interactions between mesh-cells. From a performance point of view, there are added difficulties in understanding both the processing time on a single processor and the scaling characteristics when using large-scale parallel systems. In this work we present a general performance model for the calculation of deterministic S-N transport on unstructured meshes that is also applicable to structured meshes. The model captures the key processing characteristics of the calculation and is parametric using both system performance data (latency, bandwidth, processing rate etc.) and application data (mesh size etc.) as input. A single formulation of the model is used to predict the performance of two quite different implementations of the same calculation. It is validated on two clusters (an HP AlphaServer and an Itanium-2 system) showing high prediction accuracy.
This special issue of The Journal of Supercomputing brings together sixresearch papers reflecting some of the latest original achievements in the field of performanceevaluation of large-scale grid and cluster computin...
详细信息
This special issue of The Journal of Supercomputing brings together sixresearch papers reflecting some of the latest original achievements in the field of performanceevaluation of large-scale grid and cluster computing systems. These papers have been selected asextended and revised versions, subject to peer review, from those originally accepted forpresentation at the 3rdinternational Workshop on Performance Modeling, Evaluation, and Optimizationof parallel and distributed Systems (PMEO-PDS' 2004), which was held in conjunction with the 18thinternationalparallel and distributedprocessingsymposium, April 26-30, 2004, Santa Fe, NewMexico, USA. Each paper has been reviewed by at least two or three referees drawn from the ProgrammeCommittee and other experts worldwide. The selected papers address a range of issues encountered inthe performance evaluation of grid and cluster computing systems, and their contributions areoutlined below.
In an earlier work on the design of fine-grain, scalable classifiers for massively parallel computers, the technique of unifying cascaded networks has been demonstrated. This paper further examines the method adopted ...
详细信息
ISBN:
(纸本)0818682596
In an earlier work on the design of fine-grain, scalable classifiers for massively parallel computers, the technique of unifying cascaded networks has been demonstrated. This paper further examines the method adopted using a highly parallelprocessing architecture, entitled Unified Hierarchical Classifiers (UHC), based on the principles of Generalised Regression Neural Networks (GRNN). As with the GRNN, it has been shown that the resulting classification network can be implemented efficiently on general-purpose multiprocessor platforms without dedicated hardware for processor interconnections. Adding to this the structural simplicity, and the demonstrable potential for an effective distributed realisation on the Cray T3D, will make UHC an attractive classifier architecture in practical applications.
The proceedings contains 37 papers. Some of the topics discussed are: distributed computer systems, parallelprocessing systems, network protocols, data handling, algorithms, computer applications, time sharing system...
详细信息
The proceedings contains 37 papers. Some of the topics discussed are: distributed computer systems, parallelprocessing systems, network protocols, data handling, algorithms, computer applications, time sharing systems, multiprogramming, interactive computer systems, virtual reality, computer workstations, supercomputers, dynamic load balancing, data storage, and fault tolerant computing.
暂无评论