arXiv subjects

Eric Polizzi

Publications and source records attributed to Eric Polizzi.

28 records · Page 2Linked to original sources

arXiv2014

A New Highly Parallel Non-Hermitian Eigensolver

Calculating portions of eigenvalues and eigenvectors of matrices or matrix pencils has many applications. An approach to this calculation for Hermitian problems based on a density matrix has been proposed in 2009 and a software package called FEAST has been developed. The density-matrix approach allows FEAST's implementation to exploit a key strength of modern computer architectures, namely, multiple levels of parallelism. Consequently, the software package has been well received and subsequently commercialized. A detailed theoretical analysis of Hermitian FEAST has also been established very recently. This paper generalizes the FEAST algorithm and theory, for the first time, to tackle non-Hermitian problems. Fundamentally, the new algorithm is basic subspace iteration or Bauer bi-iteration, except applied with a novel accelerator based on Cauchy integrals. The resulting algorithm retains the multi-level parallelism of Hermitian FEAST, making it a valuable new tool for large-scale computational science and engineering problems on leading-edge computing platforms.

math.NA↗

arXiv2014

FEAST as a Subspace Iteration Eigensolver Accelerated by Approximate Spectral Projection

The calculation of a segment of eigenvalues and their corresponding eigenvectors of a Hermitian matrix or matrix pencil has many applications. A new density-matrix-based algorithm has been proposed recently and a software package FEAST has been developed. The density-matrix approach allows FEAST's implementation to exploit a key strength of modern computer architectures, namely, multiple levels of parallelism. Consequently, the software package has been well received, especially in the electronic structure community. Nevertheless, theoretical analysis of FEAST has lagged. For instance, the FEAST algorithm has not been proven to converge. This paper offers a detailed numerical analysis of FEAST. In particular, we show that the FEAST algorithm can be understood as an accelerated subspace iteration algorithm in conjunction with the Rayleigh-Ritz procedure. The novelty of FEAST lies in its accelerator which is a rational matrix function that approximates the spectral projector onto the eigenspace in question. Analysis of the numerical nature of this approximate spectral projector and the resulting subspaces generated in the FEAST algorithm establishes the algorithm's convergence. This paper shows that FEAST is resilient against rounding errors and establishes properties that can be leveraged to enhance the algorithm's robustness. Finally, we propose an extension of FEAST to handle non-Hermitian problems and suggest some future research directions.

math.NA↗

arXiv2013

Discretization of Time-Dependent Quantum Systems: Real-Time Propagation of The Evolution Operator

We discuss time dependent quantum systems on bounded domains. Our work may be viewed as a framework for several models, including linear iterations involved in time dependent density functional theory (TDDFT), the Hartree-Fock model, or other quantum models. A key aspect of the analysis of the algorithms is the use of time-ordered evolution operators, which allow for both a well-posed problem and its approximation. The approximation theorems obtained for the time-ordered evolution operators complement those in the current literature. We discuss the available theory at the outset, and proceed to apply the theory systematically in later sections via approximations and a global existence theorem for a nonlinear system, obtained via a fixed point theorem for the evolution operator. Our work is consistent with first-principle real time propagation of electronic states, aimed at finding the electronic responses of quantum molecular systems and nanostructures. We present two full 3D quantum atomistic simulations using the finite element method for discretizing the real-space, and the FEAST eigenvalue algorithm for solving the evolution operator at each time step. These numerical experiments are representative of the theoretical results.

math.DS↗

arXiv2013

Non-linear eigensolver-based alternative to traditional SCF methods

The self-consistent procedure in electronic structure calculations is revisited using a highly efficient and robust algorithm for solving the non-linear eigenvector problem i.e. H({ψ})ψ = Eψ. This new scheme is derived from a generalization of the FEAST eigenvalue algorithm to account for the non-linearity of the Hamiltonian with the occupied eigenvectors. Using a series of numerical examples and the DFT-Kohn/Sham model, it will be shown that our approach can outperform the traditional SCF mixing-scheme techniques by providing a higher converge rate, convergence to the correct solution regardless of the choice of the initial guess, and a significant reduction of the eigenvalue solve time in simulations.

physics.comp-ph↗

arXiv2012

FEAST fundamental framework for electronic structure calculations: Reformulation and solution of the muffin-tin problem

In a recent article [1], the FEAST algorithm has been presented as a general purpose eigenvalue solver which is ideally suited for addressing the numerical challenges in electronic structure calculations. Here, FEAST is presented beyond the "black-box" solver as a fundamental modeling framework which can naturally address the original numerical complexity of the electronic structure problem as formulated by Slater in 1937 [2]. The non-linear eigenvalue problem arising from the muffin-tin decomposition of the real-space domain is first derived and then reformulated to be solved exactly within the FEAST framework. This new framework is presented as a fundamental and practical solution for performing both accurate and scalable electronic structure calculations, bypassing the various issues of using traditional approaches such as linearization and pseudopotential techniques. A finite element implementation of this FEAST framework along with simulation results for various molecular systems are also presented and discussed.

cond-mat.mtrl-sci↗

arXiv2010

Spectral-based Propagation Schemes for Time-Dependent Quantum Systems with Application to Carbon Nanotubes

Effective modeling and numerical spectral-based propagation schemes are proposed for addressing the challenges in time-dependent quantum simulations of systems ranging from atoms, molecules, and nanostructures to emerging nanoelectronic devices. While time-dependent Hamiltonian problems can be formally solved by propagating the solutions along tiny simulation time steps, a direct numerical treatment is often considered too computationally demanding. In this paper, however, we propose to go beyond these limitations by introducing high-performance numerical propagation schemes to compute the solution of the time-ordered evolution operator. In addition to the direct Hamiltonian diagonalizations that can be efficiently performed using the new eigenvalue solver FEAST, we have designed a Gaussian propagation scheme and a basis transformed propagation scheme (BTPS) which allow to reduce considerably the simulation times needed by time intervals. It is outlined that BTPS offers the best computational efficiency allowing new perspectives in time-dependent simulations. Finally, these numerical schemes are applied to study the AC response of a (5,5) carbon nanotube within a 3D real-space mesh framework.

physics.comp-ph↗

arXiv2009

A Density Matrix-based Algorithm for Solving Eigenvalue Problems

A new numerical algorithm for solving the symmetric eigenvalue problem is presented. The technique deviates fundamentally from the traditional Krylov subspace iteration based techniques (Arnoldi and Lanczos algorithms) or other Davidson-Jacobi techniques, and takes its inspiration from the contour integration and density matrix representation in quantum mechanics. It will be shown that this new algorithm - named FEAST - exhibits high efficiency, robustness, accuracy and scalability on parallel architectures. Examples from electronic structure calculations of Carbon nanotubes (CNT) are presented, and numerical performances and capabilities are discussed.

cs.CE↗

arXiv2005

A Theoretical Investigation of Surface Roughness Scattering in Silicon Nanowire Transistors

In this letter, we report a three-dimensional (3D) quantum mechanical simulation to investigate the effects of surface roughness scattering (SRS) on the device characteristics of Si nanowire transistors (SNWTs). We treat the microscopic structure of the Si/SiO2 interface roughness directly by using a 3D finite element technique. The results show that 1) SRS reduces the electron density of states in the channel, which increases the SNWT threshold voltage, and 2) the SRS in SNWTs becomes more effective when more propagating modes are occupied, which implies that SRS is more important in planar metal-oxide-semiconductor field-effect-transistors with many transverse modes occupied than in small-diameter SNWTs with few modes conducting.

cond-mat.mes-hall↗

arXiv2004

A Quantum Mechanical Approach for the Simulation of Si/SiO2 Interface Roughness Scattering in Silicon Nanowire Transistors

In this work, we present a quantum mechanical approach for the simulation of Si/SiO2 interface roughness scattering in silicon nanowire transistors (SNWTs). The simulation domain is discretized with a three-dimensional (3D) finite element mesh, and the microscopic structure of the Si/SiO2 interface roughness is directly implemented. The 3D Schrodinger equation with open boundary conditions is solved by the non-equilibrium Green's function method together with the coupled mode space approach. The 3D electrostatics in the device is rigorously treated by solving a 3D Poisson equation with the finite element method. Although we mainly focus on computational techniques in this paper, the physics of SRS in SNWTs and its impact on the device characteristics are also briefly discussed.

cond-mat.mes-hall↗

arXiv2004

A Three-Dimensional Quantum Simulation of Silicon Nanowire Transistors with the Effective-Mass Approximation

The silicon nanowire transistor (SNWT) is a promising device structure for future integrated circuits, and simulations will be important for understanding its device physics and assessing its ultimate performance limits. In this work, we present a three-dimensional quantum mechanical simulation approach to treat various SNWTs within the effective-mass approximation. We begin by assuming ballistic transport, which gives the upper performance limit of the devices. The use of a mode space approach (either coupled or uncoupled) produces high computational efficiency that makes our 3D quantum simulator practical for extensive device simulation and design. Scattering in SNWTs is then treated by a simple model that uses so-called Buttiker probes, which was previously used in metal-oxide-semiconductor field effect transistor (MOSFET) simulations. Using this simple approach, the effects of scattering on both internal device characteristics and terminal currents can be examined, which enables our simulator to be used for the exploration of realistic performance limits of SNWTs.

cond-mat.other↗