Search

Reactive molecular dynamics is a powerful simulation method for describing chemical reactions. Here, we introduce a new generalized polarizable reactive force-field (ReaxPQ+) model to significantly improve the accuracy by accommodating the reorganization of surrounding media. The increased computation is accelerated by (1) extended Lagrangian approach to eliminate the speed-limiting charge iteration, (2) shift-collapse computation of many-body renormalized n-tuples, which provably minimizes data transfer, (3) multithreading with round-robin data privatization, and (4) data reordering to reduce computation and allow vectorization. The new code achieves (1) weak-scaling parallel efficiency of 0.989 for 131,072 cores, and (2) eight-fold reduction of time-to-solution (T2S) compared with the original code, on an Intel Knights Landing-based computer. The reduced T2S has for the first time allowed purely computational synthesis of atomically-thin transition metal dichalcogenide layers assisted by machine learning to discover a novel synthetic pathway.

Electronic structure calculations are a widely used tool in materials science and large consumer of supercomputing resources. Traditionally, the software packages for these kind of simulations have been implemented in compiled languages, where Fortran in its different versions has been the most popular choice. While dynamic, interpreted languages, such as Python, can increase the effciency of programmer, they cannot compete directly with the raw performance of compiled languages. However, by using an interpreted language together with a compiled language, it is possible to have most of the productivity enhancing features together with a good numerical performance. We have used this approach in implementing an electronic structure simulation software GPAW using the combination of Python and C programming languages. While the chosen approach works well in standard workstations and Unix environments, massively parallel supercomputing systems can present some challenges in porting, debugging and profiling the software. In this paper we describe some details of the implementation and discuss the advantages and challenges of the combined Python/C approach. We show that despite the challenges it is possible to obtain good numerical performance and good parallel scalability with Python based software.

The quantum chemistry package General Atomic and Molecular Electronic Structure System (GAMESS) is employed in the first-principles modeling of complex molecular systems by using the density functional theory (DFT) as well as a number of other post-Hartree-Fock (HF) methods. Both DFT and time-dependent DFT (TDDFT) are of particular interest to the DoD Computational Biology, Chemistry, and Materials Science (CCM). Millions of CPU hours per year are expended by GAMESS calculations on DoD high performance computing (HPC) systems. Therefore, any reduction in wall-clock time for these calculations will represent a significant saving in CPU hours. As part of this work, three areas for improvement were identified: (1) replacement of the exchange-correlation (XC) integration grid, (2) TDDFT parallelization, and (3) profiling and optimization of the DFT and TDDFT codes. We summarize the work performed in these task areas and present the resulting speed-up. Many of our software enhancements are available to the general public in the 11APR2008R1 version of GAMESS with the anticipation that all of them will be available in a future release version of GAMESS., See also ADM002187. Presented at the Army Science Conference (26th) held in Orlando, Florida on 1-4 December 2008. Published in the Proceedings of the Army Science Conference (26th), 2008. This project, CCM-KY6-002, was funded by the DoD High Performance Computing and Modernization Program (HPCMP) under the User Productivity Enhancement and Technology Transfer (PET) program. The original document contains color images.

Quantum mechanical calculations based on density functional theory (DFT) are used to study dynamic behavior of shocked polymeric nitrogen, a novel energetic material. We report results on system sizes in excess of 3,000 atoms. Such calculations on system sizes within the 1,000 atom range remain problematic using standard implementations of DFT. We evaluate the feasibility of using several available DFT codes for this work through comparison of scalability and resource requirements. In this study, we utilize a recently developed highly-scalable localized orbital DFT code, CP2K, designed to treat large systems. Scaling and performance benchmarks of the CP2K on several Department of Defense (DoD) high performance computing (HPC) computers are presented for a variety of system sizes and shapes. Additionally, we report preliminary calculations on the conventional explosive nitromethane. In those calculations in excess of 3,500, atoms are treated., The original document contains color images. All DTIC reproductions will be in black and white. Presented at the DoD High Performance Computer Modernization Program (HPCMP) Users Group Conference (USG)(2007): A Bridge to Future Defense held in Pittsburgh, PA on 18-21 June 2007. Published in Proceedings of the DoD High Performance Computer Modernization Program (HPCMP) Users Group Conference (USG), p104-109, June 2007. Publisher: IEEE Computer Society, Conference Publishing Services (CPS). ISBN 0-7695-3088-5 and ISBN 978-0-7695-3088-8. Sponsored in part by the Department of Defense. This article is from ADA488707 Proceedings of the HPCMP Users Group Conference 2007. High Performance Computing Modernization Program: A Bridge to Future Defense held 18-21 June 2007 in Pittsburgh, Pennsylvania