Controls and Computing

High-Performance Computing

Paper Title Page
FPAT079 Data Base Extension for the Ensemble Model Using a Flexible Implementation 4036
  • W. Ackermann, T. Weiland
    TEMF, Darmstadt
  Funding: Work supported by DESY, Hamburg.

To guarantee an adequate design and a proper functionality of various machine components it is of great importance to perform detailed studies of charged particle transport. However, it is often not necessary to initiate individual kinetic simulations. When the evolution of integral quantities is of research interest, it is worth treating an investigated particle ensemble as a whole and applying a macroscopic formulation. Using a collision-less kinetic approach, the simplified model is derived from the well-known Vlasov equation. Instead of solving directly this equation, one can use moments of the density function obtained by means of an averaging process. This formalism had been implemented into the beam dynamics simulation program V-Code and a fundamental database of various beam line elements like cavities, drift spaces, solenoids, quadrupoles and steerers was set up. A flexible realization of the C++ code representing the cavities and the drift spaces can be automatically used for an arbitrary order of moments applying a symbolic algebra program. A useful extension to the remaining beam line elements together with appropriate simulation results is presented in the paper.

FPAT080 Simulations of Beam Injection and Extraction into Ion Sources 4069
  • M. Cavenago
    INFN/LNL, Legnaro, Padova
  Funding: INFN-LNL

Charge breeding, consistiting of injecting singly charged ion into ECRIS(Electron Cyclotron Resonance Ion Sources) to extract an highly charged ion beam, is a promising technique for rare or radioactive ion beam. Efficiency and extracted beam temperature are dominated by the strong collisional diffusion of charged ion inside source. A computer code, named BEAM2ECR, written to simulate details of the injection, ionization, collision and extraction processes is described.* A model of injection plasma sheath and of source fringe field were recently added. Neutral injection is also supported, for comparison with other techniques, like gas feeding or metal vapor injection. Results, clearly favouring near axis injection for most cases are described. Code is written in C-language and possibility of concurrent execution over a Linux cluster was recently added.

*M. Cavenago, O. Kester, T. Lamy and P. Sortais, Rev. Sci. Instrum. 73, 537 (2002).

FPAT081 A New Version of SixTrack with Collimation and Aperture Interface 4084
  • G. Robert-Demolaize, R.W. Assmann, S. Redaelli, F. Schmidt
    CERN, Geneva
  Simulations of collimation and beam cleaning were so far often performed with simplified computer models. However, the increase in available CPU power has opened the possibility for far more realistic simulations. For large accelerators like LHC it is now possible to track millions of particles, element by element over hundreds of turns. The well established SixTrack code treats the full six-dimensional phase space and considers the non-linear magnet components up to very high order. This code is being used for all LHC tracking simulations and has well developed linear and non-linear error models. SixTrack was extended for tracking of large ensembles of halo particles, taking into account halo interaction with arbitrarily placed collimators. An interface to a program for aperture analysis allows obtaining beam loss maps in the machine aperture. A standardized and portable SixTrack version is now available, providing all functionality of the old SixTrack, as well as the newly added support for halo tracking, collimation and aperture loss maps.  
FPAT082 From Visualisation to Data Mining with Large Data Sets 4114
  • A. Adelmann
    PSI, Villigen
  • R.D. Ryne, J.M. Shalf, C. Siegerist
    LBNL, Berkeley, California
  In 3D particle simulations, the generated 6D phase space data are can be very large due to the need for accurate statistics, sufficient noise attenuation in the field solver and tracking of many turns in ring machines or accelerators. There is a need for distributed applications that allow users to peruse these extremely large remotely located datasets with the same ease as locally downloaded data. This paper will show concepts and a prototype tool to extract useful physical information out of 6D raw phase space data. ParViT allows the user to project 6D data into 3D space by selecting which dimensions will be represented spatially and which dimensions are represented as particle attributes, and the construction of complex transfer functions for representing the particle attributes. It also allows management of time-series data. An HDF5-based parallel-I/O library, with C++, C and Fortran bindings simplifies the interface with a variety of codes. A number of hooks in ParVit will allow it to connect with a parallel back-end that is able to provide remote file access, progressive streaming, and even parallel rendering of particle sets in excess of 1Billion particles.  
FPAT083 H5Part: A Portable High Performance Parallel Data Interface for Particle Simulations 4129
  • A. Adelmann
    PSI, Villigen
  • R.D. Ryne, J.M. Shalf, C. Siegerist
    LBNL, Berkeley, California
  Largest parallel particle simulations, in six dimensional phase space generate wast amont of data. It is also desirable to share data and data analysis tools such as ParViT (Particle Visualization Toolkit) among other groups who are working on particle-based accelerator simulations. We define a very simple file schema built on top of HDF5 (Hierarchical Data Format version 5) as well as an API that simplifies the reading/writing of the data to the HDF5 file format. HDF5 offers a self-describing machine-independent binary file format that supports scalable parallel I/O performance for MPI codes on a variety of supercomputing systems and works equally well on laptop computers. The API is available for C, C++, and Fortran codes. The file format will enable disparate research groups with very different simulation implementations to share data transparently and share data analysis tools. For instance, the common file format will enable groups that depend on completely different simulation implementations to share custom data analysis tools like ParViT without modification. We will show examples and benchmak data for various platforms.  
FPAT085 The TAO Accelerator Simulation Program 4159
  • D. Sagan
    Cornell University, Department of Physics, Ithaca, New York
  • J.C. Smith
    Cornell University, Laboratory for Elementary-Particle Physics, Ithaca, New York
  Funding: NSF and DOE.

A new accelerator design and analysis simulation environment based on the BMAD relativistic charged particle dynamics library is in development at Cornell University. Called TAO (Tool for Accelerator Optimization), it is a machine independent program that implements the essential ingredients needed to solve simulation problems. This includes the ability to: 1. Design lattices subject to constraints, 2. Simulate errors and changes in machine parameters, and 3. Simulate machine commissioning including simulating data measurement and correction. TAO is designed to be easily customizable so that extending it to solve new and different problems is straight forward. The capability to simultaneously model multiple accelerator lattices, both linacs and storage rings, and injection from one lattice to another allows for the design and commissioning of large multi stage accelerators. It can also simultaneously model multiple configurations of a single lattice. Single particle, particle beam and macroparticle tracking is implemented. Use of TAO with both the International Linear Collider and the Cornell Energy Recovery Linac are provided as examples.

FPAT086 Lucretia: A Matlab-Based Toolbox for the Modeling and Simulation of Single-Pass Electron Beam Transport Systems 4197
  • P. Tenenbaum
    SLAC, Menlo Park, California
  We report on Lucretia, a new simulation tool for the study of single-pass electron beam transport systems. Lucretia supports a combination of analytic and tracking techniques to model the tuning and operation of bunch compressors, linear accelerators, and beam delivery systems of linear colliders and linac-driven Free Electron Laser (FEL) facilities. Extensive use of Matlab scripting, graphics, and numerical capabilities maximize the flexibility of the system, and emphasis has been placed on representing and preserving the fixed relationships between elements (common girders, power supplies, etc.) which must be respected in the design of tuning algorithms. An overview of the code organization, some simple examples, and plans for future development are discussed.  
FPAT087 elegantRingAnalysis: An Interface for High-Throughput Analysis of Storage Ring Lattices Using elegant 4200
  • M. Borland
    ANL, Argonne, Illinois
  Funding: Work supported by the U.S. Department of Energy, Office of Basic Energy Sciences, under Contract No. W-31-109-ENG-38.

The code {\tt elegant} is widely used for simulation of linacs for drivers for free-electron lasers. Less well known is that elegant is also a very capable code for simulation of storage rings. In this paper, we show a newly-developed graphical user interface that allows the user to easily take advantage of these capabilities. The interface is designed for use on a Linux cluster, providing very high throughput. It can also be used on a single computer. Among the features it gives access to are basic calculations (Twiss parameters, radiation integrals), phase-space tracking, nonlinear dispersion, dynamic aperture (on- and off-momentum), frequency map analysis, and collective effects (IBS, bunch-lengthening). Using a cluster, it is easy to get highly detailed dynamic aperture and frequency map results in a surprisingly short time.

FPAT088 Advanced Beam-Dynamics Simulation Tools for RIA 4218
  • R.W. Garnett, J.A. Billen, T.P. Wangler
    LANL, Los Alamos, New Mexico
  • K.R. Crandall
    TechSource, Santa Fe, New Mexico
  • P.N. Ostroumov
    ANL, Argonne, Illinois
  • J. Qiang, R.D. Ryne
    LBNL, Berkeley, California
  • R.C. York, Q. Zhao
    NSCL, East Lansing, Michigan
  Funding: U.S. Department of Energy Contract W-7405-ENG-36.

We are developing multuparticle beam-dynamics simulation codes for RIA driver linac simulations extending from the low-energy beam transport line to the end of the linac. These codes run on the NERSC parallel supercomputing platforms at LBNL, which allow us to run simulations with large numbers of macroparticles. The codes have physics capabilities needed for RIA, including transport and acceleration of multiple-charge-state beams, beam-line elements such as high-voltage platforms within the linac, interdigital accelerating structures, charge-stripper foils, and capabilities for handling the effects of machine errors and other off-normal conditions. In this paper we present the status of the work, describe some recent additions to the codes, and show preliminary end-to-end simulation results for a representative driver-linac design.

FPAT089 A Parallel Simplex Optimizer and Its Application to High-Brightness Storage Ring Design 4230
  • H. Shang, M. Borland
    ANL, Argonne, Illinois
  Funding: Work supported by the U.S. Department of Energy, Office of Basic Energy Sciences, under Contract No. W-31-109-ENG-38.

Optimization is commonly used in accelerator design to find linear optics solutions. Such optimizations are usually fairly fast as linear optics computations are themselves fast. For high-brightness storage rings, optimization of nonlinear elements (e.g., sextupoles) is also important in obtaining sufficient dynamic aperture. However, this can be very time onsuming as the basic calculations are time consuming. We have developed an efficient parallel Simplex optimizer that runs on a Linux cluster. It can optimize the result of running essentially any program or script that returns a penalty function value. We have used this optimizer with elegant to optimize dynamic aperture of storage ring designs. We discuss the optimization algorithm and performance, design of penalty functions, and optimization results.

FPAT091 LiTrack: A Fast Longitudinal Phase Space Tracking Code with Graphical User Interface 4266
  • P. Emma, K.L.F. Bane
    SLAC, Menlo Park, California
  Funding: Work supported by U.S. Department of Energy contract DE-AC02-76SF00515.

Many linear accelerators, such as linac-based light sources and linear colliders, apply longitudinal phase space manipulations in their design, including electron bunch compression and wakefield-induced energy spread control. Several computer codes handle such issues, but most require detailed information on the transverse focusing lattice. In fact, in most linear accelerators, the transverse distributions do not significantly affect the longitudinal, and can be ignored initially. This allows the use of a fast 2D code to study longitudinal aspects without time-consuming considerations of the transverse focusing. LiTrack is based on a 15-year old code (same name) originally written by one of us (KB), which is now a MATLAB-based code with additional features, such as a graphical user interface and output plotting. The single-bunch tracking includes RF acceleration, bunch compression to 3rd order, geometric and resistive wakefields, aperture limits, synchrotron radiation, and flexible output plotting. The code was used to design both the LCLS and the SPPS projects at SLAC and typically runs in <1 minute. We describe the features, show some examples, and provide access to the code.

FPAT092 Optimized Beam Matching Using Extremum Seeking 4269
  • E. Schuster
    Lehigh University, Bethlehem, Pennsylvania
  • C.K. Allen
    LANL, Los Alamos, New Mexico
  • M. Krstic
    UCSD, La Jolla, California
  The transport and matching problem for a low energy transport system is approached from a control theoretical viewpoint. The beam dynamics and transport section is modeled using the KV envelope equations. Principles of optimal control are applied to this model to formulate techniques which aid in the design of the transport and matching section. Multi-Parameter Extremum Seeking, a real-time non-model based optimization technique, is considered in this work for the lens tuning. Numerical simulations illustrate the effectiveness of this approach.