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 1,045 records · Page 58

C-SAW: a framework for graph sampling and random walk on GPUs

Many applications require to learn, mine, analyze and visualize large-scale graphs. These graphs are often too large to be addressed efficiently using conventional graph processing technologies. Fortunately, recent research efforts find out graph sampling and random walk, which significantly reduce the size of original graphs, can benefit the tasks of learning, mining, analyzing and visualizing large graphs by capturing the desirable graph properties. This paper introduces C-SAW, the first framework that accelerates Sampling and Random Walk framework on GPUs. Particularly, C-SAW makes three contributions: First, our framework provides a generic API which allows users to implement a wide range of sampling and random walk algorithms with ease. Second, offloading this framework on GPU, we introduce warp-centric parallel selection, and two novel optimizations for collision migration. Third, towards supporting graphs that exceed the GPU memory capacity, we introduce efficient data transfer optimizations for out-of-memory and multi-GPU sampling, such as workload-aware scheduling and batched multi-instance sampling. Taken together, our framework constantly outperforms the state of the art projects in addition to the capability of supporting a wide range of sampling and random walk algorithms.

97 MATHEMATICS AND COMPUTING↗

A Step Towards the Final Frontier: Lessons Learned from Acceptance Testing of the First HPE/Cray EX 3000 System at ORNL

In this paper, we summarize the deployment of the Air Force Weather (AFW) HPC11 system at Oak Ridge National Laboratory (ORNL) including the process followed to successfully complete acceptance testing of the system. HPC11 is the first HPE/Cray EX 3000 system that has been successfully released to its user community in a federal facility. HPC11 consists of two identical 800-node supercomputers, Fawbush and Miller, with access to two independent and identical Lustre parallel file systems. HPC11 is equipped with Slingshot 10 interconnect technology and relies on the HPE Performance Cluster Manager (HPCM) software for system configuration. ORNL has a clearly defined acceptance testing process used to ensure that every new system deployed can provide the necessary capabilities to support user workloads. We worked closely with HPE and AFW to develop a set of tests that used the United Kingdom’s Meteorological Office’s Unified Model (UM) and 4DVAR. We also included benchmarks and applications from the Oak Ridge Leadership Computing Facility (OLCF) portfolio to fully exercise the HPE/Cray programming environment and evaluate the functionality and performance of the system. Acceptance testing of HPC11 required parallel execution of each element on Fawbush and Miller. In addition, careful coordination was needed to ensure successful acceptance of the newly deployed Lustre file systems alongside the compute resources. In this work, we present test results from specific system components and provide an overview of the issues identified, challenges encountered, and the lessons learned along the way.

Melesse Vergara, Veronica↗

Using HEP experiment workflows for the benchmarking and accounting of WLCG computing resources

Benchmarking of CPU resources in WLCG has been based on the HEP-SPEC06 (HS06) suite for over a decade. It has recently become clear that HS06, which is based on real applications from non-HEP domains, no longer describes typical HEP workloads. The aim of the HEP-Benchmarks project is to develop a new benchmark suite for WLCG compute resources, based on real applications from the LHC experiments. By construction, these new benchmarks are thus guaranteed to have a score highly correlated to the throughputs of HEP applications, and a CPU usage pattern similar to theirs. Linux containers and the CernVM-FS filesystem are the two main technologies enabling this approach, which had been considered impossible in the past. In this paper, we review the motivation, implementation and outlook of the new benchmark suite.

73 NUCLEAR PHYSICS AND RADIATION PHYSICS↗

Reusability First: Toward FAIR Workflows

The FAIR principles of open science (Findable, Accessible, Interoperable, and Reusable) have had transformative effects on modern large-scale computational science. In particular, they have encouraged more open access to and use of data, an important consideration as collaboration among teams of researchers accelerates and the use of workflows by those teams to solve problems increases. How best to apply the FAIR principles to workflows themselves, and software more generally, is not yet well understood. We argue that the software engineering concept of technical debt management provides a useful guide for application of those principles to workflows, and in particular that it implies reusability should be considered as ‘first among equals’. Moreover, our approach recognizes a continuum of reusability where we can make explicit and selectable the tradeoffs required in workflows for both their users and developers.To this end, we propose a new abstraction approach for reusable workflows, with demonstrations for both synthetic workloads and real-world computational biology workflows. Through application of novel systems and tools that are based on this abstraction, these experimental workflows are refactored to rightsize the granularity of workflow components to efficiently fill the gap between end-user simplicity and general customizability. Our work makes it easier to selectively reason about and automate the connections between trade-offs across user and developer concerns when exposing degrees of freedom for reuse. Additionally, by exposing fine-grained reusability abstractions we enable performance optimizations, as we demonstrate on both institutional-scale and leadership-class HPC resources.

Wolf, Matthew↗

Leaky Buddies: Cross-Component Covert Channels on Integrated CPU-GPU Systems

Graphics Processing Units (GPUs) are ubiquitous components used across the range of today’s computing plat- forms, from phones and tablets, through personal computers, to high-end server class platforms. With the increasing importance of graphics and video workloads, recent processors are shipped with GPU devices that are integrated on the same chip. Integrated GPUs share some resources with the CPU and as a result, there is a potential for microarchitectural attacks from the GPU to the CPU or vice versa. We consider the potential for covert channel attacks that arise either from shared microarchitectural components (such as caches) or through shared contention domains (e.g., shared buses). We illustrate these two types of channels by developing two reliable covert channel attacks. The first covert channel uses the shared LLC cache in Intel’s integrated GPU architectures. The second is a contention based channel targeting the ring bus connecting the CPU and GPU to the LLC. This is the first demonstrated microarchitectural attack crossing the component boundary (GPU to CPU or vice versa). Cross-component channels introduce a number of new challenges that we had to overcome since they occur across heterogeneous components that use different computation models and are interconnected using asymmetric memory hierarchies. We also exploit GPU parallelism to increase the bandwidth of the communication, even without relying on a common clock. The LLC based channel achieves a bandwidth of 120 kbps with a low error rate of 2%, while the contention based channel delivers up to 400 kbps with a 0.8% error rate. We also demonstrate a proof- of-concept prime-and-probe side channel attack that probes the full LLC from the GPU.

Cyber secruity, high performance comptuing, Hetero↗

Dynamic Human-in-the-Loop Simulated Nuclear Power Plant Thermal Dispatch System Demonstration and Evaluation Study

An Idaho National Laboratory research team performed a human-in-the-loop study with two formerly licensed retired operators to evaluate thermal power dispatch operations supported by the modified GSE Systems GPWR plant simulator. Virtual representations of the analog control panels were presented on touch-screen bays configured to mimic the control room layout in the newly renovated Human Systems Simulation Laboratory. The operators performed 15 scenarios covering normal evolutions to transition the plant from full turbine operation to joint turbine and thermal power dispatch operations in addition to transient response scenarios induced with simulated faults to evaluate the impact of thermal power dispatch system on operator and plant responses. A prototype human-system interface (HSI) was developed and displayed in tandem with the virtual analog panels to support the operators executing the procedurally drive evolutions and transient responses. An interdisciplinary team of operations experts, nuclear engineers, and human factors experts observed the operators performing the scenarios to evaluate the operations. Only a preliminary analysis of the results has been performed. The full analysis will be shared in a milestone report scheduled for release at the end of September; however, some high-level conclusions were readily available. Two high level findings for the system design were captured in the study. The manual control supported by the HSI to transition from standard operations to thermal power dispatch operation imposed a considerable amount of workload on the operators due to tedious manual valve manipulations and system monitoring required to verify their intended effect. An additional operator would be required in the control room to support the daily evolution. Automatic control for the transition was deemed a requirement for plant adoption without imposing additional staffing costs. The second finding was the necessity for an automatic thermal power dispatch system trip isolation function linked to a turbine and reactor trip signal. The operators completed scenarios with automatic isolation functionality and manually required actuation of the thermal power dispatch system. The operator response was sufficiently slower in the manual trip condition, such that operators were unable to manually actuate key post trip safety functions an indicate a degraded control capability that should be avoided. With an automatic trip signal little to no impact of the thermal power dispatch system was identified on the primary plant response and therefore the system could be readily and safely adopted. Together these two findings represent the need to support the adoption of thermal power dispatch capability into existing operations by leveraging automation to augment any additional operator tasking required to control and monitor an additional system beyond existing operations.

99 GENERAL AND MISCELLANEOUS↗

NREL Stratus - Enabling Workflows to Fuse Data Streams, Modeling, Simulation, and Machine Learning

Integrating cloud services into advanced computing facilities provides significant new capabilities over focusing solely on traditional high performance computing (HPC) workloads. This brings complementary capabilities as well as enabling new focused roles for HPC. They are especially potent for workflows that fuse data streams, modeling and simulation ('modsim') and machine learning. A key challenge to adopting a hybrid edge-cloud-HPC model is to align optimal capability, data, and user intent on the right resources for each step in a workflow.?The NREL Stratus service provides a basis for this: Stratus layers capabilities needed to make?cloud services accessible to a lab-based scientific community on commercial offerings, and; currently supports upwards of 200 projects ranging from IOT integration to traditional modeling and simulation. This provides a real-world inventory of scientific workflow elements. A growing knowledge base enables placing these elements appropriately between the edge, cloud, and traditional HPC. This paper outlines a vision via reference architecture and the application of that architecture in a typical workflow highlighting multiple components: sensor data intake, cleaning and transforming (edge/cloud suitable); generation of synthetic data through modsim, computationally heavy ML training and hyperparameter optimization (HPC suitable), and; inference and deployment (cloud ideal). Every step in such a workflow involves a cost-benefit analysis regarding the data movement, computational efficiency, availability, latency, and resource capabilities. The reference architecture and examples outlined allow for understanding new opportunities in the context of emerging workflows that combine IOT, cloud, and HPC to bolster scientific productivity.

AI↗

28 NREL Stratus - Enabling Workflows to Fuse Data Streams, Modeling, Simulation, and Machine Learning: Preprint

Integrating cloud services into advanced computing facilities provides significant new capabilities over focusing solely on traditional high performance computing (HPC) workloads. This brings complementary capabilities as well as enabling new focused roles for HPC. They are especially potent for workflows that fuse data streams, modeling and simulation ('modsim') and machine learning. A key challenge to adopting a hybrid edge-cloud-HPC model is to align optimal capability, data, and user intent on the right resources for each step in a workflow.?The NREL Stratus service provides a basis for this: Stratus layers capabilities needed to make?cloud services accessible to a lab-based scientific community on commercial offerings, and; currently supports upwards of 200 projects ranging from IOT integration to traditional modeling and simulation. This provides a real-world inventory of scientific workflow elements. A growing knowledge base enables placing these elements appropriately between the edge, cloud, and traditional HPC. This paper outlines a vision via reference architecture and the application of that architecture in a typical workflow highlighting multiple components: sensor data intake, cleaning and transforming (edge/cloud suitable); generation of synthetic data through modsim, computationally heavy ML training and hyperparameter optimization (HPC suitable), and; inference and deployment (cloud ideal). Every step in such a workflow involves a cost-benefit analysis regarding the data movement, computational efficiency, availability, latency, and resource capabilities. The reference architecture and examples outlined allow for understanding new opportunities in the context of emerging workflows that combine IOT, cloud, and HPC to bolster scientific productivity.

AI↗

Report on Internship Project: Risk-Informed Cost Reduction in Nuclear Power Plants

Several Nuclear Power Plants (NPPs) in USA are facing premature closures due to financial difficulties. A significant contributing factor in the NPP electricity generation cost is Operational and Maintenance (O&M) cost. This research aims to reduce O&M cost by developing a framework which creates flexibility in NPP maintenance program. This framework relies on FLEX portable equipment to reduce NPP risk and to substitute and/or backup fixed equipment in different NPP operation modes. Probabilistic Risk Assessment (PRA) methodology is used to quantify the change of plant risk to justify the framework's applicability. The framework enables plant managers to shift maintenance workload from refueling outage period into online maintenance programs. As a result, it is expected to increase NPP capacity factor while also reducing the risk of unexpected outage. These results are expected to increase NPP electricity generation revenue and decrease O&M costs while complying to existing nuclear safety regulations.

22 GENERAL STUDIES OF NUCLEAR REACTORS↗

Union: A Unified HW-SW Co-Design Ecosystem in MLIR for Evaluating Tensor Operationson Spatial Accelerators

To meet the extreme compute demands for deep learning across commercial and scientific applications, dataflow accelerators are becoming increasingly popular. While these“domain-specific” accelerators are not fully programmable like CPUs and GPUs, they retain varying levels of flexibility with respect to data orchestration, i.e., dataflow and tiling optimizations to enhance efficiency. There are several challenges when designing new algorithms and mapping approaches to execute the algorithms for a target problem on new hardware. Previous works have addressed these challenges individually. To address this challenge as a whole, in this work, we present an HW-SW co-design ecosystem for spatial accelerators called Union within the popular MLIR compiler infrastructure. Our framework allows exploring different algorithms and their mappings on several accelerator cost models. Union also includes a plug-and-play library of accelerator cost models and mappers which can easily be extended. The algorithms and accelerator cost models are connected via a novel mapping abstraction that captures the map space of spatial accelerators which can be systematically pruned based on constraints from the hardware, workload, and mapper. We demonstrate the value of Union for the community with several case studies which examine offloading different tensor operations (CONV/GEMM/Tensor Contraction) on diverse accelerator architectures using different mapping schemes.

Jeong, Geonhwa↗

The HUNTER Dynamic Human Reliability Analysis Tool: Development of a Module for Performance Shaping Factors

Human Unimodel for Nuclear Technology to Enhance Reliability (HUNTER) is a framework to support dynamic human reliability analysis (HRA) in communication with a variety of methods and tools. In this paper, how we have developed one of the HUNTER modules, the individual module for evaluating performance shaping factors (PSFs), is introduced. The PSF refers to any factor that influences human performance such as workload or complexity. It has been used for highlighting human errors and adjusting the error probabilities in the existing HRA. We consider the eight PSFs suggested in the Standardized Plant Analysis Risk-HRA (SPAR-H) method, which is the representative HRA method widely used in the nuclear field. To support our dynamic modeling using the eight SPAR-H PSFs, we reviewed human performance literature and developed data-based mathematical models to rate and quantify PSFs in the context of dynamic HRA. We also design the individual module composed of the two functions: 1) the PSF qualification function that automatically or manually evaluates a PSF level, and 2) the PSF quantification function that dynamically or statically determines the PSF multiplier values and integrate them to adjust human error probabilities (HEPs). How each function works with the SPAR-H PSFs and how the PSFs adjust the HEPs are investigated through literature and discussed in this paper.

99 GENERAL AND MISCELLANEOUS↗

An Experimental Investigation of Students? Learning Effects When Using a Simplified Nuclear Simulator

This study focuses on investigating students' learning effects and performance trends over a certain period using the Rancor Microworld Simulator. Specifically, it aims to determine the training required to collect HRA data from non-experts (i.e., students) using Rancor Microworld and the differences in human performance measures between students and professional operators. A longitudinal experiment is conducted with sixteen undergraduate students, using four Rancor Microworld scenarios in each of the four experiment trials. The study considers four human performance measurements workload, situation awareness, time, and error. Finally, the trend of students' performance is compared with operator data collected from the previous experiment. Overall, this research complements previous studies by providing insights into how much training is required to collect HRA data from non-experts and the differences in human performance measures between students and professional operators.

99 GENERAL AND MISCELLANEOUS↗

An Experimental Investigation of Students’ Learning Effects When Using a Simplified Nuclear Simulator

Securing enough data has been a main challenge in human reliability analysis (HRA). Many researchers and institutes have made a lot of efforts for collecting HRA data to produce reasonable human error probabilities (HEPs) as well as reduce the uncertainty of HRA quantification. Representatively, U.S. Nuclear Regulatory Commission (U.S. NRC), Korea Atomic Energy Research Institute (KAERI) and Idaho National Laboratory (INL) have led lots of empirical research regarding the HRA data collection. The U.S. NRC and KAERI have mainly carried out full-scope simulator research collecting HRA data through experiments using full-scope simulators with actual operators. In contrast, INL has experimentally collected the data using simplified simulators and student operators. INL has proposed the Simplified Human Error Experimental Program (SHEEP) framework to complement full-scope data collection efforts by suggesting a way to infer full-scope data based on experimental data collected from students operating simplified simulators, specifically the Rancor Microworld Simulator (Rancor) and the Compact Nuclear Simulator (CNS). The aim of the SHEEP framework is to lower the entry point for collecting useful HRA data by securing large sample sizes at a reasonable amount of cost and labor while also guaranteeing a high degree of freedom when designing experiments. The authors’ previous research investigated whether data collected from the SHEEP framework could support a representative full-scope study. Besides, human performance differences between professional and student operators when using Rancor and CNS have been analyzed to understand the lack of fidelity of the simplified simulators and student operators within the SHEEP study. As a follow up research, this study experimentally investigates students’ learning effects and the performance trends over a certain period when using Rancor. This study aims to find out 1) how much training or education is required to collect HRA data from non-experts (i.e., students) when using Rancor and 2) how much differences there are in human performance measures between students and professional operators. In this study, a longitudinal experiment is developed. The four experiment trials with two weeks interval are carried out for sixteen undergraduate students majoring nuclear engineering at Chosun University. Totally four scenarios randomly selected from ten Rancor scenarios are used in each experiment trial. Four human performance measurements (i.e., workload, situation awareness, time and error) are considered in the experiment. Lastly, the trend of students’ performance is compared with operator data having been collected from the previous experiment.

99 GENERAL AND MISCELLANEOUS↗

Comparison of Human Performance between Operators and Students Using Rancor Microworld Simulator: A Preliminary Result

This paper describes an experiment using the Rancor Microworld simulator to probe into differences in human performance between students and actual operators. Six human performance measures – 1) time, 2) error, 3) workload, 4) situational awareness, 5) attention patterns, and 6) number of manipulations – are measured from experiments. The experiment has been conducted for 11 operators and 20 students so far. More experiments and analyses will be continued. Further experimental studies using Rancor Microworld are expected to be able to collect additional data that may be missing from full-scope studies and contribute to items that can support human reliability data.

99 GENERAL AND MISCELLANEOUS↗

Optimized self-designing key-value storage engine

Embodiments of the invention utilize an optimized key-value storage engine to strike the optimal balance between cloud-cost and performance and supports queries, including updates, lookups, range queries, inserts, and read-modify-writes. Cloud cost is manifested in purchasing both storage and processing resources. The improved approach has the ability to self-design and instantiate holistic configurations given a workload, a cloud budget, and optionally performance goals and a set of Service Level Agreement (SLA) specifications. A configuration reflects an optimized storage engine design in terms of, for example, the individual data structures design (in-memory and on-disk) in the engine as well as their algorithms and interactions, a cloud provider, and the exact virtual machines to be used.

Idreos, Stratos↗

The HUNTER Dynamic Human Reliability Analysis Tool: Development of a Module for Performance Shaping Factors

The Human Unimodel for Nuclear Technology to Enhance Reliability (HUNTER) is a framework to support dynamic human reliability analysis (HRA) in communication with a variety of methods and tools. This paper explores how we developed one of the HUNTER modules: namely, the Individual module for evaluating performance shaping factors (PSFs). A PSF is any factor that influences human performance (e.g., workload or complexity). In the existing HRA, they are used to highlight human errors and adjust error probabilities. We consider the eight PSFs suggested in the Standardized Plant Analysis Risk-HRA (SPAR-H) method, a representative HRA method widely used in the nuclear field. To support our dynamic modeling using the eight SPAR-H PSFs, we reviewed the human performance literature and developed data-based mathematical models to rate and quantify PSFs in the context of dynamic HRA. We also designed the Individual module to consist of two functions: (1) the PSF qualification function for automatically or manually evaluating PSF levels, and (2) the PSF quantification function for dynamically or statically determining PSF multiplier values and integrating them to adjust human error probabilities (HEPs). How each function works in regard to the SPAR-H PSFs, and how the PSFs serve to adjust the HEPs, were investigated via literature review and are discussed in this paper.

99 GENERAL AND MISCELLANEOUS↗

symPACK: A GPU-Capable Fan-Out Sparse Cholesky Solver

Sparse symmetric positive definite systems of equations are ubiquitous in scientific workloads and applications. Parallel sparse Cholesky factorization is the method of choice for solving such linear systems. Therefore, the development of parallel sparse Cholesky codes that can efficiently run on today’s large-scale heterogeneous distributed-memory platforms is of vital importance. Modern supercomputers offer nodes that contain a mix of CPUs and GPUs. To fully utilize the computing power of these nodes, scientific codes must be adapted to offload expensive computations to GPUs. We present symPACK, a GPU-capable parallel sparse Cholesky solver that uses one-sided communication primitives and remote procedure calls provided by the UPC++ library. We also utilize the UPC++ "memory kinds" feature to enable efficient communication of GPU-resident data. We show that on a number of large problems, symPACK outperforms comparable state-of-the-art GPU-capable Cholesky factorization codes by up to 14x on the NERSC Perlmutter supercomputer.

Bellavita, Julian↗

Experiences readying applications for Exascale

The advent of Exascale computing invites an assessment of existing best practices for developing application readiness on the world's largest supercomputers. This work details observations from the last four years in preparing scientific applications to run on the Oak Ridge Leadership Computing Facility's (OLCF) Frontier system. This paper addresses a range of topics in software including programmability, tuning, and portability considerations that are key to moving applications from existing systems to future installations. A set of representative workloads provides case studies for general system and software testing. We evaluate the use of early access systems for development across several generations of hardware. Finally, we discuss how best practices were identified and disseminated to the community through a wide range of activities including user-guides and trainings. We conclude with recommendations for ensuring application readiness on future leadership computing systems.

Gottiparthi, Kalyan↗