检索结果-内蒙古大学图书馆

Supercomputing Conference

作者： G.V. Wilson N.B. MacDonald C. Thornborrow C.M. Brough Edinburgh Parallel Computing Centre University of Edinburgh Edinburgh UK

ISBN: (纸本)9780818666056

Between 1987 and 1994, more than 100 students in a broad range of disciplines worked as summer scholars at Edinburgh parallel computing centre. Many of these students have since taken their parallel computing skills into graduate work and industry, and over a quarter of EPCC's technical staff are alumni of the Programme. This report describes the evolution and present operation of the Summer Scholarship Programme, and its costs and benefits.< >

关键词： parallel processing Scholarships Recruitment Computer industry Industrial training Concurrent computing Programming profession Collaborative work Physics Adders

来源：评论

学校读者我要写书评

暂无评论

CHIMP AND PUL - SUPPORT FOR PORTABLE parallel computing

引用

FUTURE GENERATION COMPUTER SYSTEMS 1995年第2期11卷 211-219页

作者： BRUCE, RAA CHAPPLE, S MACDONALD, NB TREW, AS TREWIN, S Edinburgh Parallel Computing Centre The University of Edinburgh James Clerk Maxwell Building The King's Buildings May field Road Edinburgh EH9 3JZ UK

The lack of a portable programming interface for parallel computers has inhibited the development of applications for such systems, and thus restricted the exploitation of this technology. In this paper we discuss various efforts to provide a portable interface to parallel computers, and present the Edinburgh CHIMP (Common High-level Interface to Message-Passing), and PUL (parallel Utilities Library) activities in this context. The concepts behind these projects are described, with details of the range of parallel systems on which they are available. An important consequence of the portability offered by the CHIMP message passing layer is the ability for applications programmers to reuse entire programs, as well as code modules. PUL assists this by providing a set of reusable components for implementing parallel applications. Developed on top of CHIMP, the PUL utilities are themselves portable across a range of platforms. Finally, this paper summarises experience using CHIMP and PUL to implement parallel applications in collaboration with both industrial and academic groups.

关键词： MESSAGE PASSING parallel LIBRARIES PORTABILITY REUSABILITY SCALABILITY

来源：评论

学校读者我要写书评

暂无评论

Design of a distributed software architecture for an intelligent planning system

引用

IEE PROCEEDINGS-CONTROL THEORY AND APPLICATIONS 1996年第2期143卷 125-131页

作者： Flavell, SJ Winter, SC Wilson, DR Hart, ER Centre for Parallel Computing University of Westminster London United Kingdom

The paper describes the design of a distributed software architecture for an intelligent planning system based on a process model of cognitive systems (PMCS). PMCS is described in terms of its problem-solving behaviour and structure, and is illustrated through a simple case study in robotics. In the PMCS autonomous knowledge sources, implementing a component of the functionality, co-operate by message-passing in order to maintain a distributed model of the system's environment and to synthesise decisions based on the model. Simulation tools developed at the University of Westminster for the rapid prototyping of parallel systems have been used to simulate the software architecture based on PMCS. Simulation results for the planning case study are presented in order to demonstrate the effectiveness of PMCS in solving planning problems in a distributed manner.

关键词： planning systems high-performance computing distributed processing software architecture co-operating knowledge-based systems computer system simulation

来源：评论

学校读者我要写书评

暂无评论

Minimizing transformer inference overhead using controlling element on Shenwei AI accelerator

引用

Frontiers of Information Technology and Electronic Engineering 2025年第4期26卷 605-622页

作者： Zhao, Yulong Wu, Chunzhi Wang, Yizhuo Zhang, Lufei Zhang, Yaguang Shen, Wenyuan Fan, Hao Fang, Hankang Qin, Yi Liu, Xin State Key Laboratory of Mathematical Engineering and Advanced Computing Wuxi214000 China School of Non-Commissioned Officer Space Engineering University Beijing100004 China National Supercomputing Center in Wuxi Wuxi214000 China Zhejiang Lab Hangzhou310000 China National Research Centre of Parallel Computer Engineering and Technology Beijing100081 China

Transformer models have become a cornerstone of various natural language processing (NLP) tasks. However, the substantial computational overhead during the inference remains a significant challenge, limiting their deployment in practical applications. In this study, we address this challenge by minimizing the inference overhead in transformer models using the controlling element on artificial intelligence (AI) accelerators. Our work is anchored by four key contributions. First, we conduct a comprehensive analysis of the overhead composition within the transformer inference process, identifying the primary bottlenecks. Second, we leverage the management processing element (MPE) of the Shenwei AI (SWAI) accelerator, implementing a three-tier scheduling framework that significantly reduces the number of host-device launches to approximately 1/10 000 of the original PyTorch-GPU setup. Third, we introduce a zero-copy memory management technique using segment-page fusion, which significantly reduces memory access latency and improves overall inference efficiency. Finally, we develop a fast model loading method that eliminates redundant computations during model verification and initialization, reducing the total loading time for large models from 22 128.31 ms to 1041.72 ms. Our contributions significantly enhance the optimization of transformer models, enabling more efficient and expedited inference processes on AI accelerators. © Zhejiang University Press 2025.

关键词： Transformer inference optimization Three-tier scheduling Zero-copy memory management Fast model loading

来源：评论

学校读者我要写书评

暂无评论

parallel PROCESSING FOR ABINITIO TOTAL ENERGY PSEUDOPOTENTIAL CALCULATIONS

引用

THEORETICA CHIMICA ACTA 1993年第4-5期84卷 325-334页

作者： CLARKE, LJ Edinburgh Parallel Computing Centre The University of Edinburgh Edinburgh UK

The total energy pseudopotential method is well suited to parallel processing. This paper discusses a procedure for calculating the valence electronic wavefunction in a given ionic configuration, and considers the exploitation of parallel processing using a data parallel approach. The implementation of this procedure on two message passing i860 based machines, containing up to 64 nodes, is described, and the prospects for massively parallel execution are examined.

关键词： TOTAL ENERGY PSEUDOPOTENTIAL VALENCE ELECTRONIC WAVE-FUNCTION

来源：评论

学校读者我要写书评

暂无评论

Progress toward a mining strategy based on mineral leaching with secondary recovery

引用

APPLIED MATHEMATICAL MODELLING 1996年第1期20卷 16-25页

作者： Forbes, LK Centre for Industrial and Applied Mathematics and Parallel Computing (CIAMP) Department of Mathematics University of Queensland Queensland Australia

The leaching of a mineral from the ground is considered using injection and recovery holes drilled into the rock. In this simplified analysis, the holes are arranged vertically and an appropriate caustic leaching fluid is introduced into the uppermost hole. The leaching liquor percolates down through the rock, dissolving the mineral of interest, and is pumped out when it arrives at the second hole. Water is introduced at the bottom hole in an attempt to increase the feasible recovery fraction of mineral-bearing liquor. Tire model problem is solved numerically using a boundary-integral formulation. Maximum feasible recovery fractions for the leaching fluid are obtained and their dependence upon the separation distance between the injection and recovery points and upon the pumping rate for water at the bottom hole are studied The volume of rock inundated by the leaching liquor is assessed, and a possible practical strategy for in situ leaching of low-grade ores is proposed.

关键词： mineral leaching water balloon boundary-integral method

来源：评论

学校读者我要写书评

暂无评论

FULLY NONLINEAR 2-LAYER FLOW OVER ARBITRARY TOPOGRAPHY

引用

JOURNAL OF ENGINEERING MATHEMATICS 1993年第4期27卷 419-432页

作者： BELWARD, SR FORBES, LK 1. Centre for Industrial and Applied Mathematics and Parallel Computing Department of Mathematics The University of Queensland 4072 Queensland Australia

Steady, two-dimensional, two-layer flow over an arbitrary topography is considered. The fluid in each layer is assumed to be inviscid and incompressible and flows irrotationally. The interfacial surface is found using a boundary integral formulation, and the resulting integrodifferential equations are solved iteratively using Newton's method. A linear theory is presented for a given topography and the non-linear theory is compared against this to show how the non-linearity affects the problem.

关键词： Two phase flow

来源：评论

学校读者我要写书评

暂无评论

MIGRATION OF SEISMIC DATA

引用

JOURNAL OF STATISTICAL PHYSICS 1994年第1-2期76卷 703-723页

作者： BONOMI, E CABITZA, G 1. Parallel Computing Centre for Advanced Studies Research and Development in Sardinia I-09123 Cagliari Italy

Prospecting for oil and gas resources poses the problem of determining the geological structure of the earth's crust from indirect measurements. Seismic migration is an acoustic image reconstruction technique based on the inversion of the scalar wave equation. Extensive computation is necessary before reliable information can be extracted from large sets of recorded data. In this paper a collection of ''industrial'' migration techniques, each giving rise to a data parallel algorithm, is outlined. Computer simulations on synthetic seismic data illustrate the problem and the approach.

关键词： SEISMOLOGY MIGRATION SCALAR WAVE EQUATION parallel computing

来源：评论

学校读者我要写书评

暂无评论

AN ANALYTICAL AND NUMERICAL STUDY OF THE FORCED VIBRATION OF A SPHERICAL CAVITY

引用

JOURNAL OF SOUND AND VIBRATION 1994年第4期172卷 471-489页

作者： FORBES, LK Centre for Industrial Applied Mathematics Parallel Computing (CIAMP) Department of Mathematics The University of Queensland St. Lucia Queensland 4072 Australia

Forced motion of a spherical bubble in an incompressible viscous fluid is considered. The system is assumed to be governed by the Rayleigh-Plesset equation, and the forcing occurs by means of a harmonically varying pressure field at infinity. A perturbation solution is presented, which extends previously published results in that it gives a complete qualitative description of the behaviour near the primary resonance, as well as for all the superharmonic and subharmonic resonances. These low order results are confirmed and extended to large amplitude motion by means of a numerical shooting algorithm. The method is capable of computing stable and unstable periodic solutions with equal ease, and it determines the stability automatically by using a numerical implementation of Floquet theory. The application of this numerical method to a non-analytical model of bubble behaviour is briefly discussed.

关键词： Viscous flow

来源：评论

学校读者我要写书评

暂无评论

ADAPTIVE DEADLOCK-FREE PACKET ROUTEING IN TRANSPUTER-BASED MULTIPROCESSOR INTERCONNECTION NETWORKS

引用

COMPUTER JOURNAL 1991年第6期34卷 493-502页

作者： SON, NT PAKER, Y Centre for Parallel Computing Queen Mary and Westfield College University of London Mile End Road London El 4NS

This paper presents a deadlock-free routeing algorithm for multiprocessor interconnection networks based on store-and-forward (S/F) communication. The adaptive nature of the method proposed encourages using light traffic paths. Furthermore it has the properties of avoiding blocked communication (deadlock), reducing communication delay time between source and destination, using efficiently message buffers as network resources and being able to control communication traffic flow from each processor of the network. The routeing algorithm has been implemented on a 64-node transputer network (T-Rack) configured as a number of well known topologies to evaluate the behaviour of the algorithm and some performance figures have been derived.

关键词： multiprocessor interconnection networks PROCESSOR Routeing destination algorithms Web resources

来源：评论

学校读者我要写书评

暂无评论

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案：

请选择收藏分类：

通借通还

建议与咨询 留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案： 新增检索档案 确定 取消

请选择收藏分类： 新增自定义分类 确定 取消

通借通还

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

请选择保存的检索档案：

请选择收藏分类：