Peter Messmer - Academia.edu (original) (raw)

Papers by Peter Messmer

Computing and Software for Big Science, 2022

The IceCube Neutrino Observatory is a cubic kilometer neutrino detector located at the geographic... more The IceCube Neutrino Observatory is a cubic kilometer neutrino detector located at the geographic South Pole designed to detect high-energy astrophysical neutrinos. To thoroughly understand the detected neutrinos and their properties, the detector response to signal and background has to be modeled using Monte Carlo techniques. An integral part of these studies are the optical properties of the ice the observatory is built into. The simulated propagation of individual photons from particles produced by neutrino interactions in the ice can be greatly accelerated using graphics processing units (GPUs). In this paper, we (a collaboration between NVIDIA and IceCube) reduced the propagation time per photon by a factor of up to 3 on the same GPU. We achieved this by porting the OpenCL parts of the program to CUDA and optimizing the performance. This involved careful analysis and multiple changes to the algorithm. We also ported the code to NVIDIA OptiX to handle the collision detection. T...

The broad-scale reductions and shifts that may be expected under climate change in the availabili... more The broad-scale reductions and shifts that may be expected under climate change in the availability and quality of stopover habitat for long-distance migrants is an area of increasing concern for conservation biologists. Researchers generally have taken two broad approaches to the modeling of migration behaviour to understand the impact of these changes on migratory bird populations. These include models based on causal processes and their response to environmental stimulation, "mechanistic models", or models that primarily are based on observed animal distribution patterns and the correlation of these patterns with environmental variables, i.e. "data driven" models. Investigators have applied the latter technique to forecast changes in migration patterns with changes in the environment, for example, as might be expected under climate change, by forecasting how the underlying environmental data layers upon which the relationships are built will change over time. ...

A 4D version of BNL's spin tracking code SPINK [1] with limited elements has been successfull... more A 4D version of BNL's spin tracking code SPINK [1] with limited elements has been successfully ported to a C++/GPU platform using GPULib [2]. This prototype used only quadrupoles, simple snakes, dipoles and drifts. We present the approach used to track spin and orbit degrees of freedom of polarized proton beams simultaneously. We also present recent results of prototyping a general-purpose particle tracking on GPUs, discussing our CUDA im-plementation of maps for single-particle dynamics in the Argonne National Lab's accelerator code ELEGANT [3] where a 40x speedup was achieved for single-particle-dynamics elements.

Recent studies demonstrated high potential of the dielectric barrier discharge (DBD) plasmas for ... more Recent studies demonstrated high potential of the dielectric barrier discharge (DBD) plasmas for medical applications, such as sterilization or tissue regeneration. Despite intensive experimental studies have been conducted, the mechanism of plasma-tissue interaction still remains unclear. One of the open questions for the plasma-medical applications is the mechanism of filamentary structures formation in plasma and their interaction with tissues. Since formation of filaments is a purely kinetic effect, this issue needs to be addressed using kinetic, Particle-In-Cell simulation approach. We will present results of such numerical study. We performed 2D simulations of multiple streamers generation in atmospheric air using Tech-X's 2D/3D hybrid simulation tool VORPAL. We will demonstrate the resolution of the filamentary structure and will report the plasma properties. We will also address the plasma-induced effects on the tissue.

accelconf.web.cern.ch

The proposed RHIC-II luminosity upgrade includes a novel electron cooling section, using ∼55 MeV ... more The proposed RHIC-II luminosity upgrade includes a novel electron cooling section, using ∼55 MeV electrons to cool fully-ionized gold. We present simulations of the dynam-ical friction force exerted on the Au ions. Rather than a strong solenoid, a long helical wiggler magnet ...

Bulletin of the American Physical Society, Nov 15, 2007

Bulletin of the American Physical Society, Oct 24, 2005

Modeling complex structures and boundaries on a Cartesian grid is a challenge for many Finite Dif... more Modeling complex structures and boundaries on a Cartesian grid is a challenge for many Finite Difference Time Domain (FDTD) electromagnetic PIC codes. The simulation of a variety of devices such as accelerating cavities, plasma processing chambers, and antennas at the edge of tokamaks require conformal (curve fitting) boundaries. Since these devices are fundamentally three dimensional, the capability to run in parallel on large numbers of processors is needed. We have recently added conformal boundaries using the method of Zagorodnov to the plasma simulation code VORPAL. Our boundary approximation can be viewed with a 3D VRML viewer. Also these complex devices often include open boundaries. VORPAL includes Perfectly Matched Layer (PML) boundaries which efficiently absorb outgoing waves of any frequency and angle of incidence. VORPAL's FDTD algorithms scale to thousands of processors allowing for large 3D simulations. Simulations of complex structures using VORPAL will be presented.

The 33rd IEEE International Conference on Plasma Science, 2006. ICOPS 2006. IEEE Conference Record - Abstracts., 2006

We present a simple modification of the semi-classical over-barrier model of ion-atom charge exch... more We present a simple modification of the semi-classical over-barrier model of ion-atom charge exchange for the case of electron transfer between positive ions. The charge exchange cross-sections have been incorporated into an electromagnetic particle-in-cell plasma model which also includes schemes for electron impact ionization and electron-ion recombination. The influence of charge exchange on the plasma ionization state is explored for selected light elements under various temperature and density regimes.

abstract-Particle plasma models, including the particle-in-cell (PIC) model class, generally foll... more abstract-Particle plasma models, including the particle-in-cell (PIC) model class, generally follow the trajectories of a large number of macro-particles under the influence of both external and internal electromagnetic fields. The charge and current density terms in the Maxwell equations are obtained by prescribing a shape function that is centered on each particle. The PIC methods are among the most efficient for plasmas out of local thermodynamic equilibrium, but, through the use of grid interpolations, suffer from numerical grid heating (or cooling). In this poster we describe a spectral particle method that conserves both energy and momentum, and is thus devoid of grid heating, but at the computational cost of scaling as the product of the number of particles and the number of spectral grid points, or wave vectors. The algorithm is derived from a discrete action principle which ensures that it is symplectic, or that it conserves phase space volume. We present some simple numeri...

High intensity, high charge-state beams for a broad variety of ions are a requirement for next-ge... more High intensity, high charge-state beams for a broad variety of ions are a requirement for next-generation heavy-ion beam accelerators. As the intensities produced by current Electron Cyclotron Resonance (ECR) sources insufficient for many ions, the ion beam production has to be optimized. Efficient loading of the neutrals into the ECR plasma is one of the key elements for optimizing the

2008 Seventh International Conference on Machine Learning and Applications, 2008

Laser wakefield particle accelerators have shown the potential to generate electric fields thousa... more Laser wakefield particle accelerators have shown the potential to generate electric fields thousands of times higher than those of conventional accelerators. The resulting extremely short particle acceleration distance could yield a potential new compact source of energetic electrons and radiation, with wide applications from medicine to physics. Physicists investigate laser-plasma internal dynamics by running particle-in-cell simulations; however, this generates a large dataset that requires time-consuming, manual inspection by experts in order to detect key features such as beam formation. This paper describes a framework to automate the data analysis and classification of simulation data. First, we propose a new method to identify locations with high density of particles in the space-time domain, based on maximum extremum point detection on the particle distribution. We analyze high density electron regions using a lifetime diagram by organizing and pruning the maximum extrema as nodes in a minimum spanning tree. Second, we partition the multivariate data using fuzzy clustering to detect time steps in a experiment that may contain a high quality electron beam. Finally, we combine results from fuzzy clustering and bunch lifetime analysis to estimate spatially confined beams. We demonstrate our algorithms successfully on four different simulation datasets.

2008 SC - International Conference for High Performance Computing, Networking, Storage and Analysis, 2008

One of the central challenges in modern science is the need to quickly derive knowledge and under... more One of the central challenges in modern science is the need to quickly derive knowledge and understanding from large, complex collections of data. We present a new approach that deals with this challenge by combining and extending techniques from high performance visual data analysis and scientific data management. This approach is demonstrated within the context of gaining insight from complex, time-varying datasets produced by a laser wakefield accelerator simulation. Our approach leverages histogram-based parallel coordinates for both visual information display as well as a vehicle for guiding a data mining operation. Data extraction and subsetting are implemented with state-of-the-art index/query technology. This approach, while applied here to accelerator science, is generally applicable to a broad set of science applications, and is implemented in a production-quality visual data analysis infrastructure. We conduct a detailed performance analysis and demonstrate good scalability on a distributed memory Cray XT4 system.

Physical Review Special Topics - Accelerators and Beams, 2006

Capillary channels of 3 cm in length and with plasma densities 10 18 cm ÿ3 are a promising altern... more Capillary channels of 3 cm in length and with plasma densities 10 18 cm ÿ3 are a promising alternative to the much shorter, higher-density gas jets for GeV-scale laser wakefield acceleration of electrons. However, the large discrepancy between length scales of the plasma and the laser presents a major computational challenge for particle-in-cell (PIC) simulations. Methods are therefore sought that relax the need to concurrently resolve both length scales. For example, the commonly used ''moving window'' algorithm enables a reduction of the computational domain to a few plasma wavelengths, which is orders of magnitude smaller than the full length of the laser-plasma interaction. In addition, ponderomotive guiding center methods enable relaxation of the constraint to resolve the laser wavelength. These averaging methods split the laser-induced current into a rapidly varying part and a slowly varying envelope. The average over fast time scales is performed in a semianalytic way, leaving the evolution of the laser envelope and the plasma response to be modeled numerically. Here, we present a ponderomotive guiding center algorithm and demonstrate its applicability to model capillary channels by comparing it with fully kinetic PIC simulations.

Journal of Physics: Conference Series, 2007

Accelerators are the largest and most costly scientific instruments of the Department of Energy, ... more Accelerators are the largest and most costly scientific instruments of the Department of Energy, with uses across a broad range of science, including colliders for particle physics and nuclear science and light sources and neutron sources for materials studies. COMPASS, the Community Petascale Project for Accelerator Science and Simulation, is a broad, four-office (HEP, NP, BES, ASCR) effort to develop computational tools for the prediction and performance enhancement of accelerators. The tools being developed can be used to predict the dynamics of beams in the presence of optical elements and space charge forces, the calculation of electromagnetic modes and wake fields of cavities, the cooling induced by comoving beams, and the acceleration of beams by intense fields in plasmas generated by beams or lasers. In SciDAC-1, the computational tools had multiple successes in predicting the dynamics of beams and beam generation. In SciDAC-2 these tools will be petascale enabled to allow the inclusion of an unprecedented level of physics for detailed prediction. use of the large computing facilities extant and future to the accelerator physics community of the Department of Energy. Accelerators are the largest experimental instruments supported by the Department of Energy, with costs upwards of $1B for construction, and they are fundamental to the research carried out by the Offices of High Energy Physics (HEP), Nuclear Physics (NP), and Basic Energy Sciences. Hence, COMPASS is a multi-office project to support the needs of all three of those offices. The main thrust areas of this project are: • Beam Dynamics (BD): The evolution of beams through beam optics systems, including self forces and other forces of interaction.

Computer Physics Communications, 2004

Here we report on the addition of a parallel electrostatic (ES) field solver to the code VORPAL. ... more Here we report on the addition of a parallel electrostatic (ES) field solver to the code VORPAL. Developed at the University of Colorado at Boulder and at Tech-X Corporation, VORPAL is a parallel, arbitrary-dimensional hybrid PIC/fluid code. At present, VORPAL has only electromagnetic (EM) field solvers. It has been used extensively for laser-plasma interactions and is now being applied to electron-cooling and other wide-ranging problems.

We present a set of atomic process models for inclusion in electron-Uranium ion plasma simulation... more We present a set of atomic process models for inclusion in electron-Uranium ion plasma simulations. These include an adaptation of the binary encounter dipole model of Kim and Rudd (1994) for electron impact ionization, the Burgess general formula for dielectronic recombination rates, and the semiclassical impact parameter approximation for ion-atom charge exchange. The orbital binding energies and oscillator strengths are estimated with a multi-configuration Dirac Fock Model. Comparisons are made to existing experimental measurements of electron impact ionization for neutral Uranium (Halle et al., 1981) and for U(10+), U(13+) and U(16+) (Gregory et al., 1990). These models have been developed for use in simulations of U(28+) through U(35+) production at the the Versatile Electron-Cyclotron-Resonance Ion Source for Nuclear Science (VENUS) at Lawrence Berkeley Laboratory.

Aps Meeting Abstracts, Oct 1, 2005

Florida-Basic research in nuclear physics and astrophysics requires the acceleration and collisio... more Florida-Basic research in nuclear physics and astrophysics requires the acceleration and collision of highly charged rare isotopes. Conventional ion source technology based on electron cyclotron resonance (ECR) cannot produce the necessary beam current for many of these rare isotopes. Laser ionization of micron-sized doped water droplets generates plasmas which can produce accelerated ions. The droplet and laser parameters can be tuned to ionize a precisely controlled number of heavy atoms. Laser-droplet plasmas therefore seem an attractive enhancement for the heavy ion current in ECR sources. Here we present results of particle-in-cell (PIC) simulations of the interaction of a strong laser pulse with µm-sized droplets, using the plasma simulation code VORPAL[1]. The code features perfectly matched layer boundary condition which allows to avoid spurious reflections off the simulation domain walls. Droplet ionization is modeled using the quasi-static ADK tunneling ionization model implemented in the ionpack library[2]. The simulation parameters are chosen close to the experimental parameters.

Proceedings of the 2005 Particle Accelerator Conference

Capillary channels of cm-length and at plasma density low compared to gas jets are promising setu... more Capillary channels of cm-length and at plasma density low compared to gas jets are promising setups for low noise laser wakefield acceleration. Computationally, however, the large discrepancy of the length scales of the plasma and the laser are a big challenge. Methods are therefore sought that relax the need to concurrently resolve both length scales. Average methods, which split the electromagnetic field into a fast and a slowly varying part, allow to relax the constraint to resolve the laser wavelength. Such an envelope model is currently being incorporated into the VORPAL plasma simulation code. Simulation results for benchmark cases and for laser pulse propagation in a Ñscale channel are presented.