This subproject is one of many research subprojects utilizing the resources provided by a Center grant funded by NIH/NCRR. The subproject and investigator (PI) may have received primary funding from another NIH source, and thus could be represented in other CRISP entries. The institution listed is for the Center, which is not necessarily the institution for the investigator. Information Technology (IT) has transformed computational science to the extent that realistic, atom-based simulations of key processes in chemistry, geoscience, nanoscience science and engineering, and biology can now be tackled using Car-Parrinello ab initio molecular dynamics1 (CPAIMD, or more simply, CPMD) implemented on high performance computing (HPC) platforms. These calculations, which can model complex events such as chemical bond forming and breaking, have enormous potential to impact Science and Technology, e.g., through (1) the design of new industrial catalysts that generate novel polymers with improved properties; (2) the study of materials under extreme conditions found in Earth's mantle and cores of planets; (3) the analysis of nanoscale, and even molecular scale, devices, which could revolutionize the electronics and computer industries; and (4) examination of the mechanism of enzyme catalysis to enable the design of cheap, artificial catalysts (biomimetics) of high efficiency. Although CPMD is very powerful, it is too computationally intensive to be performed routinely on current computing platforms, which limit accessible time and spatial scales. The goal of this project is to design and build a framework to seamlessly enable new CPMD applications on future HPC platforms and thereby greatly expand the domain of utility of this critically important applications software. In light of CPMD's tremendous promise to advance science and technology, the exciting corresponding IT research challenges, and the educational opportunities these challenges embody, it is appropriate and timely to form a tightly knit, diverse Collaborative which combines application scientists (Car, Klein, Martyna, Tuckerman, Selloni) with both hardware (Torrellas) and software computer scientists (Kal, Nystrom) to aggressively advance the state of the art. The Collaborative will: (1) improve the methodology and sampling algorithms to permit larger scale simulations of higher accuracy; (2) employ next generation software engineering tools to design and implement parallel algorithms, which scale on thousands of processors in the near term and lead to scaling on much larger machines in the far term; (3) make available new, extensible object-oriented open source software modules for CPMD to the community via the Web with annual upgrades; (4) regularly carry out performance analyses, which will be used to fuel the design of novel hardware architectures; (5) use the new techniques and software to carry out applications projects in technologically important areas; (6) transfer the knowledge thus gained to the community through diversity- conscious outreach and education programs ranging from the secondary to graduate levels as well as through proven international academic and industrial collaborations.
Showing the most recent 10 out of 292 publications
