Increasingly in industry, software design and implementation is object-oriented, developed in C++ or Java, and relies heavily on pre-existing software libraries (e.g. the Microsoft Foundation Classes for C++, the Java API for Java). A similar but more tentative trend is developing in high-performance parallel scientific computing. The transition from serial to parallel application development considerably increases the need for library support: task creation and management, data distribution and dynamic redistribution, and inter-process and inter-processor communication and synchronization must be supported. PADRE is a library to support the interoperability of parallel applications. We feel there is significant need for just such a tool to compliment the many domain-specific application frameworks presently available today, but which are generally not interoperable.

Modern antennas especially arrays are being placed in layers of materials on complex environments. This technique produces aesthetically pleasing structures if necessary, allows for more freedom in structure planning, and can improve antenna performance. In the past, buried antennas have been studied by numerous authors such as in Reference. Recent work on this subject uses spectral and/or numerical moment method formulations. For high frequency analysis it is important to find efficient and accurate methods for design purposes. A rigorous recursive method for plane waves reflection and transmission coefficients by Richmond has been used in the past for dipoles above multilayer slabs. This solution is modified in this paper to account for forward and backward traveling rays with appropriate spread factors for a dipole in the media. Extensive validation for this approximate method shows good agreement with a Method of Moments code. This code is developed at Lawrence Livermore National Laboratory. The geometry for these comparisons uses a dipole in nontruncated dielectric multilayer slabs.

An initial design of the optics required for producing gamma-gamma collisions was produced for the NLC Zeroth Order Design Report (ZDR) submitted to the 1996 Snowmass workshop. The design incorporated only loose constraints from the interaction region requirements. In this paper we report progress on a design of a gamma-gamma interaction region which incorporates all constraints.

The effect of multiple and delayed jet impact and penetration on the borehole diameter in concrete targets is discussed in this paper. A first-order principle of shaped-charge jet penetration is that target hole volume is proportional to the energy deposited in the target by the jet. This principle is the basis for the relation that target borehole diameter at any depth along the penetration path is proportional to the jet energy deposited in the target at that location. Our current research shows that the 'jet energy per unit hole volume constant' for concrete can be substantially altered by the use of multiple and delayed jet impacts. It has been shown that enhanced entrance crater formation results from the simultaneous impact and penetration of three shaped-charge jets. We now demonstrate that enhanced borehole diameter is also observed by the simultaneous impact and penetration of multiple shaped-charge jets followed by the delayed impact and penetration of a single shaped-charge jet.

The performance of commercial high-performance fibers is examined for application to flywheel power supplies. It is shown that actual delivered performance depends on multiple factors such as inherent fiber strength, strength translation and stress-rupture lifetime. Experimental results for recent stress-rupture studies of carbon fibers will be presented and compared with other candidate reinforcement materials. Based on an evaluation of all of the performance factors, it is concluded that carbon fibers are preferred for highest performance and E-glass fibers for lowest cost. The inferior performance of the low-cost E-glass fibers can be improved to some extent by retarding the stress-corrosion of the material due to moisture and practical approaches to mitigating this corrosion are discussed. Many flywheel designs are limited not by fiber failure, but by matrix-dominated failure modes. Unfortunately, very few experimental results for stress-rupture under transverse tensile loading are available. As a consequence, significant efforts are made in flywheel design to avoid generating any transverse tensile stresses. Recent results for stress-rupture of a carbon fiber/epoxy composite under transverse tensile load reveal that these materials are surprisingly durable under the transverse loading condition and that some radial tensile stress could be tolerated in flywheel applications.

An Achilles heel for the performance of thick-section, cylindrical fiber composite flywheels is the poor interlaminar properties of the material. Methods that have been used to minimize or eliminate radial tensile stresses include prestressing concentric cylinders and mass loading. There can also be significant interlaminar shear stresses at the edges of mass-loaded flywheels and in flywheels for high-power density applications where abrupt braking results in high torque levels. To specify adequate safety factors for thick-section flywheels used in these applications, the failure envelope and fatigue behavior under combined interlaminar stresses are required. Using a hollow cylindrical specimen, which was subjected to combined axial compression and torsion, results for fatigue and failure were generated for several flywheel material systems. Interlaminar compression resulted in significant enhancements to the interlaminar shear strength and results were compared to the predictions of proposed three-dimensional composite failure models. The interlaminar shear fatigue behavior of a carbodepoxy system was also studied and compression was found to greatly enhance fatigue life. The results demonstrate that radial compression stresses can yield improvements in the interlaminar shear strength and fatigue lifetimes of composite flywheel rotors.

In this paper, we present an efficient solution approach for fully implicit, large-scale, nonlinear radiation diffusion problems. The fully implicit approach is compared to a semi-implicit solution method. Accuracy and efficiency are shown to be better for the fully implicit method on both one- and three-dimensional problems with tabular opacities taken from the LEOS opacity library.

An alternative approach is described to evaluate the statistical nature of the breakup of shaped charge liners. Experimental data from ductile and brittle copper jets are analyzed in terms of velocity gradient, deviation of {Delta}V from linearity, R/S analysis, and the Hurst exponent within the coupled map lattice model. One-dimensional simulations containing 600 zones of equal mass and using distinctly different force-displacement curves are generated to simulate ductile and brittle behavior. A particle separates from the stretching jet when an element of material reaches the failure criterion. A simple model of a stretching rod using brittle, semi-brittle, and ductile force-displacement curves is in agreement with the experimental results for the Hurst exponent and the phase portraits and indicates that breakup is a correlated phenomenon.

Data obtained through scientific observations are often contaminated by noise and artifacts from various sources. As a result, a first step in mining these data is to isolate the signal of interest by minimizing the effects of the contaminations. Once the data has been cleaned or de-noised, data mining can proceed as usual. In this paper, we describe our work in denoising astronomical images from the Faint Images of the Radio Sky at Twenty-Centimeters (FIRST) survey. We are mining this survey to detect radio-emitting galaxies with a bent-double morphology. This task is made difficult by the noise in the images caused by the processing of the sensor data. We compare three different approaches to de-noising: thresholding of wavelet coefficients advocated in the statistical community, traditional Altering methods used in the image processing community, and a simple thresholding scheme proposed by FIRST astronomers. While each approach has its merits and pitfalls, we found that for our purpose, the simple thresholding scheme worked relatively well for the FIRST dataset.

The 1.5 Tesla superconducting solenoid is part of the BABAR Detector located in the PEP-II B-Factory machine at the Stanford Linear Accelerator Center. The solenoid has a 2.8 m bore and is 3.7 m long. The two layer solenoid is wound with an aluminum stabilized conductor which is graded axially to produce a {+-} 3% field uniformity in the tracking region. The 24 month fabrication, 3 month installation and 1 month commissioning of the solenoid were completed on time and budget. This paper summarizes the culmination of a 3 year design, fabrication and testing program of the BABAR superconducting solenoid. The work was completed by an international collaboration between Ansaldo, INFN, LLNL, and SLAC. Critical current measurements of the superconducting strand, cable and conductor, cool-down, operation with the thermo-siphon cooling, fast and slow discharges, and magnetic forces are discussed in detail.

The authors discuss the decay of axion walls bounded by strings and present numerical simulations of the decay process. In these simulations, the decay happens immediately, in a time scale of order the light travel time, and the average energy of the radiated axions is <w{sub a}> {approx_equal} 7m{sub a} for v{sub a}/m{sub a} {approx_equal} 500. <w{sub a}> is found to increase approximately linearly with ln(v{sub a}/m{sub a}). Extrapolation of this behavior yields <w{sub a}> {approx_equal} 60 m{sub a} in axion models of interest.

Computer-aided design and manufacture (CAD/CAM) have enabled advances in the design and manufacture of many accelerator components, though government procurement rules tend to inhibit its use. We developed and executed a method that provides adequate documentation for the procurement process, industrial vendor manufacturing processes, and laboratory installation activities. We detail our experiences in the design and manufacture of 60 separate and unique PEP-II Low Energy Ring Interaction Region vacuum chambers totaling {approx} 140m in length as an example of how we used this technique, reducing design effort and manufacturing risk while streamlining the production process. We provide ''lessons learned'' to better implement and execute the process in subsequent iterations. We present our study to determine the estimated savings in the design and production of the Spallation Neutron Source room temperature linac if this process were utilized.

The cost/performance ratio of networks of workstations has been constantly improving. This trend is expected to continue in the near future. The aggregate peak rate of such systems often matches or exceeds the peak rate offered by the fastest parallel computers. This has motivated research towards using a network of computers, interconnected via a fast network (cluster system) or a simple Local Area Network (LAN) (distributed system), for high performance concurrent computations. Some of the important research issues arise such as (1) Optimal problem partitioning and virtual interconnection topology mapping; (2) Optimal execution scheduling and load balancing. CFD codes have been efficiently implemented on homogeneous parallel systems in the past. In particular, the helicopter aerodynamics CFD code TURNS has been implemented with MPI on the IBM SP with parallel relaxation and Krylov iterative methods used in place of more traditional recursive algorithms to enhance performance. In this implementation the space domain is divided into equal subdomain which are mapped to the processors. We consider the implementation of TURNS on a LAN of heterogeneous workstations. In order to deal with the problem of load balancing due to the different processor speeds we propose a suboptimal algorithm of dividing the space domain …

The advent of CCD cameras and computerized data recording has spurred the development of several new cameras and techniques for recording nanosecond images. We have made a side by side comparison of three nanosecond frame cameras, examining them for both performance and operational characteristics. The cameras include; Micro-Channel Plate/CCD, Image Diode/CCD and Image Diode/Film; combinations of gating/data recording. The advantages and disadvantages of each device will be discussed.

Plastic deformation in bcc metals at low temperatures and high-strain rates is controlled by the motion of a/2<111> screw dislocations, and understanding the fundamental atomistic processes of this motion is essential to develop predictive multiscale models of crystal plasticity. The multiscale modeling approach presented here for bcc Ta is based on information passing, where results of simulations at the atomic scale are used in simulations of plastic deformation at mesoscopic length scales via dislocation dynamics (DD). The relevant core properties of a/2<111> screw dislocations in Ta have been obtained using quantum-based interatomic potentials derived from model generalized pseudopotential theory and an ab-initio data base together with an accurate Green's-function simulation method that implements flexible boundary conditions. In particular, the stress-dependent activation enthalpy for the lowest-energy kink-pair mechanism has been calculated and fitted to a revealing analytic form. This is the critical quantity determining dislocation mobility in the DD simulations, and the present activation enthalpy is found to be in good agreement with the previous empirical form used to explain the temperature dependence of the yield stress.

Adaptive mesh refinement is complicated by both the algorithms and the dynamic nature of the computations. In parallel the complexity of getting good performance is dependent upon the architecture and the application. Most attempts to address the complexity of AMR have lead to the development of library solutions, most have developed object-oriented libraries or frameworks. All attempts to date have made numerous and sometimes conflicting assumptions which make the evaluation of performance of AMR across different applications and architectures difficult or impracticable. The evaluation of different approaches can alternatively be accomplished through simulation of the different AMR processes. In this paper we outline our research work to simulate the processing of adaptive mesh refinement grids using a distributed array class library (P++). This paper presents a combined analytic and empirical approach, since details of the algorithms can be readily predicted (separated into specific phases), while the performance associated with the dynamic behavior must be studied empirically. The result, AMRSim, provides a simple way to develop bounds on the expected performance of AMR calculations subject to constraints given by the algorithms, frameworks, and architecture.

The mechanisms of dislocation-defect interactions are of practical importance for developing quantitative structure-property relationships, mechanistic understanding of plastic flow localization and predictive models of mechanical behavior in metals under irradiation. In copper and other face centered cubic metals, high-energy particle irradiation produces hardening and shear localization. Post-irradiation microstructural examination in Cu reveals that irradiation has produced a high number density of nanometer sized stacking fault tetrahedra. Thus, the resultant irradiation hardening and shear localization is commonly attributed to the interaction between stacking fault tetrahedra and mobile dislocations, although the mechanism of this interaction is unknown. In this work, we present a comprehensive molecular dynamics simulation study that characterizes the interaction and fate of moving dislocations with stacking fault tetrahedra in Cu using an EAM interatomic potential. This work is intended to produce atomistic input into dislocation dynamics simulations of plastic flow localization in irradiated materials.

None

The Next Linear Collider (NLC) offers an opportunity to use high energy photon collisions to probe basic particle structures. The production of high energy gammas from collisions between the NLC high energy electrons and low energy photons places difficult requirements on the laser low energy photon source [1]. The photon wavelength must be approximately one micron since longer wavelengths will decrease the electron utilization efficiency, and shorter wavelengths will open a loss channel for gammas through electron pair creation. The laser pulse format must match the electron generation format of the NLC. The electrons are produced in macropulses at 120 Hz. Each macro-pulse consists of around 100 subpulses separated by 2.8 nanoseconds. To interact efficiently with the electrons the laser subpulses must have approximately a 2 picosecond pulse duration. Analysis of the photon densities required for efficient utilization of the electrons and the focusing capabilities of the final photon injection optics leads to a required photon sub-pulse energy of approximately one joule. Thus the laser macro-pulse energy must be 100 joules at 120 Hz. The laser average power will be 12 kW.

This paper investigates whether it is better to use a certain constant amount of computational resources in a single run with a large population, or in multiple runs with smaller populations. The paper presents the primary tradeoffs involved in this problem and identifies the conditions under which there is an advantage to use multiple small runs. The paper uses an existing model that relates the quality of the solutions reached by a GA with its population size. The results suggest that in most cases a single run with the largest population possible reaches a better solution than multiple isolated runs. The findings are validated with experiments on functions of varying difficulty.

We describe a method of improving the appearance of a low vertex count mesh in a manner that is guided by rendered images of the original, detailed mesh. This approach is motivated by the fact that greedy simplification methods often yield meshes that are poorer than what can be represented with a given number of vertices. Our approach relies on edge swaps and vertex teleports to alter the mesh connectivity, and uses the downhill simplex method to simultaneously improve vertex positions and surface attributes. Note that this is not a simplification method--the vertex count remains the same throughout the optimization. At all stages of the optimization the changes are guided by a metric that measures the differences between rendered versions of the original model and the low vertex count mesh. This method creates meshes that are geometrically faithful to the original model. Moreover, the method takes into account more subtle aspects of a model such as surface shading or whether cracks are visible between two interpenetrating parts of the model.

The lifetime of optics in high-fluence UV laser applications is typically limited by the initiation of damage and its subsequent growth. We have measured the growth rate of laser-induced damage on fused silica surfaces in both air and vacuum. The data shows exponential growth in the lateral size of the damage site with shot number above a threshold fluence. The concurrent growth in depth follows a linear dependence with shot number. The size of the initial damage influences the threshold for growth; the morphology of the initial site depends strongly on the initiating fluence. We have found only a weak dependence on pulse length for growth rate. Most of the work has been on bare substrates but the presence of a sol-gel AR coating has no significant effect.

Libraries arise naturally from the increasing complexity of developing scientific applications, the optimization of libraries is just one type of high-performance optimization. Many complex applications areas can today be addressed by domain-specific object-oriented frameworks. Such object-oriented frameworks provide an effective compliment to an object-oriented language and effectively permit the design of what amount to essentially domain-specific languages. The optimization of such a domain-specific library/language combination however is particularly complicated due to the inability of the compiler to optimize the use of the libraries abstractions. The recognition of the use of object-oriented abstractions within user applications is a particularly difficult but important step in the optimization of how objects are used within expressions and statements. Such recognition entails more than just complex pattern matching. The approach presented within this paper uses specially built grammars to parse the C++ representation. The C++ representation is itself obtained using a modified version of the SAGE II C/C++ source code restructuring tool which is inturn based upon the Edison Design Group (EDG) C++ front-end. ROSETTA is a tool which automatically builds grammars and parsers from class definitions, associated parsers parse abstract syntax trees (ASTs) of lower level grammars into ASTs of higher level grammars. The …

We provide direct experimental evidence for a non-random distribution of atomic constituents in Zr-based multi-component bulk metallic glasses using positron annihilation spectroscopy. The Ti content around the open-volume regions is significantly enhanced at the expense of Cu and Ni, indicating that Cu and Ni occupy most of the volume bounded by their neighboring atoms while Ti and Zr are less closely packed and more likely to be associated with open-volume regions. Temperature-dependent measurements indicate the presence of at least two different characteristic sizes for the open volume regions. Measurements on hydrogen-charged samples show that the larger open-volume regions can be filled by hydrogen up to a critical density. Beyond this critical density, local atomic-scale open-volume damage is created in the sample to accommodate additional hydrogen. The onset of this local damage in positron annihilation data coincides with the onset of volume expansion in X-ray diffraction data.