Graph edge partition models have recently become an appealing alternative to graph vertex partition models for parallel and distributedcomputing due to their flexibility in balancing loads and their performance in re...
详细信息
ISBN:
(纸本)9781450365239
Graph edge partition models have recently become an appealing alternative to graph vertex partition models for parallel and distributedcomputing due to their flexibility in balancing loads and their performance in reducing communication cost [1, 3]. In this paper, we introduce a simple yet effective graph edge partitioning model for GPU computing. In practice, our model yields high partition quality (better than or the same as the state-of-the-art edge partition approaches, at least for power-law graphs) with low partition overhead. In theory, previous work [1] showed that an approximation factor of O(d(max) root logn log k) apply to the graphs with m = O(k(2)) edges (k is the number of partitions). Our model extends this result to all graphs. We demonstrate how graph edge partition model can be applied to GPU computing. We draw our examples from GPU program for locality enhancement both over time and (processor) space. For the first time, we demonstrate the effectiveness of edge partition for modeling data reuse in a many-core processors, both in theory and in practice.
Computational grids are a promising platform for executing large-scale resource intensive applications. However, resource management and scheduling in the grid environment is a complex undertaking as resources are (ge...
详细信息
The Galaxy science portal is a popular gateway to data analysis and computational tools for a broad range of life sciences communities. While Galaxy enables users to overcome the complexities of integrating diverse to...
详细信息
ISBN:
(纸本)9781479908981
The Galaxy science portal is a popular gateway to data analysis and computational tools for a broad range of life sciences communities. While Galaxy enables users to overcome the complexities of integrating diverse tools into unified workflows, it has only limited capabilities to execute those tools on the parallel and often distributed high-performance resources that the life sciences fields increasingly requires. We outline here an approach to meet this pressing requirement with the Swift parallel scripting language and its distributed runtime system. Swift's model of computation - implicitly parallel functional dataflow - is an elemental abstraction to which the core computing model of Galaxy maps very closely. We describe an integration between Galaxy and Swift that is transforming Galaxy into a much more powerful science gateway, retaining its user-friendly nature while extending its power to execute highly scalable workflows on diverse parallel environments.
In this paper we describe a vector and a parallel implementation of a stochastic simulation method to solve optimization problems in the field of many particle systems. We use a case-study where the (energetically) op...
详细信息
In this paper we describe a vector and a parallel implementation of a stochastic simulation method to solve optimization problems in the field of many particle systems. We use a case-study where the (energetically) optimal distribution of particles on a closed surface is studied, Crystallization on a closed surface is an interesting sub-domain since such a topology causes lattice defects. To obtain the optimal distribution of particles on a sphere we use the simulated annealing algorithm. Simulated annealing is an application of the Markov chain simulation method which, in principle, guarantees that the minimum in energy of our system of particles is found. However, the time for the algorithm to converge increases rapidly with system size. In order to find the best performing implementation we have made vectorized and parallelized implementations. We parallelize the simulated annealing method in several ways. Here we use two types of parallelization in conjunction, a systolic decomposition of the Markov chains and a functional decomposition of the energy calculations. The sequential nature of the simulated annealing algorithm is hard to parallelize and is therefore an important research topic to study the functional differences between parallel and sequential implementations. Results show that the parallelization influences the accuracy of the iterative process. In this paper we give a comparison between the vectorized and the parallelized implementation. It is shown that the current parallel implementation on the Parsytec GC parallel transputer platform is not capable of outruning our vector implementation on the GRAY Y-MP.
In this paper we describe a preferential admission control and load balancing algorithm for distributed Internet server clusters. Clients initiate sessions consisting of a series of transactions. The clients may be aw...
详细信息
ISBN:
(纸本)0769510116
In this paper we describe a preferential admission control and load balancing algorithm for distributed Internet server clusters. Clients initiate sessions consisting of a series of transactions. The clients may be aware of the various clusters of the distributed server and have preferences as to which cluster should process their session request. The scheme consists of a dispatcher which receives session requests and either admits or rejects those requests. Admitted requests are routed to their preferred cluster when the cluster is not congested. The algorithm also handles the case where a number of clients served have no cluster preference. We describe simulation results which demonstrate that the algorithm provides effective session admission control and load balancing, while maximizing the number of clients preferred by their most preferred cluster.
The grid provides inftastructure that allows an arbitrary application to be executed on a range of different computational resources. When input files are very large, or when fault tolerance is important, the data may...
详细信息
ISBN:
(纸本)0769524486
The grid provides inftastructure that allows an arbitrary application to be executed on a range of different computational resources. When input files are very large, or when fault tolerance is important, the data may be replicated Existing grid data replication middleware suffers from two shortcomings. First, it typically requires modification to existing applications. Second, there is limited support on automatic resource selection and a user usually chooses the replica manually to optimize the performance of the system. In this paper we discuss a middleware layer called the griddLeS Replication Service (GRS) that sits above existing replication services, solving both of these shortcomings. Two case studies are presented that illustrate the effectiveness of the approach.
The Soil and Water Assessment Tool (SWAT) has been used widely for large scale applications, reaching entire continents. Within the EU funded Envirogrids project, a detailed application of SWAT on the Black Sea Basin ...
详细信息
ISBN:
(纸本)9788890357411
The Soil and Water Assessment Tool (SWAT) has been used widely for large scale applications, reaching entire continents. Within the EU funded Envirogrids project, a detailed application of SWAT on the Black Sea Basin is envisaged using high resolution data. In order to support the computation, the model is run on a computer grid. The use of the SWAT allowed for such computations with little adaptations to the source. A 3-step procedure is needed. In the first step, a program is run in order to split the model into several sub-models. Afterwards, the sub-models are run in parallel. In a last step, the outputs of the sub-basins are collected at a central computer and the routing is performed. High computations are also needed when simulations have to be repeated, such as for sensitivity, calibration and uncertainty analysis. In these cases, the simulations are repeated for different parameter sets. In this paper, we discuss the gridification of the algorithm "LH-OAT" that performs sensitivity analysis and has been linked to the SWAT model. The results show a clear improvement in calculation time. Nevertheless, it is concluded that the parallelcomputing of a distributed model is mainly beneficial for large scale applications with high resolution, while running the sensitivity analysis algorithm has more general and obvious benefit. In a next step, the gridification will be optimised depending on the application and the overheads that are due to submission and receiving of files, as well as potential waiting times for executions on the grid.
Manual deployment of the application usually requires expertise both about the underlying system and the application. Automatic service deployment can improve deployment significantly by using on-demand deployment and...
详细信息
ISBN:
(纸本)9780769530895
Manual deployment of the application usually requires expertise both about the underlying system and the application. Automatic service deployment can improve deployment significantly by using on-demand deployment and self-healing services. To support these features this paper describes an extension the Globus Workspace Service [10]. This extension includes creating virtual appliances for grid services, service deployment from a repository, and influencing the service schedules by altering execution planning services, candidate set generators or information systems.
D-ReServE increases reliability of SOA-based systems in case of failure occurrence. The fault-tolerant information in D-ReServE is stored in the Stable Storage, which available space depletes with time. Thus, in this ...
详细信息
ISBN:
(纸本)9780769548791
D-ReServE increases reliability of SOA-based systems in case of failure occurrence. The fault-tolerant information in D-ReServE is stored in the Stable Storage, which available space depletes with time. Thus, in this paper we propose a garbage collection protocol for D-ReServE that allows the periodic purging of the Stable Storage, and discuss the challenges of garbage collection due to the nature of SOA systems.
作者：
Li, HuiBuyya, RajkumarLeiden Univ
Leiden Inst Adv Comp Sci POB 9512 NL-2333 CA Leiden Netherlands Univ Melbourne
Dept CSSE Grid Comp & Distributed Syst Lab Melbourne Vic 3010 Australia
Simulation studies of grid scheduling strategies require representative workloads to produce dependable results. Real production grid workloads have shown diverse correlation structures and scaling behavior, which are...
详细信息
ISBN:
(纸本)9780769530642
Simulation studies of grid scheduling strategies require representative workloads to produce dependable results. Real production grid workloads have shown diverse correlation structures and scaling behavior, which are different than the characteristics of the available supercomputer workloads and cannot be captured by Poisson or simple distribution-based models. We present models that are able to reproduce various correlation structures, including pseudo-periodicity and long range dependence. By conducting model-driven simulation, we quantitatively evaluate the performance impacts of workload correlations in grid scheduling. The results indicate that autocorrelations in workloads result in worse system performance, both at the local and the grid level. It is shown that realistic workload modeling is not only possible, but also necessary to enable dependable grid scheduling studies.
暂无评论