Graphical processing unit accelerated state-averaged complete active space self-consistent field for large-scale quantum dynamics simulations
Abstract/Contents
- Abstract
- Photochemical reactions typically occur under nonequilibrium conditions on the femtosecond to picosecond timescale. In many cases, the nonequilibrium dynamical behavior of the system has a significant impact on the reaction, and statistical and reaction path methods are unreliable for their study. As a result, nonadiabatic dynamics simulations are important for the study of photochemistry. The most generally applicable and flexible approach is based on ab initio molecular dynamics, where the electronic Schrödinger equation is solved alongside the dynamics to obtain the required potential energy surfaces and nonadiabatic couplings as needed. However, these ab initio nonadiabatic dynamics simulations are computationally expensive. In particular, the necessary electronic structure calculations are the computational bottleneck, exhibiting both a high prefactor and unfavorable scaling behavior. This scaling is especially unfavorable for systems that cannot be well described by a single electronic configuration, and the multiconfiguration/multireference methods needed to describe these electronic excited states have been limited to small molecular systems. We alleviate this problem by implementing graphical processing unit (GPU) accelerated state-averaged complete active space self-consistent field (SA-CASSCF). Our algorithms are ~10-100x faster than the nearest competitive code and exhibit effective scaling of O(N^2), versus formal scaling of O(N^4). We will show how SA-CASSCF response properties, including nuclear gradients and nonadiabatic coupling vectors, can be efficiently computed for arbitrary, tractable active spaces. To demonstrate the utility of this development to nonadiabatic dynamics simulations, we have applied GPU accelerated SA-CASSCF coupled to ab initio multiple spawning (AIMS) to unravel the photodynamics of provitamin D3, the largest simulation of its kind. Finally, we will discuss how empirical corrections can be employed to bring SA-CASSCF simulations into close agreement with more sophisticated, but computationally intractable, multireference electronic structure methods, further expanding the utility of the SA-CASSCF method.
Description
| Type of resource | text |
|---|---|
| Form | electronic; electronic resource; remote |
| Extent | 1 online resource. |
| Publication date | 2017 |
| Issuance | monographic |
| Language | English |
Creators/Contributors
| Associated with | Snyder Jr, James William |
|---|---|
| Associated with | Stanford University, Department of Chemistry. |
| Primary advisor | Martinez, Todd J. (Todd Joseph), 1968- |
| Thesis advisor | Martinez, Todd J. (Todd Joseph), 1968- |
| Thesis advisor | Boxer, Steven G. (Steven George), 1947- |
| Thesis advisor | Markland, Thomas E |
| Advisor | Boxer, Steven G. (Steven George), 1947- |
| Advisor | Markland, Thomas E |
Subjects
| Genre | Theses |
|---|
Bibliographic information
| Statement of responsibility | James William Snyder Jr. |
|---|---|
| Note | Submitted to the Department of Chemistry. |
| Thesis | Thesis (Ph.D.)--Stanford University, 2017. |
| Location | electronic resource |
Access conditions
- Copyright
- © 2017 by James William Snyder Jr
- License
- This work is licensed under a Creative Commons Attribution Non Commercial 3.0 Unported license (CC BY-NC).
Also listed in
Loading usage metrics...