Date of Award

January 2018

Document Type

Dissertation

Degree Name

Doctor of Philosophy (PhD)

Department

Computer Science

First Advisor

Mark Hoffmann

Second Advisor

Hassan Reza

Abstract

Configuration interaction (CI) is a post Hartree–Fock method that is commonly used for solving the nonrelativistic Schrödinger equation for quantum many-electron systems of molecular scale. CI includes instantaneous electron correlation and it can deal with the ground state as well as multiple excited states.

The CI matrix is a sparse matrix, and the bigger the CI matrix, the more electron correlation can be captured. However, due to the large size of the CI sparse matrix that is involved in CI computations, a good amount of the time spent on the eigenvalue computations is associated with the multiplication of the CI sparse matrix by numerous dense vectors, which is basically known as Sparse matrix-vector multiplication (SpMV).

Sparse matrix-vector multiplication (SpMV) can be used to solve diverse-scaled linear systems and eigenvalue problems that exist in numerous and varying scientific applications. One of the scientific applications that SpMV is involved in is Configuration Interaction (CI).

In this work, we have developed a new hybrid approach to deal with CI sparse matrices. The proposed model includes a newly-developed hybrid format for storing CI sparse matrices on the Graphics Processing Unit (GPU). In addition to the new developed format, the proposed model includes the SpMV kernel for multiplying the CI matrix (proposed format) by a vector using the C language and the CUDA platform. The proposed SpMV kernel is a vector kernel that uses the warp approach. We have gauged the newly developed model in terms of two primary factors, memory usage and performance.

Our proposed kernel was compared to the cuSPARSE library and the CSR5 (Compressed Sparse Row 5) format and already outperformed both. Our proposed kernel outperformed the CSR5 format by 250.7% and the cuSPARSE library by 395.1%

Keywords— CI, SpMV, Linear System, GPU, Kernel, CUDA.

Download

Share

COinS