High Performance Computations on Ionic Liquids, Deep Eutectic Propellants and Propellant Degradation
Abstract
Proposed herein is a heterogeneous computer cluster that is designed to (a) enable the PIs to accomplish challenging calculations that are relevant to the Air Force and (b) perform an extensive set of benchmark calculations to inform the DOD user community with regard to the efficacy of two very different accelerators, graphical processing units (GPU) and ARM processors. The calculations that will be accomplished on the proposed cluster will be a combination of high-level electronic structure (ES) and classical molecular dynamics (MD) simulations on ionic liquids, deep eutectic propellants, and the mechanisms of the aging of stored propellants. The ES calculations will primarily make use of the General Atomic and Molecular Electronic Structure System (GAMESS), while the MD simulations will primarily use LAAMPS. Both of these program suites are available for free downloads. GAMESS and its C++ LibCChem-CUDA library has many of its key functionalities optimized for GPUs, and the GAMESS developers have extensive experience with both GPU and ARM architectures. Likewise, the essential features of LAAMPS have been implemented on GPU architectures. The proposed cluster system will be applied to problems that are important to the Air Force. These include studies of energetic ionic liquids (EILs), deep eutectic solvents-propellants (DeEP), and the mechanisms by which stored propellants age. All of these problems require significant computer power. The EIL and DeEP calculations will be performed in collaboration with Jerry Boatz and his colleagues at AFRL-Edwards who will have access to the cluster, and in coordination with Dr. Laura Brown at the High Performance Computing Modernization Office. Several of the functionalities in GAMESS have already been implemented on GPUs. These functionalities include closed shell Hartree-Fock, (HF), second order perturbation theory (MP2), MP2 in the resolution of the identity (RI) implementation and coupled cluster theory (CCSD(T) and RI-CCSD(T)) energies. Analytic gradients for these methods are currently being implemented and will be available by the time the proposed computer cluster becomes available. Also in progress is the implementation of our EFP and EFMO fragmentation methods that are designed to take great advantage of highly parallel computer systems. The ARM system already works for GAMESS. The LAMMPS GPU package was developed by Mike Brown (now at Intel Corp.) and his collaborators, particularly Trung Nguyen (now at Northwestern). Support for AMD GPUs via HIP was added by Vsevolod Nikolskiy and coworkers at HSE University. The ionic liquid applications include the role of the cation in hypergolicity, the effect of water on the utility of ionic liquids as fuels, and the role of B-containing additives. Likewise, the deep eutectic propellant (DeEP) investigations will address the hydrogen bonding networks between the component ionic liquids and the additive as a function of the size of the molecular clusters, in addition to molecular dynamics simulations to predict properties of DeEP. The analysis of aging propellants will consider the most likely decomposition products, as well as the associated barrier heights and mechanisms. The new theory that is already or about to be implemented include a new, very fast Hartree-Fock self-consistent field method that takes optimal advantage of graphical processing units (GPUs), new second order perturbation theory and coupled cluster theory implementations on both CPUs and GPUs, the development of new force fields using machine learning based on quantum-based potentials. The research proposed herein will provide solutions to the aforementioned types of problems by applying new state-of-the-art methods that combine new approaches in quantum chemistry and classical force field-based simulations, in an integrated quantum-classical approach, while making use of modern heterogeneous computer architectures.
Document Details
- Document Type
- DoD Grant Award
- Publication Date
- Feb 29, 2024
- Source ID
- FA95502310052
Entities
People
- Mark S. Gordon
Organizations
- Air Force Office of Scientific Research
- Iowa State University
- United States Air Force