Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “Data Compression”

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 1,765 records · Page 98

Dynamic compression of dense oxide (Gd 3 Ga 5 O 12 ) from 0.4 to 2.6 TPa: Universal Hugoniot of fluid metals

Materials at high pressures and temperatures are of great current interest for warm dense matter physics, planetary sciences, and inertial fusion energy research. Shock-compression equation-of-state data and optical reflectivities of the fluid dense oxide, Gd 3 Ga 5 O 12 (GGG), were measured at extremely high pressures up to 2.6 TPa (26 Mbar) generated by high-power laser irradiation and magnetically-driven hypervelocity impacts. Above 0.75 TPa, the GGG Hugoniot data approach/reach a universal linear line of fluid metals, and the optical reflectivity most likely reaches a constant value indicating that GGG undergoes a crossover from fluid semiconductor to poor metal with minimum metallic conductivity (MMC). These results suggest that most fluid compounds, e.g., strong planetary oxides, reach a common state on the universal Hugoniot of fluid metals (UHFM) with MMC at sufficiently extreme pressures and temperatures. Lastly, the systematic behaviors of warm dense fluid would be useful benchmarks for developing theoretical equation-of-state and transport models in the warm dense matter regime in determining computational predictions.

70 PLASMA PHYSICS AND FUSION TECHNOLOGY↗

Transition of temporal scaling behavior in percolation assisted shear-branching structure during plastic deformation

This study explores the temporal scaling behavior induced shear-branching structure in response to variant temperatures and strain rates during plastic deformation of Zr-based bulk metallic glass (BMG). The data analysis based on the compression tests suggests that there are two states of shear-branching structures: the fractal structure with a long-range order at an intermediate temperature of 223 K and a larger strain rate of 2.5 × 10 –2 s –1 ; the disordered structure dominated at other temperature and strain rate. It can be deduced from the percolation theory that the compressive ductility, ec, can reach the maximum value at the intermediate temperature. Furthermore, a dynamical model involving temperature is given for depicting the shear-sliding process, reflecting the plastic deformation has fractal structure at the temperature of 223 K and strain rate of 2.5 × 10 –2 s –1 .

36 MATERIALS SCIENCE↗

Planck intermediate results: XXIX. All-sky dust modelling with Planck , IRAS, and WISE observations

In this paper, we present all-sky modelling of the high resolution Planck, IRAS, and WISE infrared (IR) observations using the physical dust model presented by Draine & Li in 2007 (DL, ApJ, 657, 810). We study the performance and results of this model, and discuss implications for future dust modelling. The present work extends the DL dust modelling carried out on nearby galaxies using Herschel and Spitzer data to Galactic dust emission. We employ the DL dust model to generate maps of the dust mass surface density Σ Md , the dust optical extinction A V , and the starlight intensity heating the bulk of the dust, parametrized by U min . The DL model reproduces the observed spectral energy distribution (SED) satisfactorily over most of the sky, with small deviations in the inner Galactic disk and in low ecliptic latitude areas, presumably due to zodiacal light contamination. In the Andromeda galaxy (M31), the present dust mass estimates agree remarkably well (within 10%) with DL estimates based on independent Spitzer and Herschel data. We compare the DL optical extinction A V for the diffuse interstellar medium (ISM) with optical estimates for approximately 2 × 10 5 quasi-stellar objects (QSOs) observed inthe Sloan Digital Sky Survey (SDSS). The DL A V estimates are larger than those determined towards QSOs by a factor of about 2, which depends on U min . The DL fitting parameter U min , effectively determined by the wavelength where the SED peaks, appears to trace variations in the far-IR opacity of the dust grains per unit A V , and not only in the starlight intensity. These results show that some of the physical assumptions of the DL model will need to be revised. To circumvent the model deficiency, we propose an empirical renormalization of the DL A V estimate, dependent of U min , which compensates for the systematic differences found with QSO observations. This renormalization, made to match the A V estimates towards QSOs, also brings into agreement the DL A V estimates with those derived for molecular clouds from the near-IR colours of stars in the 2 micron all sky survey (2MASS). The DL model and the QSOs data are also used to compress the spectral information in the Planck and IRAS observations for the diffuse ISM to a family of 20 SEDs normalized per A V , parameterized by U min , which may be used to test and empirically calibrate dust models. Finally, the family of SEDs and the maps generated with the DL model are made public in the Planck Legacy Archive.

79 ASTRONOMY AND ASTROPHYSICS↗

BEYONDPLANCK III. Commander3

We describe the computational infrastructure for end-to-end Bayesian cosmic microwave background (CMB) analysis implemented by the BeyondPlanck Collaboration. The code is called Commander3. It provides a statistically consistent framework for global analysis of CMB and microwave observations and may be useful for a wide range of legacy, current, and future experiments. The paper has three main goals. Firstly, we provide a high-level overview of the existing code base, aiming to guide readers who wish to extend and adapt the code according to their own needs or re-implement it from scratch in a different programming language. Secondly, we discuss some critical computational challenges that arise within any global CMB analysis framework, for instance in-memory compression of time-ordered data, fast Fourier transform optimization, and parallelization and load-balancing. Thirdly, we quantify the CPU and RAM requirements for the current BEYONDPLANCK analysis, finding that a total of 1.5 TB of RAM is required for efficient analysis and that the total cost of a full Gibbs sample for LFI is 170 CPU-hrs, including both low-level processing and high-level component separation, which is well within the capabilities of current low-cost computing facilities. The existing code base is made publicly available under a GNU General Public Library (GPL) license.

79 ASTRONOMY AND ASTROPHYSICS↗

Integration of RNTuple in ATLAS Athena

After using ROOT’s TTree I/O subsystem for over two decades and storing more than an exabyte of compressed High Energy Physics (HEP) data, advances in technology have motivated a complete redesign, RNTuple, which breaks backward-compatibility to take better advantage of these storage options. The RNTuple I/O subsystem has been designed to address performance bottlenecks and other shortcomings of TTree. Specifically, RNTuple comes with an updated, more compact binary data format that can be stored both in ROOT files and natively in object stores. It is designed for modern storage hardware (e.g. high-throughput low-latency NVMe SSDs), and provides robust and easy to use interfaces. The binary format of RNTuple is scheduled to become production grade in 2024, and recently has become mature enough to start exploring the integration into software used by HEP experiments. In this contribution, we discuss the developments to support the features as required by the ATLAS analysis Event Data Model (EDM) in RNTuple, which will enable its integration into the Athena software framework. With these developments in place, we evaluate the performance of the current most recent versions of RNTuple-based ATLAS data sets and compare this to that of TTree.

72 PHYSICS OF ELEMENTARY PARTICLES AND FIELDS↗

The high pressure phase transformation behavior of silicon nanowires

Si nanowires of 80–150 nm and 200–250 nm diameter are pressurized up to 22 GPa using a diamond anvil cell. Raman and x-ray diffraction data were collected during both compression and decompression. Electron microscopy images reveal that the nanowires retain a nanowire-like morphology (after high pressure treatment). On compression, dc-Si was observed to persist at pressures up to 19 GPa compared to ~11 GPa for bulk-Si. On decompression, the metallic β-Sn phase was found to be more stable for Si nanowires compared with bulk-Si when lowering the pressure and was observed as low as 6 GPa. For the smallest nanowires studied (80–150 nm), predominately a-Si was obtained on decompression, whereas for larger nanowires (200–250 nm), clear evidence for the r8/bc8-Si phase was obtained. We suggest that the small volume of the individual Si nanowires compared with bulk-Si inhibits the nucleation of the r8-Si phase on decompression. This study shows that there is a size dependence in the high pressure behavior of Si nanowires during both compression and decompression.

36 MATERIALS SCIENCE↗

Automated Fitting of a Semi-empirical Multiphase Equation of State for Carbon

The equation of state (EOS) of carbon is important in high explosive, geophysical, and inertial confinement fusion applications. Within the semi-empirical Sesame framework, the EOS of each phase is represented by a sum of cold, vibrational, and thermal electronic Helmholtz free energy contributions. Each phase has ~5-10 independent parameters that are adjusted to reproduce single-phase data (e.g., thermal expansion, isothermal compression) as well as experimentally- and computationally- derived phase boundaries. Manual calibration of the full multiphase EOS is arduous. We present our progress in development of automated EOS parameterization based on minimization of an objective function. Here, this function encodes deviation of model EOS results from experimental/computational benchmarks. Optimization is implemented as a combination of global (particle swarm) and local optimization techniques.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗

CMB distance priors revisited: effects of dark energy dynamics, spatial curvature, primordial power spectrum, and neutrino parameters

As a physical and sufficient compression of the full CMB data, the CMB distance priors, or shift parameters, have been widely used and provide a convenient way to include CMB data when obtaining cosmological constraints. In this paper, we revisit this data vector and examine its stability under different cosmological models. We find that the CMB distance priors are an accurate substitute for the full CMB data when probing dark energy dynamics. This is true when the primordial power spectrum model is directly generalized from the power spectrum of the model used in the derivation of the distance priors from the CMB data. We discover a difference when a non-flat model with the untilted primordial inflation power spectrum is used to measure the distance priors. This power spectrum is a radical change from the more conventional tilted primordial power spectrum and violates fundamental assumptions for the reliability of the CMB shift parameters. We also investigate the performance of CMB distance priors when the sum of neutrino masses Σm ν and the effective number of relativistic species N eff are allowed to vary. Our findings are consistent with earlier results: the neutrino parameters can change the measurement of the sound horizon from CMB data, and thus the CMB distance priors. We find that when the neutrino model is allowed to vary, the cold dark matter density ω c and N eff need to be included in the set of parameters that summarize CMB data, in order to reproduce the constraints from the full CMB data. As a result, we present an updated and expanded set of CMB distance priors which can reproduce constraints from the full CMB data within 1σ, and are applicable to models with massive neutrinos, as well as non-standard cosmologies.

79 ASTRONOMY AND ASTROPHYSICS↗

Accelerating Ultrafast Spectroscopy with Compressive Sensing

We report ultrafast spectroscopy is an important tool for studying photoinduced dynamical processes in atoms, molecules, and nanostructures. Typically, the time to perform these experiments ranges from several minutes to hours depending on the choice of spectroscopic method. It is desirable to reduce this time overhead not only to shorten time and laboratory resources, but also to make it possible to examine fragile specimens that quickly degrade during long experiments. In this article, we motivate using compressive sensing to significantly shorten data acquisition time by reducing the total number of measurements in ultrafast spectroscopy. We apply this technique to experimental data from ultrafast transient absorption spectroscopy and ultrafast terahertz spectroscopy and show that good estimates can be obtained with as low as 15% of the total measurements, implying a sixfold reduction in the data acquisition time.

75 CONDENSED MATTER PHYSICS, SUPERCONDUCTIVITY AND↗

BCSR on GPU: A Way Forward Extreme-scale Graph Processing on Accelerator-enabled Frontier Supercomputer

Handling large graphs in a distributed environment requires effective partitioning across processors and efficient management of local partitions. In 2D partitioning, local graphs often become too sparse, making memory-efficient data structures crucial. Using the Compressed Sparse Row (CSR) format wastes space, especially for > 83% of vertices with empty edges for the sparse graphs. This study explores bit-CSR (BCSR), a modified CSR representation, on GPUs to reduce memory usage in graph computations. We achieved 16.67% memory savings on a sparse rmat dataset with 268 million vertices and 357 million edges, without performance degradation, supported by both theoretical and experimental storage savings of 33%. However, we observed a 1.7× slowdown in degree lookup times due to bitwise operations on AMD CPUs. This analysis highlights the potential of BCSR on GPUs for improving Graph500 benchmark performance on GPU-accelerated systems, such as the Frontier supercomputer.

Sattar, Naw Safrin↗

Accelerating Random Forest Classification on GPU and FPGA

Random Forests (RFs) are a commonly used machine learning method for classification and regression tasks spanning a variety of application domains, including bioinformatics, business analytics, and software optimization. While prior work has focused primarily on improving performance of the training of RFs, many applications, such as malware identification, cancer prediction, and banking fraud detection, require fast RF classification. In this work, we accelerate RF classification on GPU and FPGA. In order to provide efficient support for large datasets, we propose a hierarchical memory layout suitable to the GPU/FPGA memory hierarchy. We design three RF classification code variants based on that layout, and we investigate GPU- and FPGA-specific considerations for these kernels. Our experimental evaluation, performed on an Nvidia Xp GPU and on a Xilinx Alveo U250 FPGA accelerator card using publicly available datasets on the scale of millions of samples and tens of features, covers various aspects. First, we evaluate the performance benefits of our hierarchical data structure over the standard compressed sparse row (CSR) format. Second, we compare our GPU implementation with cuML, a machine learning library targeting Nvidia GPUs. Third, we explore the performance/accuracy tradeoff resulting from the use of different tree depths in the RF. Finally, we perform a comparative performance analysis of our GPU and FPGA implementations. Our evaluation shows that for high accuracy targets, our GPU implementation yields 5-9x speedup over CSR, and up to a 2x speedup over cuML.

FPGA, Xilinx FPGA, GPU, Random Forest classificati↗

CORE-BFS: Communication-Optimized REctangular-partitioned BFS Achieving 160.845 TeraTEPS on Frontier Supercomputer

Distributed Breadth-First Search (BFS) is fundamental to many large-scale graph applications, but its performance on parallel systems is often limited by high communication overhead. This paper presents CORE-BFS, an extremely scalable GPU-based BFS implementation that introduces a unique rectangular 2D partitioning-based design for Frontier supercomputer. To further improve performance, we propose four key optimizations: (1) Rectangular 2D-partition specific data formats that use two compressed row and one compressed column status array bitmaps combined with a Double Compressed Sparse Row (DCSR) format per partition, reducing memory footprint and inter-rank traffic; (2) Adaptive frontier & communication strategy that unifies top-down and bottom-up traversal on the rectangular layout, uses lazy synchronization in top-down levels, and switches variants based on frontier size to minimize communication overhead; (3) Frontier-split degree-aware update that maps frontier vertices to thread-centric, wavefront-centric, and block-centric kernels based on their degree to improve GPU utilization and memory coalescing; (4) Row-reduction pipeline that overlaps bottom-up adjacency list processing with row-wise bitmap reduction to hide inter-rank latency. Together, these techniques increase parallelism while reducing memory and communication overhead. On the Graph500 benchmark, CORE - BFS scales up to 9,248 Frontier nodes with scale-42 graphs and reaches 160.845 TTEPS, delivering a 5.42 × speedup over our previous Frontier implementation.

Yang, Haoshen [Rutgers University]↗

FOSWEC v2: System identification and control testing at Hinsdale Wave Research Laboratory

This report summarizes the design and execution of a wave tank test of the floating oscillating surge wave energy converter (FOSWEC) in the O.H. Hinsdale Wave Research Laboratory Directional Wave Basin at Oregon State University. This device, which uses two "flaps" that pivot about a central platform when excited by waves, has a natural frequency within the range of the waves by which it is excited. The FOSWEC was originally considered to be a 1:33 scale device, however, for the current tests, no fixed relative scale is used (i.e., the WEC is considered to be scaled for the basin’s wave environment in which it operates). The primary goal of this test was to assess the degree to which previously developed modeling, experimentation, and control design methods could be applied to a broad range of wave energy converter designs. Testing was conducted to identify a dynamic model for the impedance and excitation behavior of the device. Using these models, a series of closed loop tests were conducted using a causal impedance matching controller. This report provides a brief description of the results, as well as a summary of the device and experimental design. The results show that the methods applied to this experimental device perform well and should be broadly applicable. The data collected during testing is compressed into FOSWEC.zip. Please refer to Appendix C (pages 61-63) of the test report for descriptions of each test ID corresponding to the compressed files.

16 TIDAL AND WAVE POWER↗

Performance Data from a 1-Meter Cross-flow Turbine with High Deflection Hydrofoils

Performance data of a 1-meter diameter cross-flow tidal turbine consisting of three NACA 0018 blades with two support struts with high deflection hydrofoils. Data was collected at the University of New Hampshire Jere A. Chase Ocean Engineering Lab within the tow tank. Three turbine parameters were varied: the blade materials, blade shape, and support strut position. A detailed description of the testing set-up and data files contained within the compressed "Turbine_Performance_Data.zip" file is in the "ReadMe.txt" file. Review of the original dataset "_Ver1" found that one of the tests had issues with one of the two redundant sensors. Resources were updated by replacing the dataset with measurements from the redundant sensor and are provided as version 2 "_Ver2".

16 TIDAL AND WAVE POWER↗

A Pipeline for Large Data Processing Using Regular Sampling for Unstructured Grids

Large simulation data requires a lot of time and computational resources to compute, store, analyze, visualize, and run user studies. Today, the largest cost of a supercomputer is not hardware but maintenance, in particular energy consumption. Our goal is to balance energy consumption and cognitive value of visualizations of resulting data. This requires us to go through the entire processing pipeline, from simulation to user studies. To reduce the amount of resources, data can be sampled or compressed. While this adds more computation time, the computational overhead is negligible compared to the simulation time. We built a processing pipeline at the example of regular sampling. The reasons for this choice are two-fold: using a simple example reduces unnecessary complexity as we know what to expect from the results. Furthermore, it provides a good baseline for future, more elaborate sampling methods. We measured time and energy for each test we did, and we conducted user studies in Amazon Mechanical Turk (AMT) for a range of different results we produced through sampling.

97 MATHEMATICS AND COMPUTING↗

IoGET: Internet of Geophysical and Environmental Things

The objective of this project is to provide novel and fast reduced-order models for onboard computation at sensor nodes for real-time analysis. The approach will require that LANL perform high-fidelity numerical simulations, construct simple reduced-order models (ROMs) using machine learning and signal processing algorithms, and use real-time data analysis for ROMs and compressive sensing at sensor nodes.

54 ENVIRONMENTAL SCIENCES↗

Design and Initial Verification Testing of an In Situ Load Retention Aging Fixture

Load retention is a key metric used to evaluate long-term mechanical functionality of polymeric materials. Its measurement currently requires periodically removing the aging fixtures from the ovens, removing the specimens from the fixtures, and performing load testing on the specimens. The ability to perform load measurement in situ (i.e., without having to remove the specimens from the fixtures, and possibly even the ovens) would facilitate data acquisition/analysis, improve data quality, and enable uninterrupted compression of specimens which is more representative of the actual application.

36 MATERIALS SCIENCE↗

AGC-1 Irradiation Induced Property Changes Analysis Report: Dynamic Elastic Modulus

Recent analyses of some of the test property data from the AGC-1 creep capsule are reported. Specifically the analyses of data for the effects of compressive creep strain on the Dynamic elastic modulus (by the fundamental frequency and ultrasonic velocity/tof methods) are reported. Statistical “t” testing was used to establish the statistical significance of any differences observed between the Dynamic elastic modulus of the creep (irradiated with compressive stress) and the control (irradiated without stress) specimens. Compressive creep strain does have a small effect on Dynamic elastic modulus, reducing it at the higher doses as volume turnaround is approached.

73 NUCLEAR PHYSICS AND RADIATION PHYSICS↗