Search NASA⌕ Search

SEARCH · Search NASA

Results for “read performance”

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 109 records · Page 6
All sourcesNASA NTRSDOE OSTI
DOE OSTI2022

CAPG: comprehensive allopolyploid genotyper

Genotyping by sequencing is a powerful tool for investigating genetic variation in plants, but many economically important plants are allopolyploids, where homoeologous similarity obscures the subgenomic origin of reads and confounds allelic and homoeologous SNPs. Recent polyploid genotyping methods use allelic frequencies, rate of heterozygosity, parental cross or other information to resolve read assignment, but good subgenomic references offer the most direct information. The typical strategy aligns reads to the joint reference, performs diploid genotyping within each subgenome, and filters the results, but persistent read misassignment results in an excess of false heterozygous calls. We introduce the Comprehensive Allopolyploid Genotyper (CAPG), which formulates an explicit likelihood to weight read alignments against both subgenomic references and genotype individual allopolyploids from whole-genome resequencing data. We demonstrate CAPG in allotetraploids, where it performs better than Genome Analysis Toolkit’s HaplotypeCaller applied to reads aligned to the combined subgenomic references.

59 BASIC BIOLOGICAL SCIENCES↗
DOE OSTI2016

Storage and Read-Optimized Data Placement Structures for High-Performance Analysis [Book Chapter]

Here, we present state-of-art structures and methods for efficient data preparation and representation for analysis. Our intent is to introduce the data science and analytics communities to open source data placements, structures, and methods. These practices can make the foundational processes of data preparation and access dramatically more efficient than typical raw file or database representations and use more conservative storage. To illustrate this, we introduce two highly efficient data placement structures. We then present a tutorial, supported by step-by-step examples, of how to create, use, and access data, structured by Parquet or ORC (Optimized Row Columnar), using Apache Spark. Finally, we illustrate the benefits of using these structures with computational and storage volume benchmarks.

97 MATHEMATICS AND COMPUTING↗
NASA NTRS2009

Influence of Combined Whole-Body Vibration Plus G-Loading on Visual Performance

Recent engineering analyses of the integrated Ares-Orion stack show that vibration levels for Orion crews have the potential to be much higher than those experienced in Gemini, Apollo, and Shuttle vehicles. Of particular concern to the Constellation Program (CxP) is the 12 Hz thrust oscillation (TO) that the Ares-I rocket develops during the final ~20 seconds preceding first-stage separation, at maximum G-loading. While the structural-dynamic mitigations being considered can assure that vibration due to TO is reduced to below the CxP crew health limit, it remains to be determined how far below this limit vibration must be reduced to enable effective crew performance during launch. Moreover, this "performance" vibration limit will inform the operations concepts (and crew-system interface designs) for this critical phase of flight. While Gemini and Apollo studies provide preliminary guidance, the data supporting the historical limits were obtained using less advanced interface technologies and very different operations concepts. In this study, supported by the Exploration Systems Mission Directorate (ESMD) Human Research Program, we investigated display readability-a fundamental prerequisite for any interaction with electronic crew-vehicle interfaces-while observers were subjected to 12 Hz vibration superimposed on the 3.8 G loading expected for the TO period of ascent. Two age-matched groups of participants (16 general population and 13 Crew Office) performed a numerical display reading task while undergoing sustained 3.8 G loading and whole-body vibration at 0, 0.15, 0.3, 0.5, and 0.7 g in the eyeballs in/out (x-axis) direction. The time-constrained reading task used an Orion-like display with 10- and 14-pt non-proportional sans-serif fonts, and was designed to emulate the visual acquisition and processing essential for crew system monitoring. Compared to the no-vibration baseline, we found no significant effect of vibration at 0.15 and 0.3 g on task error rates (ER) or response times (RT). Significant degradations in both ER and RT, however, were observed at 0.5 and 0.7 g for 10-pt, and at 0.7 g for 14-pt font displays. These objective performance measures were mirrored by participants' subjective ratings. Interestingly, we found that the impact of vibration on ER increased with distance from the center of the display, but only for vertical displacements. Furthermore, no significant ER or RT aftereffects were detected immediately following vibration, regardless of amplitude. Lastly, given that our reading task required no specialized spaceflight expertise, our finding that effects were not statistically distinct between our two groups is not surprising. The results from this empirical study provide initial guidance for evaluating the display readability trade-space between text-font size and vibration amplitude. However, the outcome of this work should be considered preliminary in nature for a number of reasons: 1. The single 12 Hz x-axis vibration employed was based on earlier load-cycle models of the induced TO environment at the end of Ares-I first stage flight. Recent analyses of TO mitigation designs suggest that significant concurrent off-axis vibration may also occur. 2. The shirtsleeve environment in which we tested fails to capture the full kinematic and dynamic complexity of the physical interface between crewmember and the still-to-bematured helmet-suit-seat designs, and the impact these will have for vibration transmission and consequent performance. 3. By examining performance in this reading and number processing task, we are only assessing readability, a first and necessary step that in itself does not directly address the performance of more sophisticated operational tasks such as vehicle-health monitoring or manual control of the vehicle.

readability↗
DOE OSTI2020

PANTHER: A Programmable Architecture for Neural Network Training Harnessing Energy-Efficient ReRAM

The wide adoption of deep neural networks has been accompanied by ever-increasing energy and performance demands due to the expensive nature of training them. Additionally, numerous special-purpose architectures have been proposed to accelerate training: both digital and hybrid digital-analog using resistive RAM (ReRAM) crossbars. ReRAM-based accelerators have demonstrated the effectiveness of ReRAM crossbars at performing matrix-vector multiplication operations that are prevalent in training. However, they still suffer from inefficiency due to the use of serial reads and writes for performing the weight gradient and update step. A few works have demonstrated the possibility of performing outer products in crossbars, which can be used to realize the weight gradient and update step without the use of serial reads and writes. However, these works have been limited to low precision operations which are not sufficient for typical training workloads. Moreover, they have been confined to a limited set of training algorithms for fully-connected layers only. To address these limitations, we propose a bit-slicing technique for enhancing the precision of ReRAM-based outer products, which is substantially different from bit-slicing for matrix-vector multiplication only. We incorporate this technique into a crossbar architecture with three variants catered to different training algorithms. To evaluate our design on different types of layers in neural networks (fully-connected, convolutional, etc.) and training algorithms, we develop PANTHER, an ISA-programmable training accelerator with compiler support. Our design can also be integrated into other accelerators in the literature to enhance their efficiency. Our evaluation shows that PANTHER achieves up to 8.02×, 54.21×, and 103× energy reductions as well as 7.16×, 4.02×, and 16× execution time reductions compared to digital accelerators, ReRAM-based accelerators, and GPUs, respectively.

42 ENGINEERING↗
DOE OSTI2014

Thermal Analysis and Thermal Properties of ANPZ and DNDMP

The thermal properties of ANPZ and DNDMP are reported here. These properties were measured upon request for Pantex. The following is a list of measurements requested: Density; Decomposition Energy; Pre-exponential Factor; Activation Energy; and, Thermal Conductivity. We have reported all most in SI units with the expectation that unit conversions can easily be performed by any individual reading this report. In many cases, we have provides more information than requested as it seemed relevant or necessary for performing safety calculations on these materials. The measurement methods and experimental details are described briefly in each section of this report.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗
NASA NTRS1978

Inconsistencies in tape read and tape write programs on the I-100 image analysis system

The author has identified the following significant results. The tape read and write programs currently available on the 1-100 perform their intended functions of reading and writing tapes, but are difficult to use because they contain a number of inconsistencies. These inconsistencies can often be overcome by the use of work-around procedures and by trial and error, which is an inefficient use of expensive computer systems that should not be necessary.

Hocutt, W. T.↗
DOE OSTI2020

LOGAN: High-Performance GPU-Based X-Drop Long-Read Alignment

Pairwise sequence alignment is one of the most computationally intensive kernels in genomic data analysis, accounting for more than 90% of the runtime for key bioinformatics applications. This method is particularly expensive for third-generation sequences due to the high computational cost of analyzing sequences of length between 1Kb and 1Mb. Given the quadratic overhead of exact pairwise algorithms for long alignments, the community primarily relies on approximate algorithms that search only for high-quality alignments and stop early when one is not found. In this work, we present the first GPU optimization of the popular X-drop alignment algorithm, that we named LOGAN. Results show that our high-performance multi-GPU implementation achieves up to 181.6 GCUPS and speed-ups up to 6.6× and 30.7× using 1 and 6 NVIDIA Tesla V100, respectively, over the state-of-the-art software running on two IBM Power9 processors using 168 CPU threads, with equivalent accuracy. We also demonstrate a 2.3× LOGAN speed-up versus ksw2, a state-of-art vectorized algorithm for sequence alignment implemented in minimap2, a long-read mapping software. Furthermore, to highlight the impact of our work on a real-world application, we couple LOGAN with a many-to-many long-read alignment software called BELLA, and demonstrate that our implementation improves the overall BELLA runtime by up to 10.6×. Finally, we adapt the Roofline model for LOGAN and demonstrate that our implementation is near optimal on the NVIDIA Tesla V100s.

97 MATHEMATICS AND COMPUTING↗
NASA NTRS2017

Investigating Access Performance of Long Time Series with Restructured Big Model Data

Data sets generated by models are substantially increasing in volume, due to increases in spatial and temporal resolution, and the number of output variables. Many users wish to download subsetted data in preferred data formats and structures, as it is getting increasingly difficult to handle the original full-size data files. For example, application research users such as those involved with wind or solar energy, or extreme weather events are likely only interested in daily or hourly model data at a single point (or for a small area) for a long time period, and prefer to have the data downloaded in a single file. With native model file structures, such as hourly data from NASA Modern-Era Retrospective analysis for Research and Applications Version-2 (MERRA-2), it may take over 10 hours for the extraction of parameters-of-interest at a single point for 30 years. The NASA Goddard Earth Sciences Data and Information Services Center (GES DISC) is exploring methods to address this particular user need. One approach is to create value-added data by reconstructing the data files. Taking MERRA-2 data as an example, we have tested converting hourly data from one-day-per-file into different data cubes, such as one-month, or one-year. Performance is compared for reading local data files and accessing data through interoperable services, such as OPeNDAP. Results show that, compared to the original file structure, the new data cubes offer much better performance for accessing long time series. We have noticed that performance is associated with the cube size and structure, the compression method, and how the data are accessed. An optimized data cube structure will not only improve data access, but also may enable better online analysis services

reanalysis↗
NASA NTRS2013

Contact Thermocouple Methodology and Evaluation for Temperature Measurement in the Laboratory

Laboratory testing of advanced aerospace components very often requires highly accurate temperature measurement and control devices, as well as methods to precisely analyze and predict the performance of such components. Analysis of test articles depends on accurate measurements of temperature across the specimen. Where possible, this task is accomplished using many thermocouples welded directly to the test specimen, which can produce results with great precision. However, it is known that thermocouple spot welds can initiate deleterious cracks in some materials, prohibiting the use of welded thermocouples. Such is the case for the nickel-based superalloy MarM-247, which is used in the high temperature, high pressure heater heads for the Advanced Stirling Converter component of the Advanced Stirling Radioisotope Generator space power system. To overcome this limitation, a method was developed that uses small diameter contact thermocouples to measure the temperature of heater head test articles with the same level of accuracy as welded thermocouples. This paper includes a brief introduction and a background describing the circumstances that compelled the development of the contact thermocouple measurement method. Next, the paper describes studies performed on contact thermocouple readings to determine the accuracy of results. It continues on to describe in detail the developed measurement method and the evaluation of results produced. A further study that evaluates the performance of different measurement output devices is also described. Finally, a brief conclusion and summary of results is provided.

thermocouple↗
DOE OSTI2023

Automated gauge reading and related systems, methods, and devices

Computing devices and methods for reading gauges are disclosed. A gauge reading method includes capturing image data corresponding to a captured image of one or more gauges, detecting one or more gauges in the captured image, cropping a detected gauge in the captured image to provide a use image including the detected gauge, and classifying the detected gauge to correlate the detected gauge with a template image. The gauge reading method also includes attempting to perform feature detection rectification on the use image to produce a rectified image of the detected gauge, performing template matching rectification on the use image to produce the rectified image responsive to a failure to perform the feature detection rectification, and estimating a gauge reading responsive to the rectified image. A computing device may implement at least a portion of a gauge reading method.

46 INSTRUMENTATION RELATED TO NUCLEAR SCIENCE AND ↗
DOE OSTI2022

Scalable training of graph convolutional neural networks for fast and accurate predictions of HOMO-LUMO gap in molecules

Abstract Graph Convolutional Neural Network (GCNN) is a popular class of deep learning (DL) models in material science to predict material properties from the graph representation of molecular structures. Training an accurate and comprehensive GCNN surrogate for molecular design requires large-scale graph datasets and is usually a time-consuming process. Recent advances in GPUs and distributed computing open a path to reduce the computational cost for GCNN training effectively. However, efficient utilization of high performance computing (HPC) resources for training requires simultaneously optimizing large-scale data management and scalable stochastic batched optimization techniques. In this work, we focus on building GCNN models on HPC systems to predict material properties of millions of molecules. We use HydraGNN, our in-house library for large-scale GCNN training, leveraging distributed data parallelism in PyTorch. We use ADIOS, a high-performance data management framework for efficient storage and reading of large molecular graph data. We perform parallel training on two open-source large-scale graph datasets to build a GCNN predictor for an important quantum property known as the HOMO-LUMO gap. We measure the scalability, accuracy, and convergence of our approach on two DOE supercomputers: the Summit supercomputer at the Oak Ridge Leadership Computing Facility (OLCF) and the Perlmutter system at the National Energy Research Scientific Computing Center (NERSC). We present our experimental results with HydraGNN showing (i) reduction of data loading time up to 4.2 times compared with a conventional method and (ii) linear scaling performance for training up to 1024 GPUs on both Summit and Perlmutter.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗
NASA NTRS2001

Hiding the Disk and Network Latency of Out-of-Core Visualization

This paper describes an algorithm that improves the performance of application-controlled demand paging for out-of-core visualization by hiding the latency of reading data from both local disks or disks on remote servers. The performance improvements come from better overlapping the computation with the page reading process, and by performing multiple page reads in parallel. The paper includes measurements that show that the new multithreaded paging algorithm decreases the time needed to compute visualizations by one third when using one processor and reading data from local disk. The time needed when using one processor and reading data from remote disk decreased by two thirds. Visualization runs using data from remote disk actually ran faster than ones using data from local disk because the remote runs were able to make use of the remote server's high performance disk array.

Ellsworth, David↗
DOE OSTI2013

SDS: A Framework for Scientific Data Services

Large-scale scientific applications typically write their data to parallel file systems with organizations designed to achieve fast write speeds. Analysis tasks frequently read the data in a pattern that is different from the write pattern, and therefore experience poor I/O performance. In this paper, we introduce a prototype framework for bridging the performance gap between write and read stages of data access from parallel file systems. We call this framework Scientific Data Services, or SDS for short. This initial implementation of SDS focuses on reorganizing previously written files into data layouts that benefit read patterns, and transparently directs read calls to the reorganized data. SDS follows a client-server architecture. The SDS Server manages partial or full replicas of reorganized datasets and serves SDS Clients' requests for data. The current version of the SDS client library supports HDF5 programming interface for reading data. The client library intercepts HDF5 calls and transparently redirects them to the reorganized data. The SDS client library also provides a querying interface for reading part of the data based on user-specified selective criteria. We describe the design and implementation of the SDS client-server architecture, and evaluate the response time of the SDS Server and the performance benefits of SDS.

59 BASIC BIOLOGICAL SCIENCES↗
DOE OSTI2023

Improving Progressive Retrieval for HPC Scientific Data using Deep Neural Network

As the disparity between compute and I/O on high-performance computing systems has continued to widen, it has become increasingly difficult to perform post-hoc data analytics on full-resolution scientific simulation data due to the high I/O cost. Error-bounded data decomposition and progressive data retrieval framework has recently been developed to address such a challenge by performing data decomposition before storage and reading only part of the decomposed data when necessary. However, the performance of the progressive retrieval framework has been suffering from the over-pessimistic error control theory, such that the achieved maximum error of recomposed data is significantly lower than the required error. Therefore, more data than required is fetched for recomposition, incurring additional I/O overhead. In order to tackle this issue, we propose a DNN-based progressive retrieval framework that can better identify the minimum amount of data to be retrieved. Our contributions are as follows: 1) We provide an in-depth investigation of the recently developed progressive retrieval framework; 2) We propose two designs of prediction models (named D-MGARD and E-MGARD) to estimate the amount of retrieved data size based on error bounds. 3) We evaluate our proposed solutions using scientific datasets generated by real-world simulations from two domains. Evaluation results demonstrate the effectiveness of our solution in accurately predicting the amount of retrieval data size, as well as the advantages of our solution over the traditional approach to reducing the I/O overhead. Based on our evaluation, our solution is shown to read significantly less data (5% - 40% with D-MGARD, 20% - 80% with E-MGARD).

Wang, Jinzhen↗
NASA NTRS2004

Performance Effects of Display Incogruity in a Digital and Analog Clock Reading Task

In an era of increasing automation, it is important to design displays and input devices that minimize human error. In this context, information concerning the human response to the detection of incongruous information is important. Such incongruous information can be operationalized as unexpected (perhaps erroneous) information on which a decision by the human or operation by an automated system is based. In the aviation environment, decision making when faced with inadequate, incomplete, or incongruous information may occur in a failure scenario. An additional challenge facing the human operator in automated environments is maintaining alertness or vigilance. The vigilance issue is of particular concern as a factor that may interact with performance when faced with inadequate, incomplete, or incongruous information. From the literature on eye-scan behavior we know that the time spent looking at a particular display or indicator is a function of the type of information one is trying to discern from the display. For example, quick glances are all it takes for confirming that an indicator is in a normal position or range, whereas a continuous look of several seconds may be required for confirmation that a complex control input is having the desired effect. Important to consider is that while an extended look takes place, visual input from other sources may be missed. Much like an extended look, the interpretation of incongruous information may require extra time. The present experiment was designed to explore the performance consequences of a decision making task when incongruous information was presented. For this experiment a display incongruity was created on a subset of trials of a clock reading laboratory task. Display incongruity was made possible through presentation of 'impossible' times (e.g. 1:65 or 11:90). Subjects made 'same' 'different' decisions and keyboard responses to pairings of Analog-Analog (AA), Digital-Digital (DD), and Analog- Digital (AD), display combinations. For trials during which display incongruities were not presented, based on prior research comparing digital and analog clock displays, it would be expected that the Digital-Digital condition would result in the shortest response times and the Analog-Analog and Analog-Digital conditions would have longer response times. The performance consequence expected on trials with incongruous times would be very long response times.

J. Raymond Comstock, Jr.↗
DOE OSTI2018

Correction to: Capture and Emission of Charge Carriers by Quantum Well

The last paragraph of the article should read “This work is performed with financial support of the Ministry of Education and Science of the Russian Federation within the framework of the federal target program “Research and development in priority areas of development of the scientific and technological complex of Russia for 2014-2020”, Unique identifier of the work (project) RFMEFI57717X0266.”.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗
DOE OSTI2021

Fast and accurate metagenotyping of the human gut microbiome with GT-Pro

Single nucleotide polymorphisms (SNPs) in metagenomics are used to quantify population structure, track strains and identify genetic determinants of microbial phenotypes. However, existing alignment-based approaches for metagenomic SNP detection require high-performance computing and enough read coverage to distinguish SNPs from sequencing errors. To address these issues, we developed the GenoTyper for Prokaryotes (GT-Pro), a suite of methods to catalog SNPs from genomes and use unique k-mers to rapidly genotype these SNPs from metagenomes. Compared to methods that use read alignment, GT-Pro is more accurate and two orders of magnitude faster. Here, using high-quality genomes, we constructed a catalog of 104 million SNPs in 909 human gut species and used unique k-mers targeting this catalog to characterize the global population structure of gut microbes from 7,459 samples. GT-Pro enables fast and memory-efficient metagenotyping of millions of SNPs on a personal computer.

59 BASIC BIOLOGICAL SCIENCES↗