A5027179168- Plasma Diagnostics and Applications
- Electromagnetic Simulation and Numerical Methods
- Vacuum and Plasma Arcs
- Particle accelerators and beam dynamics
- Gyrotron and Vacuum Electronics Research
- Computational Fluid Dynamics and Aerodynamics
- Advanced Sensor Technologies Research
- Parallel Computing and Optimization Techniques
- Electromagnetic Scattering and Analysis
- Particle Accelerators and Free-Electron Lasers
- Electron and X-Ray Spectroscopy Techniques
- Aerosol Filtration and Electrostatic Precipitation
- Fluid Dynamics and Turbulent Flows
- Pulsed Power Technology Applications
- Advanced Data Storage Technologies
- Distributed and Parallel Computing Systems
- Magnetic confinement fusion research
- Cloud Computing and Resource Management
- Astrophysics and Cosmic Phenomena
- Electrostatic Discharge in Electronics
- Plasma Applications and Diagnostics
- Gas Dynamics and Kinetic Theory
- Laser-Plasma Interactions and Diagnostics
- Particle Detector Development and Performance
- Matrix Theory and Algorithms
Nvidia (United States)
2023
Sandia National Laboratories California
2012-2022
Sandia National Laboratories
2012-2021
United States Air Force Research Laboratory
2004-2010
Wright-Patterson Air Force Base
2007
Stennis Space Center
2003-2004
University of Southern Mississippi
2004
Nielsen Engineering & Research (United States)
1998-1999
Atomistic machine learning focuses on the creation of models that obey fundamental symmetries atomistic configurations, such as permutation, translation, and rotation invariances. In many these schemes, translation invariance are achieved by building scalar invariants, e.g., distances between atom pairs. There is growing interest in molecular representations work internally with higher rank rotational tensors, vector displacements atoms, tensor products thereof. Here, we present a framework...
Trilinos is an object-oriented software framework for the solution of large-scale, complex multi-physics engineering and scientific problems. While was originally designed scalable solutions large problems, fidelity needed by many simulations significantly greater than what one could have envisioned two decades ago. When problem sizes exceed a billion elements even applications solver stacks require complete revision. The second-generation employs C++ templates in order to solve arbitrarily...
In this paper we introduce EMPIRE-PIC, a finite element method particlein-cell (FEM-PIC) application developed at Sandia National Laboratories.The code has been in C++ using the Trilinos library and Kokkos Performance Portability Framework to enable running on multiple modern compute architectures while only requiring maintenance of single codebase.EMPIRE-PIC is capable solving both electrostatic electromagnetic problems two-and three-dimensions second-order accuracy space time.In validate...
Shaped cathodes enhance key magnetron performance parameters, such as output power, power efficiency, and impedance. Simulations using these novel designs showed a dramatic increase in the range of magnetic field which functions, doubling an elimination mode competition, halving startup time.
The finite-difference time-domain technique for simulation of electromagnetic and low-density plasma phenomena is computationally expensive can require tens thousands computer hours to produce one solution. Substantial gains be made through memory streamlining (factors 2.3times faster), efficient cache usage 3times improvement), better parallel design (improving scalability four times the number CPUs). These improvements are documented tested across five different supercomputing hardware...
An explicit electrostatic particle-in-cell (PIC) code with complex boundary conditions and direct simulation Monte Carlo (DSMC) particle collisions is utilized to investigate one dimensional current breakdown between two electrodes separated by air at STP. The model includes Auger neutralization cold field electron emission from the cathode as well electron-neutral elastic, ionization, excitation interactions. simulated voltages various electrode gap sizes are compared experimental data...
Trilinos is an object-oriented software framework for the solution of large-scale, complex multi-physics engineering and scientific problems. While original version was designed highly scalable solutions large problems, need increasingly higher fidelity simulations has pushed problem sizes beyond what could have been envisioned two decades ago. When exceed a billion elements even applications solver stacks require complete revision. The next-generation employs C++ templates in order to solve...
The present paper describes a new tool kit which can be used to model the time-dependent response of nonlinear systems. Indicial Prediction System (IPS) applies indicial theory solve complex unsteady problems, such as those associated with aerodynamic phenomena during maneuver aircraft. functionality this system and its capabilities are described through numerous examples. An important demonstration method is application prediction loads on 65-degree delta wing undergoing forced body-axis...
An implementation of a distributed model coupling framework is documented. This provides the infrastructure for data-flow approach solving problem numerical models sharing information. There exists centralized server which stores information such as surface fluxes. then passed to client applications (numerical models) through series filters. These filters are used transform into ready-to-use form by and specific process being performed. CORBA all communication between processes. Results...
A new algorithm was developed, which reduces the self-force in particle-in-cell codes on unstructured meshes a predictable and controllable way. This is accomplished by computing charge density weighting function for particle, reproduces Green's solution to Poisson's equation at nodes when using standard finite element method methodology. provides superior local potential allows particle-particle particle-mesh techniques be used subtract off force contributions, including fictitious...
In this work, we present a discontinuous-Galerkin method for evolving relativistic hydrodynamics. We include an exploration of analytical and iterative methods to recover the primitive variables from conserved ideal equation state Taub-Matthews approximation Synge state. also new operator enforcing physically permissible at all basis points within element while preserving volume average implement using Kokkos performance-portability library enable running performance on both CPUs GPUs. use...
This paper explores dynamic load balancing algorithms used by asynchronous many-task (AMT), or 'taskbased', programming models to optimize task placement for scientific applications with workload imbalances. AMT use overdecomposition of the computational domain. Overdecompostion provides a natural mechanism domain developers expose concurrency and break their into pieces that can be remapped different hardware. fully distributed strategies have shown great promise exascale-level computing...
This paper reports on radio frequency (RF) sources that have been virtually-prototyped with the Improved Concurrent Electromagnetic Particle-in-Cell (ICEPIC) code. ICEPIC simulates from first-principles, Maxwell's equations and Lorenz's force law, electrodynamics charged particle dynamics of RF-producing part system. Our simulations focus several proposed variants L-band A66 half height (A66-hh) relativistic magnetron HPM source. The variations A66-hh include higher designs (C-band, S-band...
A block base sparse approximate inverse preconditioner for the electric field integral equations is documented and tested. It utilized Kokkos library performance portability shows superior when compared to a direct method, 36x faster 112K DOF problem. Furthermore, due abstractions available in it allows one migrate from CPU GPU trivial way.
This paper documents the changes required to permit ICEPIC scale efficiently thousand CPU range. Substantial were made communication paradigm within code, so that only one synchronization point is now required. led increase of a factor four in number processors can productively use on real world problems