logging in or signing up tascs news Abhil Download Post to : URL : Related Presentations : Share Add to Flag Embed Email Send to Blogs and Networks Add to Channel Uploaded from authorPOINTLite Insert YouTube videos in PowerPont slides with aS Desktop Copy embed code: (To copy code, click on the text box) Embed: URL: Thumbnail: WordPress Embed Customize Embed The presentation is successfully added In Your Favorites. Views: 47 Category: Entertainment License: All Rights Reserved Like it (0) Dislike it (0) Added: October 08, 2007 This Presentation is Public Favorites: 0 Presentation Description No description available. Comments Posting comment... Premium member Presentation Transcript Component Technology Initiatives: Component Technology Initiatives How can we exploit the component environment to provide computational scientists with better ways to develop their software? Support for Emerging HPC Hardware and Software Paradigms Motivation: chemistry, biology, and groundwater Software Quality and Verification Motivation: CCA toolkit, fusion, combustion, quantum chemistry, accelerators Computational Quality of Service Motivation: parallel mesh partitioning in combustion, resource management in quantum chemistry, parallel linear solvers in core-edge fusion and high-energy acceleratorsSupport for Emerging HPC Hardware and Software Paradigms: Support for Emerging HPC Hardware and Software Paradigms Investigate how component environments can be used to enhance the user’s experience programming for new and emerging high-end systems Massive levels of parallelism (105-106+ concurrent threads) Heterogeneous processor environments FPGAs, scalar+vector+threaded, co-processors, etc. Planned activities Develop and implement management for progress groups and (multi-level) MCMD parallelism Develop techniques to support asynchronous interactions of components representing co-processors/accelerators with those running on the traditional processorsSoftware Quality and Verification: Software Quality and Verification Extend component interface definitions with semantic annotations which can be verified/enforced at composition or run time (contracts) Developers provide “executable documentation” about correct use Users protected against improper usage due to code error, unforeseen circumstances, etc. Can facilitate verification that the code is actually doing what it was intended to do Planned activities and connections Method invocation sequencing constraints Implement and demonstrate annotations in selected Toolkit components Emphasize annotations needed to support CQoS workComputational Quality of Service: Computational Quality of Service Composition, substitution, and reconfiguration of running component applications in response to changing conditions Performance, accuracy, mathematical consistency, reliability, etc. Planned activities Develop common software infrastructure for dynamic component adaptivity Analysis (when to change) Control (what to change) Connections Software Quality initiative University of Oregon: performance monitoring, modeling, and analysis TASCS Software Quality Initiative: Interface semantics CCA Environment: Event model that supports dynamic CQoS behavior PERC Project: Performance tools TOPS Project: Solver interfaces Other external tools where appropriate (e.g., Eijkhout et al., Liu & Parashar, etc.)TASCS Initiative Interactions with Apps: TASCS Initiative Interactions with AppsEnvironmentYear 1 Deliverables: Environment Year 1 Deliverables Gary KumfertCore Tools: BenInstitutions: SNL, LLNL, UU, ORNL, ANL: Core Tools: Ben Institutions: SNL, LLNL, UU, ORNL, ANL PORTING to NLCF Machines XT-3 Babel (ORNL-lead) Wael started this in Indiana Mtg. Ccaffeine (SNL-lead) Tutorial (?) Toolkit (?) Blue Gene Babel (LLNL-lead) Ccaffeine (SNL-lead) Tutorial (?) Toolkit (?) Set up Helpdesk (LLNL-lead) https://cca-forum.org/bugs/help help@cca-forum.org 14 submissions 6 “fixed” –meaning resolved 3 tests 5 unresolved Documentation CCA Specification (LLNL-lead) To “encourage” shared staffing of the helpdesk, statistics about time-to-resolution, participating institutions, etc. will be published at CCA meetings.Enhancements: TomInstitutions: LLNL, BU, UU, SNL, ORNL: Enhancements: Tom Institutions: LLNL, BU, UU, SNL, ORNL CCA Services CommandLineService (???) MPIService (SNL-lead) EventService (UU-lead) Steve will speak more on this today GuiBuilderService (ANL-lead) Inter-framework Prototypes CCA-Kepler (UU-lead) XCAT-BabelRMI (BU-lead) Services are “MUST-HAVE” for CCA Spec 1.0 Ben was right in Indiana: We should start discussing 1.0 Release Criteria (like LLNL did with Babel). Do we want a more formal specification process?Usability: CraigInstitutions: LANL, SNL, ANL, ORNL, Tech-X, PNNL, LLNL: Usability: Craig Institutions: LANL, SNL, ANL, ORNL, Tech-X, PNNL, LLNL Draft CCA-Lite Spec & Fwk (LANL-lead) Component Test Harness (ORNL or Tech-X) Document Multilanguage & Component Debugging Techniques (SNL or ANL) Toolkit News: Toolkit News Rob Armstrong Toolkit Tut Sandia National Labs Livermore, CA 94551 rob@sandia.govTelecon: time to stop talking about the CCA Base Install Build System and just do it. : Telecon: time to stop talking about the CCA Base Install Build System and just do it. Need a scalable, pluggable build system and install base. Decisions have been made: Components built against a consistent set of CCA Tools (i.e. Sumo). where components and their dependencies are mixed and matched. Components can be added and removed without breaking the build Individual component code remains separate and individually modifiable in the build tree. Wael, Boyana, Ben will talk about this later.Nightly build and testing: Nightly build and testing CCA Tools (i.e. Sumo) and Base Install must have automatic testing, nightly build style Check out CCA Tools from their repositories and build them, test them, automatically. Many political questions: Should Toolkit components be tested nightly as well? Do you trust Rob with your repository? Repository tagging for “blessed” Toolkit distro's? Rob (yours truly) will talk about this later.Timetable: Timetable There are many components that are claimed to be “ready to go” A consistent approach is all that is lacking. By next meeting we should have a pretty deployable system (?) Front-loading some time on this will leverage the much larger number of component developers sooner rather than later...User and ApplicationsOutreach and Support: User and Applications Outreach and Support David E. Bernholdt Focus LeadApplications Activities: Applications Activities ANL Doug Dechow Beam Dynamics (DOE SBIR) discussions FACETS (fusion SciDAC2) participants; use of coupling and TOPS solver components Accelerator Physics SciDAC2 proposal development Indiana, ORNL SWIM (fusion SciDAC1) participants LLNL PSI LDRD (MPMD multiscale materials science) project ran large production runs using Babel RMI intensively (680 nodes x 2 CPUs/node x ~48 hours) ORNL GNEP integrated reactor simulation prototype (ORNL LDRD) PNNL Updating NWChem MCMD to latest CCA tools Tech-X various fusion projects UMD Space weather modeling (NSF) participant Klasky/Parashar CPES (fusion SciDAC1) discussionsUser Support: User Support ANL Doug Dechow Jim Jones (Masha Sosonkina student) SNL various help desk responses Tech-X Tech-X internal usersOutreach: Outreach ACTS tutorial Ben Allan contact with European group interested in componentizing MODFLOW 2000 (groundwater) WoCo9: Grid-based Problem Solving Environments: Implications for the Development and Deployment of Numerical Software Boyana Norris et al. talk & paper “Computational Quality of Service for Scientific Component Applications” 13th Biennial Computational Techniques and Applications Conference (CTAC '06) J.W. Larson, talk & paper "Organising Principles for Coupling in Multiphysics and Multiscale Models“ J.W. Larson, et al. talk & paper "PyMCT and PyCPL: Refactoring the Community Climate System Model Using Python” MODEST-7c: Multi-Scale, Multi-Physics Software Frameworks (in MODEST and Elsewhere) David Bernholdt, talk “The Common Component Architecture: Building Frameworks for Computational Science” NSF/Physics and the Information Frontier proposal with Bernholdt and Sussman as collaboratorsNear-Term Plans: Outreach N. Wang, an invited talk at GPCE4QoS workshop “Applying GPCE in Distributed and Parallel High-Performance Scientific Computing” Tech-X poster @ APS/DPP November “Framework for Modernization and Componentization of Fusion Modules" SC06 Binghamton paper "Benchmarking XML Processors for Applications in Grid Web Services“ Global Array tutorial Scientific Data Management tutorial CompFrame BOF IBM Scientific computer users group (ScicomP 2007) will be held at the Max Plank Institute in Garching (near Munich) July 2007 Tutorial revamp cca-forum.org upgrade and migration Applications PNNL groundwater SAP ramping up ANL climate SAP ramping up Near-Term PlansHall of Shame: Hall of Shame Indiana Rob Armstrong You do not have the permission to view this presentation. In order to view it, please contact the author of the presentation.
tascs news Abhil Download Post to : URL : Related Presentations : Share Add to Flag Embed Email Send to Blogs and Networks Add to Channel Uploaded from authorPOINTLite Insert YouTube videos in PowerPont slides with aS Desktop Copy embed code: (To copy code, click on the text box) Embed: URL: Thumbnail: WordPress Embed Customize Embed The presentation is successfully added In Your Favorites. Views: 47 Category: Entertainment License: All Rights Reserved Like it (0) Dislike it (0) Added: October 08, 2007 This Presentation is Public Favorites: 0 Presentation Description No description available. Comments Posting comment... Premium member Presentation Transcript Component Technology Initiatives: Component Technology Initiatives How can we exploit the component environment to provide computational scientists with better ways to develop their software? Support for Emerging HPC Hardware and Software Paradigms Motivation: chemistry, biology, and groundwater Software Quality and Verification Motivation: CCA toolkit, fusion, combustion, quantum chemistry, accelerators Computational Quality of Service Motivation: parallel mesh partitioning in combustion, resource management in quantum chemistry, parallel linear solvers in core-edge fusion and high-energy acceleratorsSupport for Emerging HPC Hardware and Software Paradigms: Support for Emerging HPC Hardware and Software Paradigms Investigate how component environments can be used to enhance the user’s experience programming for new and emerging high-end systems Massive levels of parallelism (105-106+ concurrent threads) Heterogeneous processor environments FPGAs, scalar+vector+threaded, co-processors, etc. Planned activities Develop and implement management for progress groups and (multi-level) MCMD parallelism Develop techniques to support asynchronous interactions of components representing co-processors/accelerators with those running on the traditional processorsSoftware Quality and Verification: Software Quality and Verification Extend component interface definitions with semantic annotations which can be verified/enforced at composition or run time (contracts) Developers provide “executable documentation” about correct use Users protected against improper usage due to code error, unforeseen circumstances, etc. Can facilitate verification that the code is actually doing what it was intended to do Planned activities and connections Method invocation sequencing constraints Implement and demonstrate annotations in selected Toolkit components Emphasize annotations needed to support CQoS workComputational Quality of Service: Computational Quality of Service Composition, substitution, and reconfiguration of running component applications in response to changing conditions Performance, accuracy, mathematical consistency, reliability, etc. Planned activities Develop common software infrastructure for dynamic component adaptivity Analysis (when to change) Control (what to change) Connections Software Quality initiative University of Oregon: performance monitoring, modeling, and analysis TASCS Software Quality Initiative: Interface semantics CCA Environment: Event model that supports dynamic CQoS behavior PERC Project: Performance tools TOPS Project: Solver interfaces Other external tools where appropriate (e.g., Eijkhout et al., Liu & Parashar, etc.)TASCS Initiative Interactions with Apps: TASCS Initiative Interactions with AppsEnvironmentYear 1 Deliverables: Environment Year 1 Deliverables Gary KumfertCore Tools: BenInstitutions: SNL, LLNL, UU, ORNL, ANL: Core Tools: Ben Institutions: SNL, LLNL, UU, ORNL, ANL PORTING to NLCF Machines XT-3 Babel (ORNL-lead) Wael started this in Indiana Mtg. Ccaffeine (SNL-lead) Tutorial (?) Toolkit (?) Blue Gene Babel (LLNL-lead) Ccaffeine (SNL-lead) Tutorial (?) Toolkit (?) Set up Helpdesk (LLNL-lead) https://cca-forum.org/bugs/help help@cca-forum.org 14 submissions 6 “fixed” –meaning resolved 3 tests 5 unresolved Documentation CCA Specification (LLNL-lead) To “encourage” shared staffing of the helpdesk, statistics about time-to-resolution, participating institutions, etc. will be published at CCA meetings.Enhancements: TomInstitutions: LLNL, BU, UU, SNL, ORNL: Enhancements: Tom Institutions: LLNL, BU, UU, SNL, ORNL CCA Services CommandLineService (???) MPIService (SNL-lead) EventService (UU-lead) Steve will speak more on this today GuiBuilderService (ANL-lead) Inter-framework Prototypes CCA-Kepler (UU-lead) XCAT-BabelRMI (BU-lead) Services are “MUST-HAVE” for CCA Spec 1.0 Ben was right in Indiana: We should start discussing 1.0 Release Criteria (like LLNL did with Babel). Do we want a more formal specification process?Usability: CraigInstitutions: LANL, SNL, ANL, ORNL, Tech-X, PNNL, LLNL: Usability: Craig Institutions: LANL, SNL, ANL, ORNL, Tech-X, PNNL, LLNL Draft CCA-Lite Spec & Fwk (LANL-lead) Component Test Harness (ORNL or Tech-X) Document Multilanguage & Component Debugging Techniques (SNL or ANL) Toolkit News: Toolkit News Rob Armstrong Toolkit Tut Sandia National Labs Livermore, CA 94551 rob@sandia.govTelecon: time to stop talking about the CCA Base Install Build System and just do it. : Telecon: time to stop talking about the CCA Base Install Build System and just do it. Need a scalable, pluggable build system and install base. Decisions have been made: Components built against a consistent set of CCA Tools (i.e. Sumo). where components and their dependencies are mixed and matched. Components can be added and removed without breaking the build Individual component code remains separate and individually modifiable in the build tree. Wael, Boyana, Ben will talk about this later.Nightly build and testing: Nightly build and testing CCA Tools (i.e. Sumo) and Base Install must have automatic testing, nightly build style Check out CCA Tools from their repositories and build them, test them, automatically. Many political questions: Should Toolkit components be tested nightly as well? Do you trust Rob with your repository? Repository tagging for “blessed” Toolkit distro's? Rob (yours truly) will talk about this later.Timetable: Timetable There are many components that are claimed to be “ready to go” A consistent approach is all that is lacking. By next meeting we should have a pretty deployable system (?) Front-loading some time on this will leverage the much larger number of component developers sooner rather than later...User and ApplicationsOutreach and Support: User and Applications Outreach and Support David E. Bernholdt Focus LeadApplications Activities: Applications Activities ANL Doug Dechow Beam Dynamics (DOE SBIR) discussions FACETS (fusion SciDAC2) participants; use of coupling and TOPS solver components Accelerator Physics SciDAC2 proposal development Indiana, ORNL SWIM (fusion SciDAC1) participants LLNL PSI LDRD (MPMD multiscale materials science) project ran large production runs using Babel RMI intensively (680 nodes x 2 CPUs/node x ~48 hours) ORNL GNEP integrated reactor simulation prototype (ORNL LDRD) PNNL Updating NWChem MCMD to latest CCA tools Tech-X various fusion projects UMD Space weather modeling (NSF) participant Klasky/Parashar CPES (fusion SciDAC1) discussionsUser Support: User Support ANL Doug Dechow Jim Jones (Masha Sosonkina student) SNL various help desk responses Tech-X Tech-X internal usersOutreach: Outreach ACTS tutorial Ben Allan contact with European group interested in componentizing MODFLOW 2000 (groundwater) WoCo9: Grid-based Problem Solving Environments: Implications for the Development and Deployment of Numerical Software Boyana Norris et al. talk & paper “Computational Quality of Service for Scientific Component Applications” 13th Biennial Computational Techniques and Applications Conference (CTAC '06) J.W. Larson, talk & paper "Organising Principles for Coupling in Multiphysics and Multiscale Models“ J.W. Larson, et al. talk & paper "PyMCT and PyCPL: Refactoring the Community Climate System Model Using Python” MODEST-7c: Multi-Scale, Multi-Physics Software Frameworks (in MODEST and Elsewhere) David Bernholdt, talk “The Common Component Architecture: Building Frameworks for Computational Science” NSF/Physics and the Information Frontier proposal with Bernholdt and Sussman as collaboratorsNear-Term Plans: Outreach N. Wang, an invited talk at GPCE4QoS workshop “Applying GPCE in Distributed and Parallel High-Performance Scientific Computing” Tech-X poster @ APS/DPP November “Framework for Modernization and Componentization of Fusion Modules" SC06 Binghamton paper "Benchmarking XML Processors for Applications in Grid Web Services“ Global Array tutorial Scientific Data Management tutorial CompFrame BOF IBM Scientific computer users group (ScicomP 2007) will be held at the Max Plank Institute in Garching (near Munich) July 2007 Tutorial revamp cca-forum.org upgrade and migration Applications PNNL groundwater SAP ramping up ANL climate SAP ramping up Near-Term PlansHall of Shame: Hall of Shame Indiana Rob Armstrong