Michael Francis Pasik

A load-balancing algorithm for a parallel electromagnetic particle-in-cell code

Abstract Particle-in-cell simulations often suffer from load-imbalance on parallel machines due to the competing requirements of the field-solve and particle-push computations. We propose a new algorithm that balances the two computations independently. The grid for the field-solve computation is statically partitioned. The particles within a processors sub-domain(s) are dynamically balanced by migrating spatially-compact groups of particles from heavily loaded processors to lightly loaded ones as needed. The algorithm has been implemented in the quicksilver electromagnetic particle-in-cell code. We provide details of the implementation and present performance results for quicksilver running models with up to a billion grid cells and particles on thousands of processors of a large distributed-memory parallel machine.

Electromagnetic Analysis of Forces and Torques on the Baseline and Enhanced ITER Shield Modules due to Plasma Disruption

An electromagnetic analysis is performed on the ITER shield modules under different plasma-disruption scenarios using the OPERA-3d software. The models considered include the baseline design as provided by the International Organization and an enhanced design that includes the more realistic geometrical features of a shield module. The modeling procedure is explained, electromagnetic torques are presented, and results of the modeling are discussed.

Transient Electromagnetic Modeling of the ZR Accelerator Water Convolute and Stack

The ZR accelerator is a refurbishment of Sandia National Laboratories Z accelerator [1]. The ZR accelerator components were designed using electrostatic and circuit modeling tools. Transient electromagnetic modeling has played a complementary role in the analysis of ZR components [2]. In this paper we describe a 3D transient electromagnetic analysis of the ZR water convolute and stack using edge-based finite element techniques.

Application of the PML Absorbing Boundary Condition to the Analysis of Patch Antennas

ABSTRACT The Berenger Perfectly Matched Layer (PML) absorbing boundary condition is applied to a Finite-Difference Time-Domain analysis of rectangular patch antennas. Both microstrip line and probe-fed patch antennas are analyzed. The effects of the proximity of the PML regions on the resonant characteristics of the antennas and the mutual coupling between antennas are examined. The computational overhead associated with the use of the PML is also considered. Numerical results are validated by comparisons with measured data.

Electromagnetic analysis and modeling of the coax-to-triplate transition for the pulse-compression section the ZR accelerator

Transverse electromagnetic (TEM) wave analysis is used to estimate the efficiencies of the coax to triplate transition in Sandias Z-20 test module. The structure of both the TEM mode and higher order TE modes in the triplate transmission line are characterized. In addition, three dimensional time domain simulations are carried out and used in conjunction with the modal analysis to provide insight into the wave structure excited in the triplate transmission line.

Dynamic load-balancing for a parallel electromagnetic particle-in-cell code

QUICKSILVER is a 3-D electromagnetic particle-in-cell simulation code developed and used at Sandia to model relativistic charged particle transport. It was originally written for shared-memory, multi-processor supercomputers such as the Cray X/MP. A new parallel version of QUICKSILVER has been developed to enable large-scale simulations to be efficiently run on massively-parallel distributed memory supercomputers with thousands of processors, such as the DOE ASCI (Accelerated Strategic Computing Initiative) platforms. The new parallel code implements all features of the original QUICKSILVER and runs on any platform that supports the message-passing interface (MPI) standard as well as on single-processor workstations. The original QUICKSILVER code was based on a multiple-block grid, which provided a natural strategy for extending the code to partition a simulation among multiple processors. By adding the automated capability to divide QUICKSILVERs existing blocks into subblocks and then distribute those subblocks among processors, a simulations spatial domain can be easily and efficiently partitioned. Based upon this partitioning scheme as well as QUICKSILVERs existing particle-handling infrastructure, an efficient algorithm has been developed for dynamically rebalancing the particle workload on a timestep-by-timestep. This paper will elaborate on the strategies used and describe the algorithms developed to parallelize and dynamically load-balance the code. Results of several benchmark simulations will be presented that illustrate the codes performance and parallel efficiency for a wide variety of simulation conditions. These calculations have as many as 10/sup 8/ grid cells and 10/sup 9/ particles and were run on thousands of processors.

EMPHASIS/Nevada UTDEM User Guide Version 1.0

The Unstructured Time-Domain ElectroMagnetics (UTDEM) portion of the EMPHASIS suite solves Maxwells equations using finite-element techniques on unstructured meshes. This document provides user-specific information to facilitate the use of the code for applications of interest. UTDEM is a general-purpose code for solving Maxwells equations on arbitrary, unstructured tetrahedral meshes. The geometries and the meshes thereof are limited only by the patience of the user in meshing and by the available computing resources for the solution. UTDEM solves Maxwells equations using finite-element method (FEM) techniques on tetrahedral elements using vector, edge-conforming basis functions. EMPHASIS/Nevada Unstructured Time-Domain ElectroMagnetic Particle-In-Cell (UTDEM PIC) is a superset of the capabilities found in UTDEM. It adds the capability to simulate systems in which the effects of free charge are important and need to be treated in a self-consistent manner. This is done by integrating the equations of motion for macroparticles (a macroparticle is an object that represents a large number of real physical particles, all with the same position and momentum) being accelerated by the electromagnetic forces upon the particle (Lorentz force). The motion of these particles results in a current, which is a source for the fields in Maxwells equations.

Load-balancing techniques for a parallel electromagnetic particle-in-cell code

QUICKSILVER is a 3-d electromagnetic particle-in-cell simulation code developed and used at Sandia to model relativistic charged particle transport. It models the time-response of electromagnetic fields and low-density-plasmas in a self-consistent manner: the fields push the plasma particles and the plasma current modifies the fields. Through an LDRD project a new parallel version of QUICKSILVER was created to enable large-scale plasma simulations to be run on massively-parallel distributed-memory supercomputers with thousands of processors, such as the Intel Tflops and DEC CPlant machines at Sandia. The new parallel code implements nearly all the features of the original serial QUICKSILVER and can be run on any platform which supports the message-passing interface (MPI) standard as well as on single-processor workstations. This report describes basic strategies useful for parallelizing and load-balancing particle-in-cell codes, outlines the parallel algorithms used in this implementation, and provides a summary of the modifications made to QUICKSILVER. It also highlights a series of benchmark simulations which have been run with the new code that illustrate its performance and parallel efficiency. These calculations have up to a billion grid cells and particles and were run on thousands of processors. This report also serves as a user manual for people wishing to run parallel QUICKSILVER.

Analysis of airframe-mounted antennas using parallel and hybridized finite-element time-domain methods

The application of numerical methods to electrically large, three-dimensional geometries typically requires the use of parallel-processing techniques. The distributed parallelization of the finite-element time-domain-finite-difference time-domain (FETD-FDTD) hybrid is described. An application is provided for antenna radiation on a realistic airframe. The parallelization procedure requires two domain decompositions: one for the structured, finite-difference grid, and the other for the unstructured, finite-element grid. The two grids communicate across the interface shown. The FETD method and the interface to FDTD are described, and the parallelization implementation is presented together with applications.

New self-magnetically insulated connection of multi-level accelerators to a common load

We have developed a new type of convolute called the Clam Shell MITL (CSMITL) to couple multi-level accelerators to a common load. The CSMITL has magnetic nulls only at large radius where the cathode electric field is kept below the threshold for emission, has only a simply connected magnetic topology to avoid plasma motion along magnetic field lines into highly stressed gaps, and has electron injectors that ensure efficient electron flow even in the limiting case of self-limited MITLs. We report the first experimental results on a CSMITL, which convolutes two disk feeds on the Saturn accelerator into a single disk feed. Experiments with a high impedance electron beam load operating at twice the self-limited impedance of the CSMITL confirm key design features and demonstrate robust operation.