 |
|
 |
||||||
|
||||||
 |
Home
| Mission
|
about SciDAC
|
Contact Us |
|||||
 |
 |
 |
||||
|
Alumni ProjectBuilding Component-Based Terascale Applications Using the CCA CCTTSS Leads: SummaryAs the Center for Component Technology for Terascale Simulation Software (CCTTSS) is developing and refining the tools and techniques for component-based high-performance scientific computing, numerous applications groups are moving to adopt the Common Component Architecture (CCA). A wide range of efforts are creating standardized interfaces, reusable components, and full-blown applications in the CCA environment. CCA technology is helping to increase the productivity and capability of simulation-based research efforts. Climate Modeling Computational climate modeling is critically important for our understanding of global processes and the potential for human impact. The CCTTSS is currently collaborating with two major climate modeling efforts. The SciDAC CCSM1 is a coupled global climate model, into which components are being introduces at three different levels. At the system integration level, the Model Coupling Toolkit2 (MCT), which couples elements of the CCSM (i.e., atmosphere, ocean, sea ice), is being adapted to the CCA environment. The Community Atmosphere Model (CAM) of the CCSM is being refactored to decouple the “physics” and “dynamics” portions of the code; the various “dynamical cores” will be available as interoperable components. Finally, at the algorithmic level, MCT-derived components are being used to create a new River Transport Model for the CCSM. NASA's ESMF3 is an effort to create a specialized component-based environment supporting a wide range of climate modeling activities, and will be the infrastructure for future versions of the CCSM and other climate codes. Software has been created to automatically “wrap” ESMF components to pose as CCA components. This promising work allows CCA components to be used by the ESMF community and vice versa. Combustion SimulationsThe SciDAC CFRFS4 project is developing a component-based toolkit for flame simulations to facilitate the creation and execution of simulations by researchers with minimal knowledge of the supporting software infrastructure. The CFRFS, collaborating with the SciDAC TOPS5, PERC6, and APDEC7 centers, have developed components for time integration, structured adaptively refined meshes (SAMR), performance observation, and various physical models, allowing the assembly of a wide range of simulations. The clean, flexible interfaces have greatly simplified the task of incrementally upgrading the implementation of individual components, without impacting other parts of the toolkit. Future plans include expanding the scientific capabilities of the CFRFS and expanding to larger problem sizes. Computational ChemistryQuantum chemistry (QC) is a major tool in the accurate simulation of chemical phenomena, ranging from combustion modeling to the design of catalysts, pharmaceuticals, and novel materials. Researchers are creating interfaces and componentizing the NWChem8 and MPQC9 QC codes and the TAO10 optimization package, as well as the linear algebra capabilities within Global Arrays11 and PETSc12 , to provide chemists with an unprecedented level of flexibility in determining molecular structures. Recent work builds on an earlier prototype, providing a more complete and robust implementation with a user-friendly interface. Using the new capabilities, benchmarking studies of numerical optimization techniques for chemical problems are being performed. Future work will include interoperability at deeper levels within the QC packages and enabling higher levels of parallelism by being able to run multiple independent computations simultaneously within a single simulation. Other Application AreasNumerous additional projects, both inside and outside the SciDAC program, are adopting the CCA. Fusion and SciDAC SDM13 center researchers are investigating the CCA for the integration of coupled simulations and diverse data analysis tools, respectively. Other active areas include materials science, nanoscience, and underground radionuclide transport. Scientific ComponentsIn addition to entire applications, many “standalone” reusable components are also being developed. These include various services, tools for mesh management, discretization, linear algebra, integration, optimization, parallel data description and redistribution, visualization, and performance evaluation. Another important facet of this work is defining domain-specific abstract interfaces, which helps in realizing our vision of plug-and-play scientific components. We collaborate with the SciDAC TSTT14, APDEC, and TOPS centers, and we engage the community at large in these efforts. For further information on this subject contact: 1 Community Climate System Model , PIs: J. Drake and R. Malone, http://www.cgd.ucar.edu/csm
|
||||
Home | ASCR | Contact Us | DOE disclaimer |
|
|
