Revision as of 16:01, 28 July 2014 by Brutt (Talk | contribs)

(diff) ← Older revision | Latest revision (diff) | Newer revision → (diff)
Jump to: navigation, search

Cyberinfrastructure for ICME

The Cyberinfrastructure, serving as an integration platform, will support cooperative use of various software tools, including simulations software on high-performance multiprocessor systems, opening opportunities for optimizations and verification at all stages of the design process. The complexity of the CI will be hidden behind multifaceted user interfaces that will allow users to select the best tools for the task at the hand and to combine the tools into workflows with seamless data flows. The CI will also enable users to monitor the overall process and capturing information for storage into the knowledge databases.

In addition to presenting a cohesive and integrated interface to end users, the CI will also provide tools that facilitate the creation of parallel and distributed applications. Using these tools, domain experts will specify computations in the form of mathematical models and rules specifying the interaction of the models’ components. Computer scientists will provide descriptions of the underlying software and hardware infrastructure to the application generation tools. The tools will automatically translate the high-level models and rules into optimized program constructs and data structures based on the architectural descriptions. The optimizations will be tuned specifically for the execution system selected by the Grid services.

Architecture of the Virtual Organization for CyberDesign

As shown in the figure to the left, the Portal comprises five components: knowledge management, the repository of experimental data and material constants integrated with online model calibration tools, the repository of computational modules, and autonomic runtime environment for execution of computational workflows, such as multiscale simulations of materials.

Knowledge management is achieved by applying “architecture of participation” advocated and implemented by Web 2.0 concepts and technologies, in particular Wiki. The most important aspect of Web 2.0 is a focus on user-generated contents, as opposed to centrally managed information. We are in an early stage of this activity trying to accumulate enough information to launch “ICME Wikipedia” that would be interesting enough to attract the community to participate. In particular, the project toolbox (on the left side of the browser window) provides access to the user-provided information on the data, models, and tools relevant to the SRCLID and ICME projects.

Currently, we are concentrating on the development of the repositories to collect and disseminate through the Web data, models, and codes pertaining to computational material engineering. These repositories are accessible through services toolbox (on the left side of the browser window).

In the future, the cyberinfrastructure will provide the runtime environment for running complex multistep design optimizations that involve running multiscale simulations. The technical challenge of running these complex workflows comprising high-performance applications running in a distributed, heterogeneous environment that are necessary to fulfill the promise of ICME are intimidating for a material scientist. An intelligent, autonomic, that is, self-protecting, self-healing, self-configuring, and self-optimizing runtime environment that hides the IT complexity from the user is necessary.


The EVOWiki is maintained by the Cooperative Computing Group at Center for Advanced Vehicular Systems at Mississippi State University.

Primary contact: Tomasz Haupt (haupt@cavs.msstate.edu)

Major contributors to the Cyberinfrastructure: Anand Kalyanasundaram (alumni), Igor Zhuk (alumni), Gregory Henley, Bhargavi Parihar (alumni), Ricolindo Carino, and Florina Ciorba.

Students involved: Purnachandrika Ghantasala, Benjamin Daggolu

Personal tools

Material Models