This article describes a system for run-time prediction of applications in heterogeneous environments. To exploit the power of computational grids, scheduling systems need profound information about the job to be executed. The run-time of a job is - beside others - not only dependent of its kind and complexity but also of the adequacy and load of the remote host where it will be executed. Accounting and billing are additional aspects that have to be considered when creating a schedule. Currently predictions are achieved by using descriptive models of the applications or by applying statistical methods to former jobs mostly neglecting the behaviour of users. Motivated by this, we propose a method that is not only based on the characteristics of a job but also takes the behaviour of single users and groups of similar users respectively into account. The basic idea of our approach is to cluster users, hosts and jobs and apply multiple methods in order to detect similarities and create forecasts. This is achieved by tagging jobs with attributes and by deriving predictions for similar attributed jobs whereas the recent behaviour of a user determines which predictions are finally taken.
{"title":"An Architecture for an Adaptive Run-time Prediction System","authors":"C. Glasner, J. Volkert","doi":"10.1109/ISPDC.2008.34","DOIUrl":"https://doi.org/10.1109/ISPDC.2008.34","url":null,"abstract":"This article describes a system for run-time prediction of applications in heterogeneous environments. To exploit the power of computational grids, scheduling systems need profound information about the job to be executed. The run-time of a job is - beside others - not only dependent of its kind and complexity but also of the adequacy and load of the remote host where it will be executed. Accounting and billing are additional aspects that have to be considered when creating a schedule. Currently predictions are achieved by using descriptive models of the applications or by applying statistical methods to former jobs mostly neglecting the behaviour of users. Motivated by this, we propose a method that is not only based on the characteristics of a job but also takes the behaviour of single users and groups of similar users respectively into account. The basic idea of our approach is to cluster users, hosts and jobs and apply multiple methods in order to detect similarities and create forecasts. This is achieved by tagging jobs with attributes and by deriving predictions for similar attributed jobs whereas the recent behaviour of a user determines which predictions are finally taken.","PeriodicalId":125975,"journal":{"name":"2008 International Symposium on Parallel and Distributed Computing","volume":"154 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2008-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"122688206","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
The method GMRES is used widely to solve the large sparse linear systems. In this paper, we will present an effective parallel hybrid asynchronous method, which combines the typical parallel method GMRES with the Least Square method that needs some eigenvalues obtained from a parallel Arnoldi process. And we will apply it on a Grid Computing platform Grid5000. Grid computing in general is a special type of parallel computing. It intends to deliver high-performance computing over distributed platforms for computation and data-intensive applications by making use of a very large amount of resources. From the numeric results, we will present this hybrid method has some advantage for some matrices compared to the general method GMRES.
{"title":"Large Scale Parallel Hybrid GMRES Method for the Linear System on Grid System","authors":"Ye Zhang, Guy Bergére, S. Petiton","doi":"10.1109/ISPDC.2008.21","DOIUrl":"https://doi.org/10.1109/ISPDC.2008.21","url":null,"abstract":"The method GMRES is used widely to solve the large sparse linear systems. In this paper, we will present an effective parallel hybrid asynchronous method, which combines the typical parallel method GMRES with the Least Square method that needs some eigenvalues obtained from a parallel Arnoldi process. And we will apply it on a Grid Computing platform Grid5000. Grid computing in general is a special type of parallel computing. It intends to deliver high-performance computing over distributed platforms for computation and data-intensive applications by making use of a very large amount of resources. From the numeric results, we will present this hybrid method has some advantage for some matrices compared to the general method GMRES.","PeriodicalId":125975,"journal":{"name":"2008 International Symposium on Parallel and Distributed Computing","volume":"30 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2008-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"115432382","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
In this paper we present an extension to MONARC, a generic simulator for large scale distributed systems, which allows realistic evaluation of various actual distributed system technologies based on real-world monitored data supplied by MonALISA. The field of modelling and simulation was long-time seen as a viable solution to develop new algorithms and technologies and to enable the development of large-scale distributed systems, where analytical validations are prohibited by the nature of the encountered problems. This paper presents a novel approach to combining two distributed systems domains, monitoring and simulation, highlighting a realistic solution to the problem of accurately evaluating various distributed systems technologies using simulation. We also present a simulation study which demonstrates the interoperability between the simulation framework and the monitoring instrument, demonstrating important properties of the US LHCNet research network in the context of the LHC experiments in CERN.
{"title":"Realistic Simulation of Large Scale Distributed Systems using Monitoring","authors":"C. Dobre, Corina Stratan, V. Cristea","doi":"10.1109/ISPDC.2008.37","DOIUrl":"https://doi.org/10.1109/ISPDC.2008.37","url":null,"abstract":"In this paper we present an extension to MONARC, a generic simulator for large scale distributed systems, which allows realistic evaluation of various actual distributed system technologies based on real-world monitored data supplied by MonALISA. The field of modelling and simulation was long-time seen as a viable solution to develop new algorithms and technologies and to enable the development of large-scale distributed systems, where analytical validations are prohibited by the nature of the encountered problems. This paper presents a novel approach to combining two distributed systems domains, monitoring and simulation, highlighting a realistic solution to the problem of accurately evaluating various distributed systems technologies using simulation. We also present a simulation study which demonstrates the interoperability between the simulation framework and the monitoring instrument, demonstrating important properties of the US LHCNet research network in the context of the LHC experiments in CERN.","PeriodicalId":125975,"journal":{"name":"2008 International Symposium on Parallel and Distributed Computing","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2008-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"126786843","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
We propose a new mixed integer linear programming approach to solve the classical problem of scheduling independent parallel tasks without preemption. We propose a formulation where the goal is to minimize the makespan.Then we show the flexibility of this approach by extending the result to the contiguous case. We validate this approach with some experiments on the execution times and comparing the optimal results with the solutions provided by list algorithms.
{"title":"A MILP Approach to Schedule Parallel Independent Tasks","authors":"A. Goldman, Yanik Ngoko","doi":"10.1109/ISPDC.2008.59","DOIUrl":"https://doi.org/10.1109/ISPDC.2008.59","url":null,"abstract":"We propose a new mixed integer linear programming approach to solve the classical problem of scheduling independent parallel tasks without preemption. We propose a formulation where the goal is to minimize the makespan.Then we show the flexibility of this approach by extending the result to the contiguous case. We validate this approach with some experiments on the execution times and comparing the optimal results with the solutions provided by list algorithms.","PeriodicalId":125975,"journal":{"name":"2008 International Symposium on Parallel and Distributed Computing","volume":"40 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2008-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"124202586","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Georgiana Macariu, A. Cârstea, M. Frîncu, D. Petcu
One of the benefits of the current service-oriented architectures is the easy static composition of geographically scattered services into complex applications. Dynamic composition is more difficult to achieve with the current technologies. We propose a practical solution for dynamic composition of the facilities provided by computer algebra systems, based on Grid services and the WS-BPEL standard Web service orchestration language. Moreover, we introduce a methodology for migrating from Web services to Grid services using databases for persistence.
{"title":"Towards a Grid Oriented Architecture for Symbolic Computing","authors":"Georgiana Macariu, A. Cârstea, M. Frîncu, D. Petcu","doi":"10.1109/ISPDC.2008.46","DOIUrl":"https://doi.org/10.1109/ISPDC.2008.46","url":null,"abstract":"One of the benefits of the current service-oriented architectures is the easy static composition of geographically scattered services into complex applications. Dynamic composition is more difficult to achieve with the current technologies. We propose a practical solution for dynamic composition of the facilities provided by computer algebra systems, based on Grid services and the WS-BPEL standard Web service orchestration language. Moreover, we introduce a methodology for migrating from Web services to Grid services using databases for persistence.","PeriodicalId":125975,"journal":{"name":"2008 International Symposium on Parallel and Distributed Computing","volume":"7 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2008-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"124953999","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
We consider the problem of computing all Nash equilibria in bimatrix games (i.e., nonzero-sum two-player noncooperative games). Computing all Nash equilibria for large bimatrix games using single-processor computers is not feasible due to the exponential time required by the existing algorithms. We consider the use of parallel computing which allows us to solve larger games. We design and implement a parallel algorithm for computing all Nash Equilibria in bimatrix games. The algorithm computes all Nash equilibria by searching all possible supports of mixed strategies. We perform experiments on a cluster computing system to evaluate the performance of the parallel algorithm.
{"title":"Computing Equilibria in Bimatrix Games by Parallel Support Enumeration","authors":"J. Widger, Daniel Grosu","doi":"10.1109/ISPDC.2008.38","DOIUrl":"https://doi.org/10.1109/ISPDC.2008.38","url":null,"abstract":"We consider the problem of computing all Nash equilibria in bimatrix games (i.e., nonzero-sum two-player noncooperative games). Computing all Nash equilibria for large bimatrix games using single-processor computers is not feasible due to the exponential time required by the existing algorithms. We consider the use of parallel computing which allows us to solve larger games. We design and implement a parallel algorithm for computing all Nash Equilibria in bimatrix games. The algorithm computes all Nash equilibria by searching all possible supports of mixed strategies. We perform experiments on a cluster computing system to evaluate the performance of the parallel algorithm.","PeriodicalId":125975,"journal":{"name":"2008 International Symposium on Parallel and Distributed Computing","volume":"33 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2008-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"125010094","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
In this paper we present the architecture of a distributed framework that allows real-time accurate monitoring of large scale high-speed networks. An important component of a large-scale distributed collaboration is the complex network infrastructure on which it relies. For monitoring and controlling the networking resources an adequate instrument should offer the possibility to collect and store the relevant monitoring information, presenting significant perspectives and synthetic views of how the large distributed system performs. We therefore developed within the MonALISA monitoring framework a system able to collect, store, process and interpret the large volume of status information related to the US LHCNet research network. The system uses flexible mechanisms for data representation, providing access optimization and decision support, being able to present real-time and long-time history information through global or specific views and to take further automated control actions based on them.
{"title":"A Monitoring Architecture for High-Speed Networks in Large Scale Distributed Collaborations","authors":"Alexandru Costan, C. Dobre, V. Cristea, R. Voicu","doi":"10.1109/ISPDC.2008.33","DOIUrl":"https://doi.org/10.1109/ISPDC.2008.33","url":null,"abstract":"In this paper we present the architecture of a distributed framework that allows real-time accurate monitoring of large scale high-speed networks. An important component of a large-scale distributed collaboration is the complex network infrastructure on which it relies. For monitoring and controlling the networking resources an adequate instrument should offer the possibility to collect and store the relevant monitoring information, presenting significant perspectives and synthetic views of how the large distributed system performs. We therefore developed within the MonALISA monitoring framework a system able to collect, store, process and interpret the large volume of status information related to the US LHCNet research network. The system uses flexible mechanisms for data representation, providing access optimization and decision support, being able to present real-time and long-time history information through global or specific views and to take further automated control actions based on them.","PeriodicalId":125975,"journal":{"name":"2008 International Symposium on Parallel and Distributed Computing","volume":"87 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2008-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"117118422","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
In this paper we describe hardware implementations of generators of combinatorial objects. For implementation several systolic algorithms were selected that generate combinatorial configurations in a linear array model. The algorithms generate such objects as combinations, combinations with repetitions, t-ary trees, partitions, and variations with repetitions. The generators were implemented in VHLD with Xilinx Foundation ISE software and tested on Digilent development boards with Xilinx FPGAs. Implementation data obtained for various input parameters and FPGA devices are given.
本文描述了组合对象生成器的硬件实现。为了实现，选择了几种收缩算法，在线性阵列模型中生成组合配置。算法生成诸如组合、重复组合、t- tree、分区和重复变化等对象。发生器在VHLD中使用Xilinx Foundation ISE软件实现，并在带有Xilinx fpga的Digilent开发板上进行了测试。给出了各种输入参数和FPGA器件的实现数据。
{"title":"FPGA Generators of Combinatorial Configurations in a Linear Array Model","authors":"Zbigniew Kokosinski, Pawel Halesiak","doi":"10.1109/ISPDC.2008.48","DOIUrl":"https://doi.org/10.1109/ISPDC.2008.48","url":null,"abstract":"In this paper we describe hardware implementations of generators of combinatorial objects. For implementation several systolic algorithms were selected that generate combinatorial configurations in a linear array model. The algorithms generate such objects as combinations, combinations with repetitions, t-ary trees, partitions, and variations with repetitions. The generators were implemented in VHLD with Xilinx Foundation ISE software and tested on Digilent development boards with Xilinx FPGAs. Implementation data obtained for various input parameters and FPGA devices are given.","PeriodicalId":125975,"journal":{"name":"2008 International Symposium on Parallel and Distributed Computing","volume":"50 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2008-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"125673302","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Data intensive applications require massive data transfers between storage and processing units. VLSI scaling has increased the sizes of dynamic memories as well as speeds and capabilities of processing units to a point where, for many such applications, storage and computational processing capabilities are no longer the main limiting factors. Despite this fact, most current architectures fail to meet the performance requirements for such data intensive applications. In this paper, we describe a PIM architecture that harnesses the benefits of VLSI scaling to accelerate matrix operations that constitute the core of many data-intensive applications. We then present data partitioning and placement schemes that are efficient in terms of the computational complexities and internode communication cost. Such approaches are evaluated and analyzed under various computing environments. We also discuss on how to apply such partitioning and placement schemes to each matrix when chains of matrix operations are given as a task.
{"title":"Data Partitioning and Placement Schemes for Matrix Multiplications on a PIM Architecture","authors":"J. Cha, S. Gupta","doi":"10.1109/ISPDC.2008.7","DOIUrl":"https://doi.org/10.1109/ISPDC.2008.7","url":null,"abstract":"Data intensive applications require massive data transfers between storage and processing units. VLSI scaling has increased the sizes of dynamic memories as well as speeds and capabilities of processing units to a point where, for many such applications, storage and computational processing capabilities are no longer the main limiting factors. Despite this fact, most current architectures fail to meet the performance requirements for such data intensive applications. In this paper, we describe a PIM architecture that harnesses the benefits of VLSI scaling to accelerate matrix operations that constitute the core of many data-intensive applications. We then present data partitioning and placement schemes that are efficient in terms of the computational complexities and internode communication cost. Such approaches are evaluated and analyzed under various computing environments. We also discuss on how to apply such partitioning and placement schemes to each matrix when chains of matrix operations are given as a task.","PeriodicalId":125975,"journal":{"name":"2008 International Symposium on Parallel and Distributed Computing","volume":"44 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2008-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"131724282","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
A new parallel algorithm, based on the concept of anti diagonal wave pattern, for computing approximate inverses, is introduced for symmetric multiprocessor systems. The parallel normalized approximate inverses are used in conjunction with parallel normalized preconditioned conjugate gradient-type schemes, for the efficient solution of sparse finite element linear systems. The parallel implementation issues of the new algorithm are discussed and the parallel performance is presented, using OpenMP.
{"title":"An Improved Parallel Algorithm for Computing Approximate Inverses by Reducing Synchronizations","authors":"G. Gravvanis, K. M. Giannoutakis","doi":"10.1109/ISPDC.2008.18","DOIUrl":"https://doi.org/10.1109/ISPDC.2008.18","url":null,"abstract":"A new parallel algorithm, based on the concept of anti diagonal wave pattern, for computing approximate inverses, is introduced for symmetric multiprocessor systems. The parallel normalized approximate inverses are used in conjunction with parallel normalized preconditioned conjugate gradient-type schemes, for the efficient solution of sparse finite element linear systems. The parallel implementation issues of the new algorithm are discussed and the parallel performance is presented, using OpenMP.","PeriodicalId":125975,"journal":{"name":"2008 International Symposium on Parallel and Distributed Computing","volume":"47 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2008-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"120904617","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
京公网安备 11010802042870号
京ICP备2023020795号-1
扫码关注我们
求助内容：
应助结果提醒方式：