Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “workload”

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 73 records · Page 4
All sourcesNASA NTRSDOE OSTI
DOE OSTI2026

Dataset of Generative AI Workload Power Profiles

This dataset provides a collection of high-resolution (5/10 Hz or every 0.2/0.1 seconds) power consumption profiles for generative artificial intelligence (GenAI) workloads executed on NLR's High Performance Computing (HPC) platform Kestrel. The dataset also includes examples of representative whole-facility power profiles generated using a bottom-up, event-driven, data center energy model . This dataset is designed to support research in energy modeling, infrastructure planning, energy system integration, and sustainability analysis for AI-driven computing systems. The dataset captures time-resolved electrical power measurements across a diverse set of configurations, including variations in job type (inference vs. training), workload (LLM vs. image generation), datasets, and number of compute nodes. Power traces are provided in a standardized format and include both raw/instantaneous and aggregated files. Each profile is accompanied by metadata describing workload parameters, enabling reproducibility and cross-study comparison. The dataset is intended for use in applications such as data center infrastructure planning, energy modeling, demand response and grid impact studies, and development and validation of system-level simulation tools. By making these workload-specific power profiles publicly available, this dataset aims to address the current lack of open, empirical energy data for generative AI systems and to facilitate transparent, reproducible research on the energy and environmental impacts of large-scale AI deployment. If you use this dataset, please cite the associated publication: Vercellino et al., “Measurement of Generative AI Workload Power Profiles for Whole-Facility Data Center Infrastructure Planning,” arXiv:2604.07345 (2026).

97 MATHEMATICS AND COMPUTING↗
NASA NTRS2017

An Unobtrusive System to Measure, Assess, and Predict Cognitive Workload in Real-World Environments

Across many careers, individuals face alternating periods of high and low attention and cognitive workload, which can result in impaired cognitive functioning and can be detrimental to job performance. For example, some professions (e.g., fire fighters, emergency medical personnel, doctors and nurses working in an emergency room, pilots) require long periods of low workload (boredom), followed by sudden, high-tempo operations during which they may be required to respond to an emergency and perform at peak cognitive levels. Conversely, other professions (e.g., air traffic controllers, market investors in financial industries, analysts) require long periods of high workload and multitasking during which the addition of just one more task results in cognitive overload resulting in mistakes. An unobtrusive system to measure, assess, and predict cognitive workload could warn individuals, their teammates, or their supervisors when steps should be taken to augment cognitive readiness. In this talk I will describe an approach to this problem that we have found to be successful across work domains including: (1) a suite of unobtrusive, field-ready neurophysiological, physiological, and behavioral sensors that are chosen to best suit the target environment; (2) custom algorithms and statistical techniques to process and time-align raw data originating from the sensor suite; (3) probabilistic and statistical models designed to interpret the data into the human state of interest (e.g., cognitive workload, attention, fatigue); (4) and machine-learning techniques to predict upcoming performance based on the current pattern of events, and (5) display of each piece of information depending on the needs of the target user who may or may not want to drill down into the functioning of the system to determine how conclusions about human state and performance are determined. I will then focus in on our experimental results from our custom functional near-infrared spectroscopy sensor, designed to operate in real-world environments to be worn comfortably (e.g., positioned into a baseball cap or a surgeons cap) to measure changes in brain blood oxygenation without adding burden to the individual being assessed.

workload collection↗
NASA NTRS2017

Workload and Performance in Air Traffic Control: Exploring the Influence of Levels of Automation and Variation in Task Demand

In air traffic control, task demand and workload have important implications for the safety and efficiency of air traffic. Task demand is dynamic, however, research on demand transitions and associated controller perception and performance is limited. In addition, there is a comparatively restricted understanding of the influence of task demand transitions on workload and performance, in association with automation. This study used an air traffic control simulation to investigate the influence of task demand transitions and two conditions of varying automation, on workload and efficiency-related performance. Findings showed that a both the direction of the task demand variation, and the amount of automation, influenced the relationship between workload and performance. Further research is needed to enhance understanding of demand transition and workload history effects on operator experience and performance, in both air traffic control and other safety-critical domains.

air traffic control↗
NASA NTRS2018

Survey of Methods to Predict Controller Workload for Real-Time Monitoring of Airspace Safety

The Real Time Safety Monitoring (RTSM) approach allows for assessment and prediction of the safety margin in the National Airspace System (NAS) to help preempt incidents and accidents, rather than having to reactively mitigate them. In RTSM, the NAS is modeled using state variables, and safety metrics are defined in terms of these state variables. The safety metrics have been classified as weather-related, airspace-related, and human-related. Many of the formulated human-related safety metrics need an estimate of the controller workload for their computation. However, this computation is not trivial. Hence, in this report, we perform a literature survey to identify the different factors that enable the computation of controller workload and categorize these factors. Next, we describe studies undertaken to determine a minimal set of factors that pro- vide a correct assessment of controller workload. Lastly, we survey approaches for evaluating how well the selected factors correlate with the controllers' subjective assessment of their workload. Based on this survey, we present factors beneficial to computing and predicting controller workload in real time, and discuss the status of data sources necessary for these computations.

National Airspace System↗
DOE OSTI2018

Scaling Deep Learning workloads: NVIDIA DGX-1/Pascal and Intel Knights Landing

Deep Learning (DL) algorithms have become ubiquitous in data analytics. As a result, major computing vendors—including NVIDIA, Intel, AMD, and IBM—have architectural road maps influenced by DL workloads. Furthermore, several vendors have recently advertised new computing products as accelerating large DL workloads. Unfortunately, it is difficult for data scientists to quantify the potential of these different products. Here, this article provides a performance and power analysis of important DL workloads on two major parallel architectures: NVIDIA DGX-1 (eight Pascal P100 GPUs interconnected with NVLink) and Intel Knights Landing (KNL) CPUs interconnected with Intel Omni-Path or Cray Aries. Our evaluation consists of a cross section of convolutional neural net workloads: CifarNet, AlexNet, GoogLeNet, and ResNet50 topologies using the Cifar10 and ImageNet datasets. The workloads are vendor-optimized for each architecture. We use sequentially equivalent implementations to maintain iso-accuracy between parallel and sequential DL models. Our analysis indicates that although GPUs provide the highest overall performance, the gap can close for some convolutional networks; and the KNL can be competitive in performance/watt. We find that NVLink facilitates scaling efficiency on GPUs. However, its importance is heavily dependent on neural network architecture. Furthermore, for weak-scaling—sometimes encouraged by restricted GPU memory—NVLink is less important.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2021

Characterizing Machine Learning I/O Workloads on Leadership Scale HPC Systems

High performance computing (HPC) is no longer solely limited to traditional workloads such as simulation and modeling. With the increase in the popularity of machine learning (ML) and deep learning (DL) technologies, we are observing that an increasing number of HPC users are incorporating ML methods into their workflow and scientific discovery processes, across a wide spectrum of science domains such as biology, earth science, and physics. This gives rise to a diverse set of I/O patterns than the traditional checkpoint/restart-based HPC I/O behavior. The details of the I/O characteristics of such ML I/O workloads have not been studied extensively for large-scale leadership HPC systems. This paper aims to fill that gap by providing an in-depth analysis to gain an understanding of the I/O behavior of ML I/O workloads using darshan - an I/O characterization tool designed for lightweight tracing and profiling. We study the darshan logs of more than 23, 000 HPC ML I/O jobs over a time period of one year running on Summit - the second-fastest supercomputer in the world. This paper provides a systematic I/O characterization of ML I/O jobs running on a leadership scale supercomputer to understand how the I/O behavior differs across science domains and the scale of workloads, and analyze the usage of parallel file system and burst buffer by ML I/O workloads.

Paul, Arnab↗
NASA NTRS1971

Development of techniques for measuring pilot workload

An objective method of assessing information workload based on physiological measurements was developed. Information workload, or reserve capacity, was measured using a visual discrimination secondary task and subjective rating of task difficulty. The primary task was two axis (pitch and roll) tracking, and the independent variables in this study were aircraft pitch dynamics and wind gust disturbances. The study was structured to provide: (1) a sensitive, nonloading measure of reserve capacity, and (2) an unencumbering reliable measurement of the psychophysiological state. From these, a measured workload index (MWI) and physiological workload index (PWI) were extracted. An important measure of the success of this study was the degree to which the MWI and PWI agreed across the 243 randomly-presented, four-minute trials (9 subjects X 9 tasks X 3 replications). The electrophysiological data collected included vectorcardiogaram, respiration, electromyogram, skin impedance, and electroencephalogram. Special computer programs were created for the analysis of each physiological variable. The digital data base then consisted of 82 physiological features for each of the 243 trials. A prediction of workload based on physiological observations was formulated as a simultaneous least-squares prediction problem. A best subset of 10 features was chosen to predict the three measures of reserve capacity. The cannonical correlation coefficient was .754 with a chi squared value of 91.3 which allows rejection of the null hypothesis with p of .995.

Spyker, D. A.↗
NASA NTRS1982

Subjective scaling of mental workload in a multi-task environment

Those factors in a multi-task environment that contribute to the operators' "sense" of mental workload were identified. The subjective judgment as conscious experience of mental effort was decided to be the appropriate method of measurement. Thurstone's law of comparative judgment was employed in order to construct interval scales of subjective mental workload from paired comparisons data. An experimental paradigm (Simulated Multi-Task Decision-Making Environment) was employed to represent the ideal experimentally controlled environment in which human operators were asked to "attend" to different cases of Tulga's decision making tasks. Through various statistical analyses it was found that, in general, a lower number of tasks-to-be-processed per unit time (a condition associated with longer interarrival times), results in a lower mental workload, a higher consistency of judgments within a subject, a higher degree of agreement among the subjects, and larger distances between the cases on the Thurstone scale of subjective mental workload. The effects of various control variables and their interactions, and the different characteristics of the subjects on the variation of subjective mental workload are demonstrated.

Daryanian, B.↗
NASA NTRS1983

Comparative evaluation of workload estimation techniques in piloting tasks

Techniques to measure operator workload in a wide range of situations and tasks were examined. The sensitivity and intrusion of a wide variety of workload assessment techniques in simulated piloting tasks were investigated. Four different piloting tasks, psychomotor, perceptual, mediational, and communication aspects of piloting behavior were selected. Techniques to determine relative sensitivity and intrusion were applied. Sensitivity is the relative ability of a workload estimation technique to discriminate statistically significant differences in operator loading. High sensitivity requires discriminable changes in score means as a function of load level and low variation of the scores about the means. Intrusion is an undesirable change in the task for which workload is measured, resulting from the introduction of the workload estimation technique or apparatus.

Wierwille, W. W.↗
NASA NTRS1982

The reliability and validity of flight task workload ratings

Twelve instrument-rated general aviation pilots each flew two scenarios in a motion-base simulator. During each flight, the pilots verbally estimated their workload every three minutes. Following each flight, they again estimated workload for each flight segment and also rated their overall workload, perceived performance, and 13 specific factors on a bipolar scale. The results indicate that time (a priori, inflight, or postflight) of eliciting ratings, period to be covered by the ratings (a specific moment in time or a longer period), type of rating scale, and rating method (verbal, written, or other) may be important variables. Overall workload ratings appear to be predicted by different specific scales depending upon the situation, with activity level the best predictor. Perceived performance seems to bear little relationship to observer-rated performance when pilots rate their overall performance and an observer rates specific behaviors. Perceived workload and performance also seem unrelated.

Childress, M. E.↗
NASA NTRS1984

On the measurement of pilot perceptual workload - A comparison of assessment techniques addressing sensitivity and intrusion issues

A flight simulator-based study was conducted to examine fourteen distinct mental workload estimation measures, including opinion, secondary task, physiological, and primary task measures. Both the relative sensitivity of the measures to changes in mental workload and the differential intrusion of the changes on primary task performance were assessed. The flight task was varied in difficulty by manipulation of the presentation rate and complexity of a hazard-perception task that required each of 48 licensed pilots to rely heavily on their perceptual abilities. Three rating scales (Modified Cooper-Harper, Multi-descriptor, and Workload-Compensation-Interference/Technical Effectiveness), two secondary task measures (time estimation and tapping regularity), one physiological measure (respiration frequency), and one primary task measure (danger-condition response time) were reliable indicants of workload changes. Recommendations for applying the workload measures are presented.

Casali, J. G.↗
NASA NTRS1986

Fault and Error Latency Under Real Workload: an Experimental Study

A practical methodology for the study of fault and error latency is demonstrated under a real workload. This is the first study that measures and quantifies the latency under real workload and fills a major gap in the current understanding of workload-failure relationships. The methodology is based on low level data gathered on a VAX 11/780 during the normal workload conditions of the installation. Fault occurrence is simulated on the data, and the error generation and discovery process is reconstructed to determine latency. The analysis proceeds to combine the low level activity data with high level machine performance data to yield a better understanding of the phenomena. A strong relationship exists between latency and workload and that relationship is quantified. The sampling and reconstruction techniques used are also validated. Error latency in the memory where the operating system resides was studied using data on the physical memory access. Fault latency in the paged section of memory was determined using data from physical memory scans. Error latency in the microcontrol store was studied using data on the microcode access and usage.

Chillarege, Ram↗
NASA NTRS1987

Inflight application of three pilot workload measurement techniques

Three inflight techniques for workload measurement were tested in nine pilots flying the NASA Kuiper Airborne Observatory: subjective ratings, heart rate, and communication performance. The activities that contributed to the crew-member workload varied; the commander was responsible for aircraft control and navigation whereas the copilot handled communications. The three workload measures were found to provide different information. Pilot ratings of workload, effort, and stress were sensitive to variations in flight-related task demands across flight segments but did not reflect specific differences in the type of demands imposed on the commander and the copilot. The heart rate was sensitive to the differential impact of duties, being higher for the commander than for the copilot. The rate of communications per minute of flight proved to be the most sensitive indicator. It was related to workload, stress, effort rating, and average heart rate across flight segments.

Hart, Sandra G.↗
NASA NTRS1987

Research papers and publications (1981-1987): Workload research program

An annotated bibliography of the research reports written by participants in NASA's Workload Research Program since 1981 is presented, representing the results of theoretical and applied research conducted at Ames Research Center and at universities and industrial laboratories funded by the program. The major program elements included: 1) developing an understanding of the workload concept; 2) providing valid, reliable, and practical measures of workload; and 3) creating a computer model to predict workload. The goal is to provide workload-related design principles, measures, guidelines, and computational models. The research results are transferred to user groups by establishing close ties with manufacturers, civil and military operators of aerospace systems, and regulatory agencies; publishing scientific articles; participating in and sponsoring workshops and symposia; providing information, guidelines, and computer models; and contributing to the formulation of standards. In addition, the methods and theories developed have been applied to specific operational and design problems at the request of a number of industry and government agencies.

Hart, Sandra G.↗
NASA NTRS1987

Comparison of workload measures on computer-generated primary flight displays

Four Air Force pilots were used as subjects to assess a battery of subjective and physiological workload measures in a flight simulation environment in which two computer-generated primary flight display configurations were evaluated. A high- and low-workload task was created by manipulating flight path complexity. Both SWAT and the NASA-TLX were shown to be effective in differentiating the high and low workload path conditions. Physiological measures were inconclusive. A battery of workload measures continues to be necessary for an understanding of the data. Based on workload, opinion, and performance data, it is fruitful to pursue research with a primary flight display and a horizontal situation display integrated into a single display.

Nataupsky, Mark↗
NASA NTRS1989

Crew procedures and workload of retrofit concepts for microwave landing system

Crew procedures and workload for Microwave Landing Systems (MLS) that could be retrofitted into existing transport aircraft were evaluated. Two MLS receiver concepts were developed. One is capable of capturing a runway centerline and the other is capable of capturing a segmented approach path. Crew procedures were identified and crew task analyses were performed using each concept. Crew workload comparisons were made between the MLS concepts and an ILS baseline using a task-timeline workload model. Workload indexes were obtained for each scenario. The results showed that workload was comparable to the ILS baseline for the MLS centerline capture concept, but significantly higher for the segmented path capture concept.

Summers, Leland G.↗
NASA NTRS1991

A user-oriented synthetic workload generator

A user oriented synthetic workload generator that simulates users' file access behavior based on real workload characterization is described. The model for this workload generator is user oriented and job specific, represents file I/O operations at the system call level, allows general distributions for the usage measures, and assumes independence in the file I/O operation stream. The workload generator consists of three parts which handle specification of distributions, creation of an initial file system, and selection and execution of file I/O operations. Experiments on SUN NFS are shown to demonstrate the usage of the workload generator.

Kao, Wei-Lun↗
NASA NTRS1994

Practical guidelines for workload assessment

The practical problems that might be encountered in carrying out workload evaluations in work settings have been outlined. Different approaches have been distinguished that may determine the type of research design used and provide assistance in the difficult choice between workload assessment techniques. One approach to workload assessment is to examine the short-term consequences of combining various tasks. Theoretical models of attention allocation will underpin specific studies of interference and the consequences of task demand and task conflict for performance. A further approach with a different temporal orientation may lead us to a better understanding of the relationships between work demands and strain through the analysis of individual differences in cognitive control processes. The application of these processes may depend on individual differences in long term styles and short term strategies, but may be used to prevent decrements in work performance under difficult conditions. However, control may attract costs as well as benefits in terms of changes in effective state and physiological activity. Thus, strain associated with work demands may only be measurable in the form of tradeoffs between performance and other domains of individual activity. The methodological implications are to identify patterns of adjustment to workload variations using repeated measures and longitudinal sampling of performance as well as subjective and physiological measures. Possible enhancements to workplace design must take into account these human factors considerations of workload in order to avoid potential decrements in individual performance and associated organizational problems.

Tattersall, Andrew J.↗