Research Abstracts Online
January - December 2011
University of Minnesota Twin Cities
College of Science and Engineering
of Computer Science and Engineering
PI: Yousef Saad, Fellow
Algorithms for Inhomogeneous Quantum-Mechanical Systems; Library of Distributed-Memory Sparse Iterative Solvers; PARSEC; and Materials Informatics
These researchers are involved in three projects using MSI. The purpose of the first project is to develop new algorithms for determining the diagonal of the inverse of a Green's function, within the dynamical mean-field theory, as well as determining the trace of a function of matrix Hamiltonian, within the numerical-linked cluster approach. Functions of matrices are currently at the heart of intense research for computational physics applications, as well as for other fields of research. Sparse technical methods based on graph theory are being used within this project.
The goal of the second project, known as p-ARMS, is to investigate robust preconditioning techniques for solving general large sparse linear systems with an emphasis on parallel techniques. The software is intended to be used in various applications, such as in the work described above.
The goal of the third project is to investigate patterns in the properties of various crystalline structures using state-of-the-art data mining methods. The code PARSEC as well as crystal structure database are combined and used in conjunction with data-mining tools, such as the k-nearest neighbors algorithm, and others.
Pierre Carrier, Research Associate
Haw-Ren Fang, Research Associate
Da Gao, Research Associate
Leeor Kronik, Weizmann Institute of Science, Rehobot, Israel
Ruipeng Li, Graduate Student
Lan Luo, Graduate Student
Thanh T. Ngo, Graduate Student
Daniel Osei-Kuffor, Graduate Student
Yaugen Vecharynski, Research Associate