Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “cluster computer”

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

Automatic mesh adaptivity for CADIS and FW-CADIS neutronics modeling of difficult shielding problems

The CADIS and FW-CADIS hybrid Monte Carlo/deterministic techniques dramatically increase the efficiency of neutronics modeling, but their use in the accurate design analysis of very large and geometrically complex nuclear systems has been limited by the large number of processors and memory requirements for their preliminary deterministic calculations and final Monte Carlo calculation. Three mesh adaptivity algorithms were developed to reduce the memory requirements of CADIS and FW-CADIS without sacrificing their efficiency improvement. First, a macro-material approach enhances the fidelity of the deterministic models without changing the mesh. Second, a deterministic mesh refinement algorithm generates meshes that capture as much geometric detail as possible without exceeding a specified maximum number of mesh elements. Finally, a weight window coarsening algorithm de-couples the weight window mesh and energy bins from the mesh and energy group structure of the deterministic calculations in order to remove the memory constraint of the weight window map from the deterministic mesh resolution. The three algorithms were used to enhance an FW-CADIS calculation of the prompt dose rate throughout the ITER experimental facility. Using these algorithms resulted in a 23.3% increase in the number of mesh tally elements in which the dose rates were calculated in a 10-day Monte Carlo calculation and, additionally, increased the efficiency of the Monte Carlo simulation by a factor of at least 3.4. The three algorithms enabled this difficult calculation to be accurately solved using an FW-CADIS simulation on a regular computer cluster, obviating the need for a world-class super computer. (authors)

70 PLASMA PHYSICS AND FUSION TECHNOLOGY↗

Containers on Switches: A Cluster School Experience

Network switches, such as those from Arista and Mellanox, often have underutilized computational resources in the form of built-in processors and memory. By leveraging these untapped resources, we can optimize functionality and efficiency of computational cluster networks. Our research focuses on deploying containers directly onto these switches to execute various auxiliary tasks ranging from metric logging to system-wide management via post-boot configuration. By doing so, we can significantly enchance the capabilities of the cluster without the need for additional dedicated hardware. Our research involved five distinct scenarios where switch utilization could have a profound impact on HPC Clusters: run cloud-init services via link-local connection; configuring a Telegraf container to export metrics; deploying a caching proxy; creating a reconfigurable IPv6 DHCP/DNS provider for VLAN; and implementing a client detection with Magellan discovery. These scenarios were containerized with podman and docker, and tested both physically on the switch virtually on a QEMU VM both running SONiC OS. Testing and findings indicate that network switches can indeed be used for these scenarios. They offer a wide range of possibilities beyond these applications. They run as expected as containers on the switches, and although there were some minor issues, work-arounds were implemented. Overall, this is a positive result that can be further explored with more scenarios.

97 MATHEMATICS AND COMPUTING↗

Seismology in the Cloud: A New Streaming Workflow

Data-intensive research in seismology is experiencing a recent boom, driven in part by large volumes of available data and advances in the growing field of data science. However, there are significant barriers to processing large data volumes, such as long retrieval times from data repositories, complex data management, and limited computational resources. New tools and platforms have reduced the barriers to entry for scientific cluster computing, including the maturation of the commercial cloud as an accessible instrument for research. Here, we build a customized research cluster in the cloud to test a new workflow for large-scale seismic analysis, in which data are processed as a stream (retrieved on-the-fly and acted upon without storing), with data from the Incorporated Research Institutions for Seismology Data Management Center. We use this workflow to deploy a spectral peak detection algorithm over 5.6 TB of compressed continuous seismic data from 2074 stations of the USArray Transportable Array EarthScope network. Using a 50-node cluster in the cloud, we completed the noise survey in 80 hr, with an average data throughput of 1.7 GB per minute. By varying cluster sizes, we find the scaling of our analysis to be sublinear, due to a combination of algorithmic limitations and data center response times. The cloud-based streaming workflow represents an order-of-magnitude increase in acquisition and processing speed compared to a traditional download-store-process workflow, and offers the additional benefits of employing a flexible, accessible, and widely used computing architecture. It is limited, however, due to its reliance on Internet transfer speeds and data center service capacity, and may not work well for repeated analyses or those for which even higher data throughputs are needed. These research applications will require a new class of cloud-native approaches in which both data and analysis are in the cloud.

58 GEOSCIENCES↗

A cloud computing approach to massively-parallel distributed stochastic optimization with application to geomechanical reservoir characterization

Many search strategies used to optimize parameters during hydrologic model calibration are inherently limited in that simulations must be performed sequentially in order for the algorithm to converge. As a result, the overall time required to reach a solution increases linearly with the number of steps necessary. In contrast, sampling strategies that attempt to approximate the posterior distribution of the parameters do not require the model simulations to be run in sequence and therefore invite the use of a massively parallel approach across heterogeneous computing environments. We apply a cloud computing methodology to this problem where a head node running on Amazon Web Services continuously produces candidate models that are distributed to arbitrary compute environments to run the simulation, in our case Clemson’s Palmetto cluster computer. Thus the time required to obtain a sufficient number of samples to produce an adequate approximation of the posterior distribution of the parameters is limited by the number of available compute sites. Monte Carlo sampling of the posterior space would be duplicative and inefficient, thus we combine Latin hypercube methods with genetic algorithms to ensure that no large regions of the parameter space are left unexplored, and that high probability regions are investigated thoroughly. We independently assess the fit to different types of data sets without the need to construct a single objective function. Sets of Pareto optimal models are particularly valuable when the quality and information content of data is poorly known or conceptual model errors exist in our simulation. We illustrate our approach for a problem where in situ strain data are used to calibrate the parameters of an oil reservoir in the Avant oil field in Oklahoma. We use well injection and shut-in to stress a deep (500m) confined formation, and observe the geomechanical response using borehole strainmeters installed in the shallow subsurface (30m). These data were then used to calibrate the poroelastic properties of the reservoir and caprock along with the geometry of an elliptic channel associated with the formation. Using our approach we are able to identify issues with model error as well as substantial tradeoffs that occur within the parameters.

inversion, stochastic, geomechanics, poroelastic, ↗

Elastic Resource Management for Deep Learning Applications in a Container Cluster

The increasing demand for learning from massive datasets is restructuring our economy. Effective learning, however, involves nontrivial computing resources. Most businesses utilize commercial infrastructure providers (e.g., AWS) to host their computing clusters in the cloud, where various jobs compete for available resources. While cloud resource management is a fruitful research field that has made many advances in production, such as Kubernetes and YARN, few efforts have been invested to further optimize the system performance, especially for deep learning (DL) training jobs in a container cluster. This work introduces FlowCon, a system that is able to monitor the individual evaluation functions of DL jobs at runtime, and thus to make placement decisions on resource allocations elastically. Here, we present a detailed design and implementation of FlowCon and conduct intensive experiments over various DL models. The results demonstrate that FlowCon significantly improves DL job completion time and resource utilization efficiency, compared to default systems. According to the results, FlowCon is able to improve the completion time by up to 68.8% and meanwhile, reduce the makespan by 18.0%, in the presence of various DL job workloads.

97 MATHEMATICS AND COMPUTING↗

Predicting runtime and resource utilization of jobs on integrated cloud and HPC systems

Recent advances in virtualization technologies used in cloud computing offer performance that closely approaches bare-metal levels. Combined with specialized instance types and high-speed networking services for cluster computing, cloud platforms have become a compelling option for high-performance computing (HPC). However, most current batch job schedulers in HPC systems are designed for homogeneous clusters and make decisions based on limited information about jobs and system status. Scientists typically submit computational jobs to these schedulers with a requested runtime that is often over- or under-estimated. More accurate runtime predictions can help schedulers make better decisions and reduce job turnaround times. Here, they can also support decisions about migrating jobs to the cloud to avoid long queue wait times in HPC systems.

97 MATHEMATICS AND COMPUTING↗

DICER: Data Intensive Computing Environment and Runtime for Evaluating Unprecedented Scale of Geospatial-Temporal Human Mobility Data

With the significant increase in sources and volume of human mobility data through commercial data vendors as well as microsimulation of cities, the scale of geospatial-temporal data to analyze and assess for mobility characterization has grown to the level of Big Data. There are mobility related commercial organizations deploying scalable computing, but often the system architecture, workflow, and intermediate processing components are not fully disclosed in relevant scope. Current research literature has a notable lack of studies demonstrating architectures and workflows for human mobility analytics that are implemented on a TeraByte scale of geospatial-temporal data. In this context, this paper presents a hyperscale-level system solution named DICER (Data Intensive Computing Environment and Runtime) for processing and analytics of geospatial-temporal data at big data scale. Although the cluster computing architecture of DICER with Apache Spark job running on Kubernetes cluster is not new, there are innovations in the workflow, hierarchical processing logic, and a wide range of intermediate preprocessing and mobility metrics calculation. We have performed case studies to validate the effectiveness of DICER system solution by performing detailed analytics and assessment of human mobility microsimulation output at three different scopes and scale, including a usecase with 16.97 TeraByte and 259.2 Billion rows of data. In addition, we have presented another case study of utilizing DICER to perform the same mobility processing and comparative analytics on large-scale commercially available geospatial-temporal data. All these case studies validate the efficiency and usefulness of DICER in computing population mobility characteristics from geospatial-temporal trajectory data at an unprecedented scale (not only just data volume, but also combination of: number of user entities, temporal frequency, spatial resolution, data duration).

De, Debraj↗

Scalable Static and Dynamic Community Detection Using Grappolo

Graph clustering, popularly known as community detection, is a fundamental kernel for several applications of relevance to the Defense Advanced Research Projects Agency’s (DARPA) Hierarchical Identify Verify Exploit (HIVE) Pro- gram. Clusters or communities represent natural divisions within a network that are densely connected within a cluster and sparsely connected to the rest of the network. The need to compute clustering on large scale data necessitates the development of efficient algorithms that can exploit modern architectures that are fundamentally parallel in nature. How- ever, due to their irregular and inherently sequential nature, many of the current algorithms for community detection are challenging to parallelize. In response to the HIVE Graph Challenge, we present several parallelization heuristics for fast community detection using the Louvain method as the serial template. We implement all the heuristics in a software library called Grappolo. Using the inputs from the HIVE Challenge, we demonstrate superior performance and high quality solutions based on four parallelization heuristics. We use Grappolo on static graphs as the first step towards community detection on streaming graphs.

Community Detection↗

Reinterpreting the infrared spectrum of H + HCN: Methylene amidogen radical and its coproducts

The methylene amidogen radical (H 2 CN) plays a role in high-energy material combustion and extraterresterial atmospheres. Recent theoretical work has struggled to match experimental assignments for its CN and antisymmetric CH 2 stretching frequencies (ν 2 and ν 5 ), which were reported to occur at 1725 and 3103 cm –1 . Herein, we compute the vibrational energy levels of this molecule by extrapolating quadruples-level coupled- cluster theory to the complete basis limit and adding corrections for vibrational anharmonicity. This level of theory predicts that ν 2 and ν 5 should occur at 1646 and 2892 cm –1 , at odds with the experimental assignments. To investigate the possibility of defects in our theoretical treatment, we analyze the sensitivity of our approach to each of its contributing approximations. Our analysis suggests that the observed deviation from experiment is too large to be explained as an accumulation of errors, leading us to conclude that these transitions were misassigned. To help resolve this discrepancy, we investigate possible byproducts of the H + HCN reaction, which was the source of H 2 CN in the original experiment. In particular, we predict vibrational spectra for cis -HCNH, trans -HCNH, and H 2 CNH using high-level coupled-cluster computations. Based on these results, we reassign the transition at 1725 cm –1 to ν 3 of trans-HCNH, yielding excellent agreement. Supporting this identification, we assign a known contaminant peak at 886 cm –1 to ν 5 of the same conformer. Our computations suggest that the peak observed at 3103 cm –1 , however, does not belong to any of the aforementioned species. To facilitate further investigation, we use structure and bonding arguments to narrow the range of possible candidates. Furthermore, these arguments lead us to tentatively put forth formaldazine [(H 2 CN) 2 ] as a suggestion for further study, which we support with additional computations.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗

Massively Parallel Latent Semantic Analyzes using a Graphics Processing Unit

Latent Semantic Indexing (LSA) aims to reduce the dimensions of large Term-Document datasets using Singular Value Decomposition. However, with the ever expanding size of data sets, current implementations are not fast enough to quickly and easily compute the results on a standard PC. The Graphics Processing Unit (GPU) can solve some highly parallel problems much faster than the traditional sequential processor (CPU). Thus, a deployable system using a GPU to speedup large-scale LSA processes would be a much more effective choice (in terms of cost/performance ratio) than using a computer cluster. Due to the GPU s application-specific architecture, harnessing the GPU s computational prowess for LSA is a great challenge. We present a parallel LSA implementation on the GPU, using NVIDIA Compute Unified Device Architecture and Compute Unified Basic Linear Algebra Subprograms. The performance of this implementation is compared to traditional LSA implementation on CPU using an optimized Basic Linear Algebra Subprograms library. After implementation, we discovered that the GPU version of the algorithm was twice as fast for large matrices (1000x1000 and above) that had dimensions not divisible by 16. For large matrices that did have dimensions divisible by 16, the GPU algorithm ran five to six times faster than the CPU version. The large variation is due to architectural benefits the GPU has for matrices divisible by 16. It should be noted that the overall speeds for the CPU version did not vary from relative normal when the matrix dimensions were divisible by 16. Further research is needed in order to produce a fully implementable version of LSA. With that in mind, the research we presented shows that the GPU is a viable option for increasing the speed of LSA, in terms of cost/performance ratio.

99 GENERAL AND MISCELLANEOUS↗

On the Efficient Evaluation of the Exchange Correlation Potential on Graphics Processing Unit Clusters

The predominance of Kohn–Sham density functional theory (KS-DFT) for the theoretical treatment of large experimentally relevant systems in molecular chemistry and materials science relies primarily on the existence of efficient software implementations which are capable of leveraging the latest advances in modern high-performance computing (HPC). With recent trends in HPC leading toward increasing reliance on heterogeneous accelerator-based architectures such as graphics processing units (GPU), existing code bases must embrace these architectural advances to maintain the high levels of performance that have come to be expected for these methods. In this work, we purpose a three-level parallelism scheme for the distributed numerical integration of the exchange-correlation (XC) potential in the Gaussian basis set discretization of the Kohn–Sham equations on large computing clusters consisting of multiple GPUs per compute node. In addition, we purpose and demonstrate the efficacy of the use of batched kernels, including batched level-3 BLAS operations, in achieving high levels of performance on the GPU. We demonstrate the performance and scalability of the implementation of the purposed method in the NWChemEx software package by comparing to the existing scalable CPU XC integration in NWChem.

97 MATHEMATICS AND COMPUTING↗

Further Automate Planned Cluster Maintenance to Minimize System Downtime during Maintenance Windows

This report documents the integration and testing of the automated update process of compute clusters in LC to minimize impact to user productivity. Description: A set of scripts will be written and deployed to further standardize cluster maintenance activities and minimize downtime during planned maintenance windows. Completion Criteria: When the scripts have been deployed and used during planned maintenance windows and a timing comparison is completed between the existing process and the new more automated process, this milestone is complete. This milestone was completed on Aug 23, 2016 on the new CTS1 cluster called Jade when a request to upgrade the version of TOSS 3 was initiated while SWL jobs and normal user jobs were running. Jobs that were running when the update to the system began continued to run to completion. New jobs on the cluster started on the new release of TOSS 3. No system administrator action was required. Current update procedures in TOSS 2 begin by killing all users jobs. Then all diskfull nodes are updated, which can take a few hours. Only after the updates are applied are all nodes are rebooted, and then finally put back into service. A system administrator is required for all steps. In terms of human time spent during a cluster OS update, the TOSS 3 automated procedure on Jade took 0 FTE hours. Doing the same update without the Toss Update Tool would have required 4 FTE hours.

97 MATHEMATICS AND COMPUTING↗

Scalable deep text comprehension for Cancer surveillance on high-performance computing

Background: Deep Learning (DL) has advanced the state-of-the-art capabilities in bioinformatics applications which has resulted in trends of increasingly sophisticated and computationally demanding models trained by larger and larger data sets. This vastly increased computational demand challenges the feasibility of conducting cutting-edge research. One solution is to distribute the vast computational workload across multiple computing cluster nodes with data parallelism algorithms. In this study, we used a High-Performance Computing environment and implemented the Downpour Stochastic Gradient Descent algorithm for data parallelism to train a Convolutional Neural Network (CNN) for the natural language processing task of information extraction from a massive dataset of cancer pathology reports. We evaluated the scalability improvements using data parallelism training and the Titan supercomputer at Oak Ridge Leadership Computing Facility. To evaluate scalability, we used different numbers of worker nodes and performed a set of experiments comparing the effects of different training batch sizes and optimizer functions.Results: We found that Adadelta would consistently converge at a lower validation loss, though requiring over twice as many training epochs as the fastest converging optimizer, RMSProp. The Adam optimizer consistently achieved a close 2nd place minimum validation loss significantly faster; using a batch size of 16 and 32 allowed the network to converge in only 4.5 training epochs.Conclusions: We demonstrated that the networked training process is scalable across multiple compute nodes communicating with message passing interface while achieving higher classification accuracy compared to a traditional machine learning algorithm.

Qiu, John X.↗

Parallel Latent Semantic Analysis using a Graphics Processing Unit

Latent Semantic Analysis (LSA) can be used to reduce the dimensions of large Term-Document datasets using Singular Value Decomposition. However, with the ever expanding size of data sets, current implementations are not fast enough to quickly and easily compute the results on a standard PC. The Graphics Processing Unit (GPU) can solve some highly parallel problems much faster than the traditional sequential processor (CPU). Thus, a deployable system using a GPU to speedup large-scale LSA processes would be a much more effective choice (in terms of cost/performance ratio) than using a computer cluster. In this paper, we presented a parallel LSA implementation on the GPU, using NVIDIA Compute Unified Device Architecture (CUDA) and Compute Unified Basic Linear Algebra Subprograms (CUBLAS). The performance of this implementation is compared to traditional LSA implementation on CPU using an optimized Basic Linear Algebra Subprograms library. For large matrices that have dimensions divisible by 16, the GPU algorithm ran five to six times faster than the CPU version.

59 BASIC BIOLOGICAL SCIENCES↗

DOE site remedial management optimization utilizing on the cloud computing systems - 15676

Environmental cleanups at complex sites necessarily address many competing demands and answer questions such as: 'What is the best way and in a sustainable manner to minimize risk to human health and the environment during active restoration efforts and/or long term legacy management, while incorporating parameter uncertainty, management considerations, and stakeholder concerns into the process?' Physics Based Management Optimization (PBMO{sup TM}) provides a means to answer such questions. 'Physics Based' indicates incorporation of numerically computed groundwater flow and transport processes into the analysis. This enables optimal remedy design based on comprehensive mass removal/destruction metrics and optimal monitoring strategies. 'Management Optimization' indicates the ability to incorporate objective functions (e.g., management constraints, sustainability considerations) into the evaluation. PBMO{sup TM} components have been successfully implemented for optimal plume delineation (points of compliance) and monitoring, for optimal system design and to optimize existing active treatment systems in the US at multiple DOE, DOD and industrial sites since the mid 1990's. PBMO{sup TM} has been recently extended and implemented on multi-core, multi-CPU grid computing systems including computational clusters, local area networks and the Cloud. This extension allows for increased flexibility and adaptation of computational resources which can now support optimal remedy design and long term management strategies for complex soil and groundwater sites. In the PBMO{sup TM} -Grid case study, we evaluate flow and transport for a typical optimization search for a realistic project against sequential optimization. PBMO{sup TM}-Grid reduces the CPU time required to solve the optimization problem by more than a factor of 14 X compared to sequential optimization. PBMO{sup TM}-Grid can be deployed for deterministic or stochastic optimization and formally computes design risk and predicted degree of remedial action success. The numerical models used can consist of a single 'mega' model (aka a monolith), or a mixture of models and other calculations or expert systems. PBMO{sup T}M-Grid integrates the computational models in a seamless manner, and models run on the Cloud can be implemented in Windows or Linux platforms. (authors)

12 MANAGEMENT OF RADIOACTIVE AND NON-RADIOACTIVE W↗

A Versatile Simulated Data Transport Layer for in Situ Workflows Performance Evaluation

In situ processing does not only allow scientific applications to face the explosion in data volume and velocity but also to address the time constraints of many simulation-analysis workflows by providing scientists with early insights about their applications at runtime. Multiple frameworks implement the concept of a data transport layer (DTL) to enable such in situ workflows. These tools are very versatile, directly or indirectly access the data generated on the same node, another node of the same compute cluster, or a completely distinct node, and allow data publishers and subscribers to run on the same computing resources or not. This versatility puts on researchers the onus of taking key decisions related to resource allocation and how to transport data to ensure the most efficient execution of their in situ workflows. However, domain scientists and workflow practitioners lack the appropriate tools to assess the respective performance of particular design and deployment options. In this paper we introduce a versatile simulated DTL designed to provide researchers with insights on the respective performance of different execution scenarios of in situ workflows. This open-source, standalone library builds on the SimGrid toolkit and can be linked to any SimGrid-based simulator. It facilitates the evaluation of the performance behavior, at scale, of different data transport configurations and the study of the effects of resource allocation strategies. We demonstrate the scalability, versatility, and accuracy of this simulated DTL by reproducing the execution of two synthetic benchmarks and of a real-world in situ workflow composed of an MPI application and a parallel data analysis. Results of simulations run on a single core show that the proposed library can simulate the interactions of tens of thousands of simulated processes deployed on two interconnected commodity clusters in a few seconds, and the execution by a thousand simulated processes of an in situ workflow in less than three minutes.

Suter, Fred [ORNL] (ORCID:0000000319021955)↗

Energetics and Structure of Ag–Water Clusters Formed in Mordenite

Zeolite-supported Ag 0 clusters have broad applications from catalysis to medicine, necessitating a mechanistic understanding of the formation of Ag 0 clusters in situ. Density functional theory (DFT) simulations have been performed on silver, water, and silver–water clusters in silica mordenite (Si-MOR), to identify the role of the confinement on the structure and energetics of Ag 0 cluster formation. The most favorable binding energy in the 12-membered ring (MR) pore of the Si-MOR is a 10–15-atom Ag 0 cluster. Computational pair distribution function (PDF) data indicates that the Ag 0 and Ag 0 –H 2 O clusters formed in vacuum versus in Si-MOR exhibit structural differences. Additionally, when the Ag 0 cluster is confined, the density decreases and the surface area increases, hypothesized to be due to the limiting geometry of the 12-MR main channel. An energetic drive toward formation of larger Ag 0 clusters was also identified, with hydrated silver atoms generating higher energy structures. In conclusion, this work identifies mechanistic and structural insight into the role of nanoconfinement on formation of Ag 0 clusters in mordenite.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗