A Finite-element parallelcomputing frame-PANDA and its implementation processes are introduced. To validate the parallel performance of the PANDA frame, a series of tests were carried out to obtain the computing scal...
详细信息
ISBN:
(纸本)9783037852804
A Finite-element parallelcomputing frame-PANDA and its implementation processes are introduced. To validate the parallel performance of the PANDA frame, a series of tests were carried out to obtain the computingscale and the speedup ratios. First, three different large-scale freedom degree models (i.e. 1.83 million, 7 million and 10 million) of a typical engineering clamp were created in *** and were translated into geometric-grid files that can be identified in PANDA frame. Second, Linear static parallel computations of the three cases were successfully carried out on largeparallel computers with preconditioned conjugate gradient methods in PANDA frame. The speedup ratios of the three cases were obtained with a maximum process number of 64. The results show that the PANDA frame is competent for carrying out large-scale parallel computing of 10 million freedom degrees. In each scale, the parallelcomputing is nearly linearly accelerated along with the increase of process numbers, moreover, a super-linear speedup appears in some cases. The speedup curves show that the linear degree increases when the computingscale enlarges. The influence of different communication bandwidths on computing efficiency was also discussed. All the testing results indicate that the PANDA frame has excellent parallel performance and favorable computing scalability.
In large-scale multi-processor computing systems, global communications are typically supported by an auxiliary network (e. g., IBM Blue Gene) or with hardware support in the network (e. g., NEC Earth Simulator). We e...
详细信息
In large-scale multi-processor computing systems, global communications are typically supported by an auxiliary network (e. g., IBM Blue Gene) or with hardware support in the network (e. g., NEC Earth Simulator). We explore the potential for realizing efficient global communications that can scale beyond a million processors by harnessing the unique parallelism and wavelength routing properties of optical devices. Specifically, we use an arrayed waveguide grating router (AWGR) device as the basic building block in realizing scalable global communication. The AWGR is a passive switch fabric (wavelength router) that uses multiple wavelengths to interconnect outputs and inputs by following a specific cyclic wavelength routing (permutation) pattern. We analyze different network topologies using AWGR devices for barrier synchronization and propose techniques to pick parameters of the network for a given number of processors. We compare the performance and energy consumption for barrier synchronization with what is achievable with state-of-the-art electrical networks.
This work is motivated by the current numerical limitation in multiscale simulation of ductile fracture processes at scale down to the microstructure size and aims to overcome the difficulties in 3D complicated mesh g...
详细信息
This work is motivated by the current numerical limitation in multiscale simulation of ductile fracture processes at scale down to the microstructure size and aims to overcome the difficulties in 3D complicated mesh generation and locally extremely large strain analysis (local mesh distortion). The proposed 'conforming local meshfree approximation' directly and exactly satisfies displacement compatibility on a non-conforming assembly mesh. Local meshfree nodes, which can be freely placed and move on a finite element mesh, describe local large deformation. The improved accuracy on non-conforming mesh, the exactness in geometry representation on a structured mesh, and the good tolerance to mesh distortion are demonstrated by numerical examples. Copyright (C) 2010 John Wiley & Sons, Ltd.
There are increasingly demanding for huge computing capabilities and complex processes managing technologies along with the development of large-scaleparallel scientific computing applications. Taking Ensemble Predic...
详细信息
ISBN:
(纸本)9780769537665
There are increasingly demanding for huge computing capabilities and complex processes managing technologies along with the development of large-scaleparallel scientific computing applications. Taking Ensemble Prediction in climate domain for example, we discuss BPEL-based workflow for complex scientific computing management. The contributions of this paper are double-folds: On the framework side, we propose a job wrapping paradigm to wrap the environment depending applications with JSDL, and use a standard job submission system called GridSAM for job submitting and monitoring, which enables scientists to compose, monitor and run the applications easily with workflow system. On the algorithm side, we propose a job scheduling algorithm for the large-scaleparallel applications based on job pool in the heterogeneous environment. As a result, the resources can be utilized according to their capabilities and we can achieve the load balance of resources.
This article outlines the recent developments in the field of large-scale parallel computing applied to molecular simulations, also including some original, preliminary contributions of the authors. It is not meant to...
详细信息
This article outlines the recent developments in the field of large-scale parallel computing applied to molecular simulations, also including some original, preliminary contributions of the authors. It is not meant to be an exhaustive review paper, but rather an introductive material aimed at narrowing the "cultural gap" between the developers and users of molecular simulations (chemists, medicinal chemists and biologists-typical workstation users) and the informatics experts in massively parallelcomputing. The article starts with a brief overview of the existing molecular simulation techniques, in emphasizing the weaknesses of present approaches and the need for more computer-intensive methods. Docking procedures are the most discussed, given the high importance of this application in computer-aided drug design. An introduction to computer grids is logically pursued with the presentation of some of the most promising large-scaleparallel molecular simulations already performed. Eventually, the author's own research program, Docking@Grid, is briefly discussed.
暂无评论