Two cases demonstrating the utility of high performance computing to calculate X-ray absorption spectra (XAS) of solids are discussed. The X-ray absorption spectra of high pressure ices II, VIII, and IX have been computed with the Car Parrinello plane wave pseudopotential method. XAS for the intermediate structures obtained from uniaxial compression of hexagonal graphite along the c-axis are also studied. Whenever possible, comparisons to available experimental results are made. The reliability of the computational methods for the calculations of the XAS for these structures is discussed.
{"title":"Utility of High Performance Computing Facilities for the Calculation of the Theoretical X-ray Absorption Spectra of Solids","authors":"D. Shaw, M. Odelius, J. Tse","doi":"10.1109/HPCS.2007.33","DOIUrl":"https://doi.org/10.1109/HPCS.2007.33","url":null,"abstract":"Two cases demonstrating the utility of high performance computing to calculate X-ray absorption spectra (XAS) of solids are discussed. The X-ray absorption spectra of high pressure ices II, VIII, and IX have been computed with the Car Parrinello plane wave pseudopotential method. XAS for the intermediate structures obtained from uniaxial compression of hexagonal graphite along the c-axis are also studied. Whenever possible, comparisons to available experimental results are made. The reliability of the computational methods for the calculations of the XAS for these structures is discussed.","PeriodicalId":354520,"journal":{"name":"21st International Symposium on High Performance Computing Systems and Applications (HPCS'07)","volume":"71 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2007-05-13","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"125212647","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}
This paper is revisiting computational Grid economy models with an innovative, unrestricted and purely market regulated approach where both the consumer and provider standpoints are considered. The paper starts by outlining the actual Grid model shortcomings in regards to sharing incentive, return on investment, resource usage efficiency and quality of service accountability. Subsequently, a replacement for the traditional utility function is presented in the form of an aggregation of statistical indices computed from previous transactions. While this sharing scheme is based on economic foundations, its intent is not to transform the Grid in an expensive capitalist system. Targeted goal is rather to stimulate sharing by assessing resources dynamically in some form of abstract grid unit fostering supercomputing democratization.
{"title":"Resource Assessment using Market Indices: Toward an Economic Grid Exchange Model","authors":"Nicolas Dubé, M. Parizeau","doi":"10.1109/HPCS.2007.27","DOIUrl":"https://doi.org/10.1109/HPCS.2007.27","url":null,"abstract":"This paper is revisiting computational Grid economy models with an innovative, unrestricted and purely market regulated approach where both the consumer and provider standpoints are considered. The paper starts by outlining the actual Grid model shortcomings in regards to sharing incentive, return on investment, resource usage efficiency and quality of service accountability. Subsequently, a replacement for the traditional utility function is presented in the form of an aggregation of statistical indices computed from previous transactions. While this sharing scheme is based on economic foundations, its intent is not to transform the Grid in an expensive capitalist system. Targeted goal is rather to stimulate sharing by assessing resources dynamically in some form of abstract grid unit fostering supercomputing democratization.","PeriodicalId":354520,"journal":{"name":"21st International Symposium on High Performance Computing Systems and Applications (HPCS'07)","volume":"19 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2007-05-13","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"116360772","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}
Wesley Bland, T. Naughton, Geoffroy R. Vallée, S. Scott
The open source cluster application resources (OSCAR) toolkit is used to build and maintain HPC clusters. The OSCAR cluster installer provides a graphical user interface (GUI) "wizard" that directs the user through the installation process. This GUI is useful for general usage, but in some instances a more basic interface is desirable. Therefore, a command line interface (CLI) has been developed for the OSCAR installer. Through the CLI, the OSCAR installer is better equipped to support scripting, which aids more advanced testing scenarios and helps in cluster replication. Additionally, the CLI is better suited for remote installations using OSCAR. While a graphical user interface simplifies installation for occasional users, frequent users and the OSCAR developers will find the CLI beneficial. For developers, the CLI will make testing easier because it may now be scripted for automated testing. Furthermore, because the CLI creates logs of user input, the user may create an exact duplicate of a previous installation, which could be used in later rebuilds or for diagnostic purposes when reporting issues to the developers. This paper provides an overview of the design and capabilities of the OSCAR CLI along with a brief discussion of possible future work.
{"title":"Design and Implementation of a Menu Based OSCAR Command Line Interface","authors":"Wesley Bland, T. Naughton, Geoffroy R. Vallée, S. Scott","doi":"10.1109/HPCS.2007.14","DOIUrl":"https://doi.org/10.1109/HPCS.2007.14","url":null,"abstract":"The open source cluster application resources (OSCAR) toolkit is used to build and maintain HPC clusters. The OSCAR cluster installer provides a graphical user interface (GUI) \"wizard\" that directs the user through the installation process. This GUI is useful for general usage, but in some instances a more basic interface is desirable. Therefore, a command line interface (CLI) has been developed for the OSCAR installer. Through the CLI, the OSCAR installer is better equipped to support scripting, which aids more advanced testing scenarios and helps in cluster replication. Additionally, the CLI is better suited for remote installations using OSCAR. While a graphical user interface simplifies installation for occasional users, frequent users and the OSCAR developers will find the CLI beneficial. For developers, the CLI will make testing easier because it may now be scripted for automated testing. Furthermore, because the CLI creates logs of user input, the user may create an exact duplicate of a previous installation, which could be used in later rebuilds or for diagnostic purposes when reporting issues to the developers. This paper provides an overview of the design and capabilities of the OSCAR CLI along with a brief discussion of possible future work.","PeriodicalId":354520,"journal":{"name":"21st International Symposium on High Performance Computing Systems and Applications (HPCS'07)","volume":"53 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2007-05-13","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"122024430","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}
Flow processes that involve the displacement of a viscous fluid by a less viscous one often lead to a hydrodynamic instability known as viscous fingering. In this study the viscous fingering instability for anisotropic dispersive flows will be addressed. In order to understand the physics of the flow displacement, the basic equations of conservation of mass and momentum are solved for a two-dimensional porous medium. The linear stability of the flow is analyzed first. The flow is then modeled numerically using a highly accurate spectral method based on the Hartley transformation. The streamfunction and concentration fields are tracked using an iteration process for two dimensional flows in every time-step. In this study, different types of anisotropic dispersions are considered and their effects on finger patterns are examined. We will present physical discussion of how medium dispersivity affects hydrodynamics and could result in interesting instability schemes.
{"title":"Effect of Medium Dispersivity on the Viscous Fingering Instability in Porous Media","authors":"Karim Ghesmat, J. Azaiez","doi":"10.1109/HPCS.2007.16","DOIUrl":"https://doi.org/10.1109/HPCS.2007.16","url":null,"abstract":"Flow processes that involve the displacement of a viscous fluid by a less viscous one often lead to a hydrodynamic instability known as viscous fingering. In this study the viscous fingering instability for anisotropic dispersive flows will be addressed. In order to understand the physics of the flow displacement, the basic equations of conservation of mass and momentum are solved for a two-dimensional porous medium. The linear stability of the flow is analyzed first. The flow is then modeled numerically using a highly accurate spectral method based on the Hartley transformation. The streamfunction and concentration fields are tracked using an iteration process for two dimensional flows in every time-step. In this study, different types of anisotropic dispersions are considered and their effects on finger patterns are examined. We will present physical discussion of how medium dispersivity affects hydrodynamics and could result in interesting instability schemes.","PeriodicalId":354520,"journal":{"name":"21st International Symposium on High Performance Computing Systems and Applications (HPCS'07)","volume":"9 7","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2007-05-13","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"114010519","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}
Scientific applications written in MPI use collective communications intensively. Efficient and scalable implementation of such collective operations is therefore crucial to the performance of MPI applications running on clusters. Quadrics QsNetII is a high-performance network that implements some collectives at its Elan user-level library. Its MPI implementation uses such primitives directly. Quadrics communication software supports point-to-point message striping over multi-rail QsNetII networks. However, multi-rail collectives, other than broadcast, are not supported. In this work, we propose, design and implement a number of RDMA- based multi-port algorithms for the all-gather operation over multi-rail QsNetII clusters directly at the Elan level. Our performance results indicate that the proposed multi-port all-gather Direct algorithm gains an improvement of up to 1.49 for 32 KB messages over the native Elan_gather() collective.
{"title":"High Performance RDMA-based Multi-port All-gather on Multi-rail QsNet II","authors":"Y. Qian, A. Afsahi","doi":"10.1109/HPCS.2007.19","DOIUrl":"https://doi.org/10.1109/HPCS.2007.19","url":null,"abstract":"Scientific applications written in MPI use collective communications intensively. Efficient and scalable implementation of such collective operations is therefore crucial to the performance of MPI applications running on clusters. Quadrics QsNetII is a high-performance network that implements some collectives at its Elan user-level library. Its MPI implementation uses such primitives directly. Quadrics communication software supports point-to-point message striping over multi-rail QsNetII networks. However, multi-rail collectives, other than broadcast, are not supported. In this work, we propose, design and implement a number of RDMA- based multi-port algorithms for the all-gather operation over multi-rail QsNetII clusters directly at the Elan level. Our performance results indicate that the proposed multi-port all-gather Direct algorithm gains an improvement of up to 1.49 for 32 KB messages over the native Elan_gather() collective.","PeriodicalId":354520,"journal":{"name":"21st International Symposium on High Performance Computing Systems and Applications (HPCS'07)","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2007-05-13","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"128705511","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}
Earth science markup language (ESML) is efficient and effective in representing scientific data in an XML-based formalism. However, features of the data being represented are not accounted for in ESML. Such features might derive from events, identifications, or some other source. In order to account for features in an ESML context, they are considered from the perspective of annotation. Although it is possible to extend ESML to incorporate feature-based annotations internally, there are complicating factors identified that apply to ESML and most XML dialects. Rather than pursue the ESML-extension approach, an external representation for feature-based annotations via XML pointer language (XPointer) is developed. In previous work, it has been shown that it is possible to extract relationships from ESML-based representations, and capture the results in the resource description format (RDF). Application of this same requirement to XPointer-based annotations of ESML representations results in a revised semantic framework for the global geodynamics project (GGP).
{"title":"Semantically Enabling the Global Geodynamics Project: Incorporating Feature-Based Annotations via XML Pointer Language (XPointer)","authors":"L. Lumb, J. Lederman, J. Freemantle, K. Aldridge","doi":"10.1109/HPCS.2007.29","DOIUrl":"https://doi.org/10.1109/HPCS.2007.29","url":null,"abstract":"Earth science markup language (ESML) is efficient and effective in representing scientific data in an XML-based formalism. However, features of the data being represented are not accounted for in ESML. Such features might derive from events, identifications, or some other source. In order to account for features in an ESML context, they are considered from the perspective of annotation. Although it is possible to extend ESML to incorporate feature-based annotations internally, there are complicating factors identified that apply to ESML and most XML dialects. Rather than pursue the ESML-extension approach, an external representation for feature-based annotations via XML pointer language (XPointer) is developed. In previous work, it has been shown that it is possible to extract relationships from ESML-based representations, and capture the results in the resource description format (RDF). Application of this same requirement to XPointer-based annotations of ESML representations results in a revised semantic framework for the global geodynamics project (GGP).","PeriodicalId":354520,"journal":{"name":"21st International Symposium on High Performance Computing Systems and Applications (HPCS'07)","volume":"231 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2007-05-13","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"123262256","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 only robust general-purpose numerical methods for approximating the solution to systems of nonlinear algebraic equations (NAEs) are based on Newton's method. Many variants of Newton's method exist in order to take advantage of problem structure; it is often computationally infeasible to solve a given problem without taking some advantage of this structure. It is generally impossible to know a priori which variant of Newton's method will be optimal for a given problem. In this paper, we describe an algorithm for automatically selecting a composite Newton method, i.e., a sequential combination of Newton variants, for solving NAEs. The algorithm is based on a greedy principle that updates the current state at regular intervals according to the best performing Newton variant. Preliminary results show that it is possible for composite Newton methods to outperform optimal classical implementations of Newton's method, i.e., ones that only use one Newton variant on a given problem.
{"title":"Observations on greedy composite Newton methods","authors":"T. Ter, Matthew W. Donaldson, R. Spiteri","doi":"10.1109/HPCS.2007.22","DOIUrl":"https://doi.org/10.1109/HPCS.2007.22","url":null,"abstract":"The only robust general-purpose numerical methods for approximating the solution to systems of nonlinear algebraic equations (NAEs) are based on Newton's method. Many variants of Newton's method exist in order to take advantage of problem structure; it is often computationally infeasible to solve a given problem without taking some advantage of this structure. It is generally impossible to know a priori which variant of Newton's method will be optimal for a given problem. In this paper, we describe an algorithm for automatically selecting a composite Newton method, i.e., a sequential combination of Newton variants, for solving NAEs. The algorithm is based on a greedy principle that updates the current state at regular intervals according to the best performing Newton variant. Preliminary results show that it is possible for composite Newton methods to outperform optimal classical implementations of Newton's method, i.e., ones that only use one Newton variant on a given problem.","PeriodicalId":354520,"journal":{"name":"21st International Symposium on High Performance Computing Systems and Applications (HPCS'07)","volume":"326 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2007-05-13","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"132584596","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}
Several high-throughput distributed data-processing applications require multi-hop processing of streams of data. These applications include continual processing on data streams originating from a network of sensors, composing a multimedia stream through embedding several component streams originating from different locations, etc. These data-flow computing applications require multiple processing nodes interconnected according to the data-flow topology of the application, for on-stream processing of the data. Since the applications usually sustain for a long period, it is important to optimally map the component computations and communications on the nodes and links in the network, fulfilling the capacity constraints and optimizing some quality metric such as end-to-end latency. The mapping problem is unfortunately NP-complete and heuristics have been previously proposed to compute the approximate solution in a centralized way. However, because of the dynamicity of the network, it is practically impossible to aggregate the correct state of the whole network in a single node. In this paper, we present a distributed algorithm for optimal mapping of the components of the data flow applications. We propose several heuristics to minimize the message complexity of the algorithm while maintaining the quality of the solution.
{"title":"Towards a Decentralized Algorithm for Mapping Network and Computational Resources for Distributed Data-Flow Computations","authors":"S. Asaduzzaman, Muthucumaru Maheswaran","doi":"10.1109/HPCS.2007.32","DOIUrl":"https://doi.org/10.1109/HPCS.2007.32","url":null,"abstract":"Several high-throughput distributed data-processing applications require multi-hop processing of streams of data. These applications include continual processing on data streams originating from a network of sensors, composing a multimedia stream through embedding several component streams originating from different locations, etc. These data-flow computing applications require multiple processing nodes interconnected according to the data-flow topology of the application, for on-stream processing of the data. Since the applications usually sustain for a long period, it is important to optimally map the component computations and communications on the nodes and links in the network, fulfilling the capacity constraints and optimizing some quality metric such as end-to-end latency. The mapping problem is unfortunately NP-complete and heuristics have been previously proposed to compute the approximate solution in a centralized way. However, because of the dynamicity of the network, it is practically impossible to aggregate the correct state of the whole network in a single node. In this paper, we present a distributed algorithm for optimal mapping of the components of the data flow applications. We propose several heuristics to minimize the message complexity of the algorithm while maintaining the quality of the solution.","PeriodicalId":354520,"journal":{"name":"21st International Symposium on High Performance Computing Systems and Applications (HPCS'07)","volume":"65 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2007-05-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"130674785","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：