Search NASA⌕ Search

SEARCH · Search NASA

Results for “Parallel optimization”

Search indexed NASA NTRS and DOE OSTI research on propulsion, heat transfer, battery materials and energy systems. Follow report and document links to the original sources.

Quote a phrase for an exact phrase match. Source license links do not imply unrestricted reuse.

At least 1,315 records · Page 73

Controlled Dielectric Breakdown to Form Pinhole Passivating Contacts

This contribution explores an alternate route to forming pinhole-based poly-Si/dielectric/c-Si passivating contacts. The method utilizes controlled dielectric breakdown or electroforming to produce nanoscale pinholes in a thick (non-tunnelling) dielectric which, when annealed, allows dopant atoms to pass from doped poly-Si through the pinholes and into the c-Si wafer, forming conductive pathways. We show that the pinholes lose passivation after electroforming but can be repassivated with a forming gas anneal. N-type contacts show contact resistivities of ~20 mOhm-cm2, but p-type contacts are ~100 mOhm-cm2. Devices show a distinct kink in the J-V curve indicative of a barrier to transport. The method can be expanded to optimized dielectric passivation stacks (not just thin, single layers) and can be formed in parallel over the faces of the wafer in selected areas (pinholes only under the grid lines).

CONDENSED MATTER PHYSICS, SUPERCONDUCTIVITY AND SU↗

SAIGE-GPU: accelerating genome- and phenome-wide association studies using GPUs

Genome-wide association studies (GWAS) at biobank scale are computationally intensive, especially for admixed populations requiring robust statistical models. SAIGE is a widely used method for generalized linear mixed-model GWAS but is limited by its CPU-based implementation, making phenome-wide association studies impractical for many research groups. We developed SAIGE-GPU, a GPU-accelerated version of SAIGE that replaces CPU-intensive matrix operations with GPU-optimized kernels. The core innovation is distributing genetic relationship matrix calculations across GPUs and communication layers. Applied to 2068 phenotypes from 635 969 participants in the Million Veteran Program, including diverse and admixed populations, SAIGE-GPU achieved a 5-fold speedup in mixed model fitting on supercomputing infrastructure and cloud platforms. We further optimized the variant association testing step through multi-core and multi-trait parallelization. Deployed on Google Cloud Platform and Azure, the method provided substantial cost and time savings. Source code and binaries are available for download at https://github.com/saigegit/SAIGE/tree/SAIGE-GPU-1.3.3. A code snapshot is archived at Zenodo for reproducibility (DOI: [10.5281/zenodo.17642591]). SAIGE-GPU is available in a containerized format for use across HPC and cloud environments and is implemented in R/C++ and runs on Linux systems.

Rodriguez, Alex [Argonne National Laboratory (ANL)↗

New Effective Multithreaded Matching Algorithms

Matching is an important combinatorial problem with a number of applications in areas such as community detection, sparse linear algebra, and network alignment. Since computing optimal matchings can be very time consuming, several fast approximation algorithms, both sequential and parallel, have been suggested. Common to the algorithms giving the best solutions is that they tend to be sequential by nature, while algorithms more suitable for parallel computation give solutions of less quality. We present a new simple 1 2 -approximation algorithm for the weighted matching problem. This algorithm is both faster than any other suggested sequential 1 2 -approximation algorithm on almost all inputs and also scales better than previous multithreaded algorithms. We further extend this to a general scalable multithreaded algorithm that computes matchings of weight comparable with the best sequential algorithms. The performance of the suggested algorithms is documented through extensive experiments on different multithreaded architectures.

matching, parallel algorithms↗

TU-EF-207-01: Introductory Remarks on Recent Advances in Breast Imaging

Breast imaging technology is advancing on several fronts. In digital mammography, the major technological trend has been on optimization of approaches for performing combined mammography and tomosynthesis using the same system. In parallel, photon-counting slot-scan mammography is now in clinical use and more efforts are directed towards further development of this approach for spectral imaging. Spectral imaging refers to simultaneous acquisition of two or more energy-windowed images. Depending on the detector and associated electronics, there are a number of ways this can be accomplished. Spectral mammography using photon-counting detectors can suppress electronic noise and importantly, it enables decomposition of the image into various material compositions of interest facilitating quantitative imaging. Spectral imaging can be particularly important in intravenously injected contrast mammography and eventually tomosynthesis. The various approaches and applications of spectral mammography are discussed. Digital breast tomosynthesis relies on the mechanical movement of the x-ray tube to acquire a number of projections in a predefined arc, typically from 9 to 25 projections over a scan angle of +/−7.5 to 25 degrees depending on the particular system. The mechanical x-ray tube motion requires relatively long acquisition time, typically between 3.7 to 25 seconds depending on the system. Moreover, mechanical scanning may have an effect on the spatial resolution due to internal x-ray filament or external mechanical vibrations. New x-ray source arrays have been developed and they are aimed at replacing the scanned x-ray tube for improved acquisition time and potentially for higher spatial resolution. The potential advantages and challenges of this approach are described. Combination of digital mammography and tomosynthesis in a single system places increased demands on certain functional aspects of the detector and overall performance, particularly in the tomosynthesis mode due to lower photon fluence per projection. This may require fast-frame acquisition and symmetric or asymmetric pixel binning in some systems. Recent studies investigated the performance of increased conversion layer thickness for contrast-enhanced imaging of the breast in dual-energy acquisition mode. In other direct conversion detectors operating in the avalanche mode, sensitivities close to the single photon response are also explored for mammography and breast tomosynthesis. The potential advantages and challenges of this approach are described. Dedicated breast CT brings x-ray imaging of the breast to true tomographic 3D imaging. It can eliminate the tissue superposition problem and does not require physical compression of the breast. Using cone beam geometry and a flat-panel detector, several hundred projections are acquired and reconstructed to near isotropic voxels. Multiplanar reconstruction facilitates viewing the breast volume in any desired orientation. Ongoing clinical studies, the current state-of-the art, and research to advance the technology are described. Learning Objectives: To understand the ongoing developments in x-ray imaging of the breast To understand the approaches and applications of spectral mammography To understand the potential advantages of distributed x-ray source arrays for digital breast tomosynthesis To understand the ongoing developments in detector technology for digital mammography and breast tomosynthesis To understand the current state-of-the-art for dedicated cone-beam breast CT and research to advance the technology. Research collaboration with Koning Corporation.

60 APPLIED LIFE SCIENCES↗

Parallel Element Agglomeration Algebraic Multigrid and Upscaling Library

ParFELAG is a parallel distributed memory C++ library for numerical upscaling of finite element discretizations. It provides optimal complesity algorithms ro build multilevel hierarchies and solvers that can be used for solving a wide class of partial differential equations (elliptic, hyperbolic, saddle point problems) on general unstructured mesh (under the assumption that the topology of the agglomerated entities is correct). Additionally, a novel multilevel solver for saddle point problems with divergence constraint is implemented.

Source record↗

ML-PSA

The computer code uses a parallel simulated annealing framework with embedded machine learning components to solve multi-constrained optimization problems. The software automatically balances the execution of low and high fidelity physics models within the optimization procedure. The low fidelity model is used to rapidly explore the design space while the high fidelity physics model is executed sparingly to account for complex design constraints that are not resolved by the quickly executing low fidelity model.

Gurecky, William↗

Robust Event Simulation Variants Endowed to Ns-3 for General Exploration

Sandia's ns-3 contributions are modifications to the open source ns-3 network simulator that enable simulation speedup. One improvement, for example, removes the need for simulating the transmission and receipt of packets between nodes that are too distant to actually be able to communicate. In large-scale simulations this optimization has shown significant gains in performance. Additional contributions will pave the way for parallel discrete simulation (PDES) in ns-3. SAND2020-12456 O Sandia National Laboratories is a multimission laboratory managed and operated by National Technology & Engineering Solutions of Sandia, LLC, a wholly owned subsidiary of Honeywell International Inc., for the U.S. Department of Energy’s National Nuclear Security Administration under contract DE-NA0003525.

Dickson, Joseph↗

Expanded Capabilities and Opportunities for Virtual Beamline Code

In the 25 years since LLNL broke ground on the National Ignition Facility (NIF) - a cornerstone of the National Nuclear Security Administration’s Stockpile Stewardship Program - the NIF and Photon Science (NIF&PS) Directorate has steadily pushed the boundaries of laser physics, nonlinear optics, and photonics in service of inertial confinement fusion experiments, and advanced photon sources development. Beyond the 192 beamlines of the primary laser facility with its world-record energy and fusion yields, NIF&PS capabilities have expanded to include chirped-pulse amplification, kilojoule petawatt-class short-pulse systems to generate hard x-ray radiographic probes, high-average power lasers as secondary sources for generation or wake field particle acceleration along directed energy applications, and other scientific pursuits. The Computing Directorate partners with NIF&PS on numerous projects and technologies from diagnostic measurements and high performance control systems to data analysis, information technology infrastructure, and scientific simulation code development. “This strong collaboration enables cutting-edge science, particularly with regard to physics-based modeling and simulation,” says physicist Jean-Michel Di Nicola. One crucial capability for NIF&PS is the Virtual Beamline (VBL) laser simulation code. It can model all the major laser physics and technology involved in the design optimization, commissioning, and operations of advanced laser architectures from tabletop to NIF-scale. VBL provides researchers with high-fidelity models and high-resolution calculations of laser performance predictions—including for the entire NIF laser system, the Advanced Radiographic Capability, parts of the High Repetition-Rate Advanced Petawatt Laser System laser delivered to the Czech Republic, and the Optical Science Laser. After more than two decades of experimentally verified physics and computing enhancements, this unique workhorse code is wrapping up another major milestone: migration from Java to C++ with a flurry of user interface and optimization features, as well as laser physics and high-resolution enhancements thanks to parallel execution on Livermore Computing platforms.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗

Multivariable closed loop control analysis and synthesis for complex flight systems

A flight control system analysis and synthesis method is presented that is intended to be especially suitable for application to vehicles exhibiting complex dynamic characteristics. For such vehicles quantitative handling qualities specifications are not usually available. Howver, handling qualities objectives are specifically introduced in this method via the hypothesis of correlation between pilot ratings and the objective function of an optimal control model of the human pilot. Further, since augmentation and pilot operate in parallel, simultaneous determination of the augmentation and pilot model gains is required. Desirable augmented dynamics are obtained for a variety of complex systems and the method is experimentally verified in the case of simple pilot damper gain selection for optimum pitch tracking performance.

Schmidt, D. K.↗

Investigation of potential driver modules and transmission lines for a high frequency power system on the space station

The feasibility of using Series Resonant Inverter as the driver module for high frequency power system on the Space Station was assessed. The performance of the Series Resonant Inverter that was used in the testing of the single-phase, 2.0-kw resonant AC power system breadboard is summarized. The architecture is descirbed and the driver modules of the 5.0 kw AC power system breadboard are analyzed. An investigation of the various types of transmission lines is continued. Measurements of equivalent series resistor and inductor and equivalent parallel capacitors are presented. In particular, a simplified approach is utilized to describe the optimal transmission line.

Brush, Harold T.↗

Single-mode projection filters for modal parameter identification for flexible structures

Single-mode projection filters are developed for eigensystem parameter identification from both analytical results and test data. Explicit formulations of these projection filters are derived using the orthogonal matrices of the controllability and observability matrices in the general sense. A global minimum optimization algorithm is applied to update the filter parameters by using the interval analysis method. The updated modal parameters represent the characteristics of the test data. For illustration of this new approach, a numerical simulation for the MAST beam structure is shown by using a one-dimensional global optimization algorithm to identify modal frequencies and damping. The projection filters are practical for parallel processing implementation.

Huang, Jen-Kuang↗

Development and Demonstration of an Ada Test Generation System

In this project we have built a prototype system that performs Feasible Path Analysis on Ada programs: given a description of a set of control flow paths through a procedure, and a predicate at a program point feasible path analysis determines if there is input data which causes execution to flow down some path in the collection reaching the point so that tile predicate is true. Feasible path analysis can be applied to program testing, program slicing, array bounds checking, and other forms of anomaly checking. FPA is central to most applications of program analysis. But, because this problem is formally unsolvable, syntactic-based approximations are used in its place. For example, in dead-code analysis the problem is to determine if there are any input values which cause execution to reach a specified program point. Instead an approximation to this problem is computed: determine whether there is a control flow path from the start of the program to the point. This syntactic approximation is efficiently computable and conservative: if there is no such path the program point is clearly unreachable, but if there is such a path, the analysis is inconclusive, and the code is assumed to be live. Such conservative analysis too often yields unsatisfactory results because the approximation is too weak. As another example, consider data flow analysis. A du-pair is a pair of program points such that the first point is a definition of a variable and the second point a use and for which there exists a definition-free path from the definition to the use. The sharper, semantic definition of a du-pair requires that there be a feasible definition-free path from the definition to the use. A compiler using du-pairs for detecting dead variables may miss optimizations by not considering feasibility. Similarly, a program analyzer computing program slices to merge parallel versions may report conflicts where none exist. In the context of software testing, feasibility analysis plays an important role in identifying testing requirements which are infeasible. This is especially true for data flow testing and modified condition/decision coverage. Our system uses in an essential way symbolic analysis and theorem proving technology, and we believe this work represents one of the few successful uses of a theorem prover working in a completely automatic fashion to solve a problem of practical interest. We believe this work anticipates an important trend away from purely syntactic-based methods for program analysis to semantic methods based on symbolic processing and inference technology. Other results demonstrating the practical use of automatic inference is being reported in hardware verification, although there are significant differences between the hardware work and ours. However, what is common and important is that general purpose theorem provers are being integrated with more special-purpose decision procedures to solve problems in analysis and verification. We are pursuina commercial opportunities for this work, and will use and extend the work in other projects we are engaged in. Ultimately we would like to rework the system to analyze C, C++, or Java as a key step toward commercialization.

Source record↗

Compute as Fast as the Engineers Can Think! ULTRAFAST COMPUTING TEAM FINAL REPORT

This report documents findings and recommendations by the Ultrafast Computing Team (UCT). In the period 10-12/98, UCT reviewed design case scenarios for a supersonic transport and a reusable launch vehicle to derive computing requirements necessary for support of a design process with efficiency so radically improved that human thought rather than the computer paces the process. Assessment of the present computing capability against the above requirements indicated a need for further improvement in computing speed by several orders of magnitude to reduce time to solution from tens of hours to seconds in major applications. Evaluation of the trends in computer technology revealed a potential to attain the postulated improvement by further increases of single processor performance combined with massively parallel processing in a heterogeneous environment. However, utilization of massively parallel processing to its full capability will require redevelopment of the engineering analysis and optimization methods, including invention of new paradigms. To that end UCT recommends initiation of a new activity at LaRC called Computational Engineering for development of new methods and tools geared to the new computer architectures in disciplines, their coordination, and validation and benefit demonstration through applications.

Biedron, R. T.↗

On Improving Efficiency of Differential Evolution for Aerodynamic Shape Optimization Applications

Differential Evolution (DE) is a simple and robust evolutionary strategy that has been proven effective in determining the global optimum for several difficult optimization problems. Although DE offers several advantages over traditional optimization approaches, its use in applications such as aerodynamic shape optimization where the objective function evaluations are computationally expensive is limited by the large number of function evaluations often required. In this paper various approaches for improving the efficiency of DE are reviewed and discussed. These approaches are implemented in a DE-based aerodynamic shape optimization method that uses a Navier-Stokes solver for the objective function evaluations. Parallelization techniques on distributed computers are used to reduce turnaround times. Results are presented for the inverse design of a turbine airfoil. The efficiency improvements achieved by the different approaches are evaluated and compared.

Madavan, Nateri K.↗

Parallel Element Agglomeration Algebraic Multigrid and Upscaling Library

ParFELAG is a parallel distributed memory C++ library for numerical upscaling of finite element discretizations. It provides optimal complesity algorithms ro build multilevel hierarchies and solvers that can be used for solving a wide class of partial differential equations (elliptic, hyperbolic, saddle point problems) on general unstructured mesh (under the assumption that the topology of the agglomerated entities is correct). Additionally, a novel multilevel solver for saddle point problems with divergence constraint is implemented.

Source record↗

High Performance Interactive System Dynamics Visualization

This brochure describes a system dynamics simulation (SD) framework that supports an end-to-end analysis workflow that is optimized for deployment on ESIF facilities(Peregrine and the Insight Center). It includes (I) parallel and distributed simulation of SD models, (ii) real-time 3D visualization of running simulations, and (iii) comprehensive database-oriented persistence of simulation metadata, inputs, and outputs.

97 MATHEMATICS AND COMPUTING↗

High Performance Interactive System Dynamics Visualization

This presentation describes a system dynamics simulation (SD) framework that supports an end-to-end analysis workflow that is optimized for deployment on ESIF facilities(Peregrine and the Insight Center). It includes (I) parallel and distributed simulation of SD models, (ii) real-time 3D visualization of running simulations, and (iii) comprehensive database-oriented persistence of simulation metadata, inputs, and outputs.

97 MATHEMATICS AND COMPUTING↗