Description: This presentation will review several previous attempts at representing quantum mechanical force fields for complex chemical systems using either classical reactive forcefields or newer Machine Learning forcefields. We will discuss our experiences with DeepMD in detail and offer possible scaling arguments for its failure. We will then open the discussion so the audience can describe their experiences and/or suggest what to try next.
| Presenter: Dr. James Lill, GDIT / PET Location: Webcast Date & Time: February 18, 2025, 3:00p - 4:00p ET |
Controlled by: DoD HPCMP Controlled by: PET Program CUI Category: OPSEC Limited Dissemination Control: FEDCON POC: Mr. Ronald Hedgepeth, pet@hpc.mil |
CUI
- Presenter: James Lill
Description: This seminar will demonstrate LAMMPS scripting capabilities for performing complex simulations with loops, Boolean logic checks, and custom variables with the Python interface. The Python LAMMPS API will also be demonstrated for coupling LAMMPS simulations with external algorithms, like differential equation solvers or AI/ML models.
| Presenter(s): Dr. Mathew Boyer, GDIT/PET Location: Webcast Date & Time: May 18, 2023, 2:00p - 3:30p ET |
Controlled by: DoD HPCMP Controlled by: PET Program CUI Category: OPSEC Limited Dissemination Control: FEDCON POC: Mr. Ronald Hedgepeth, pet@hpc.mil |
CUI
Description: This seminar will demonstrate the capabilities of the PET LAMMPS Test Suite including: tools for construction of LAMMPS data files for amorphous and crystalline systems, automated equilibration routines for solids and liquids, property prediction routines, and usage of GROMACS post-process tools with .xtc binary trajectory files from LAMMPS simulations.
An example will also be presented which includes the addition of a new force field module to the test suite to showcase how users can adapt the codebase to their research problems.
| Presenter(s): Dr. Mathew Boyer, GDIT/PET Location: Webcast Date & Time: May 16, 2023, 2:00p - 3:30p ET |
Controlled by: DoD HPCMP Controlled by: PET Program CUI Category: OPSEC Limited Dissemination Control: FEDCON POC: Mr. Ronald Hedgepeth, pet@hpc.mil |
CUI
Description: NAMD is an efficient HPC molecular dynamics code designed to model biological systems. It works with GPUs and runs distributed across nodes using Charm++. The xtb code implements an efficient and accurate semi-empirical quantum mechanical method, GNF2-xTB, that is parameterized for all the elements of the periodic table up to Radon. The training will show users how to take advantage of the PET developed interface and input preparation scripts to perform mixed quantum mechanics molecular mechanics (QM/MM) simulations with these two codes. Quantum region selection, simulation options, and use of the codes in the HPCMP compute environment will be discussed. These techniques will allow DOD researchers to model free energies of reactions in biological systems and to calculate binding energies for systems where reliable force fields are not yet available.
Presenter(s): Dr. Gary Kedziora, GDIT / PET
Location: Webcast
Date & Time: June 22, 2022, 2:00p - 3:30p ET
Distribution Statement D. Distribution authorized to the Department of Defense and U.S. DoD contractors only, Administrative or Operational Use, 22 June 2022. Other requests for this document shall be referred to the High Performance Computing Modernization Office, 3909 Halls Ferry Road, Vicksburg, MS 39180.
- Presenter: Gary Kedziora