Research Projects and Softwares

The FEAST Eigenvalue Solver

The FEAST solver package is a high-performance comprehensive numerical library for solving the standard or generalized eigenvalue problem, and obtaining all the eigenvalues and eigenvectors within a given search interval. It is based on an innovative fast and stable numerical algorithm presented in [18] -- named the FEAST algorithm -- which deviates fundamentally from the traditional Krylov subspace iteration based techniques (Arnoldi and Lanczos algorithms) or other Davidson-Jacobi techniques. The FEAST algorithm takes its inspiration from the density-matrix representation and contour integration technique in quantum mechanics. The FEAST algorithm combines simplicity and efficiency and offers many important capabilities for achieving high performance, robustness, accuracy, and scalability on parallel architectures. The current version of the FEAST package focuses on solving the symmetric eigenvalue problems (real symmetric or complex Hermitian systems) on a shared-memory architecture (e.g one multicore node).

SPIKE: A Banded Linear System Solver

SPIKE is a poly-algorithm that uses many different strategies to solve large banded systems in parallel. It uses a novel decomposition method to balance computation against communication requirements. The solver is parallelized with MPI to take advantage of high-performance computing (HPC) clusters and other parallel architectures. The large number of options/decision schemes available for SPIKE created the need for the automatic generation of a sophisticated decision tree --SPIKE-ADAPT-- that has been developed by Intel. In June 2008, SPIKE and SPIKE-ADAPT have been regrouped into one comprehensive package - named Intel Adaptive Spike-Based Solver - which has been released in the Intel whatif website. SPIKE offers HPC users a new and valuable tool for solving large linear systems.

NESSIE: A Nanoelectronics Simulation Environment

NESSIE is a multidimensional finite element nanoelectronics simulation environment for solving the self-consistent transport/electrostatics problem for arbitrary devices. From 1998 to 2005, NESSIE has been used for exploring new modeling techniques for multidimensional transport models and the essential numerical parallel algorithms. The effective mass approach has been used to study a wide range of characteristics (current-voltage, etc.) of many realistic quantum structures (2D MOSFET and DG-MOSFET, 3D III-V interference devices, 3D Silicon nanowire transistors, 3D carbon nanotubes etc.). Elements of NESSIE have been used for developing the packages Nanowires and CNTFET Lab which enable on-line simulations of 3D silicon nanowires and carbon nanotube transistors. Since 2006, my group and I have been working on a more fundamental approach to redesign the whole NESSIE real-space mesh framework with the aim to simulate arbitrary atomistic structures and devices (composed of thousand to million of atoms) using first-principle calculations.