Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “file system”

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 235 records · Page 13
All sourcesNASA NTRSDOE OSTI
DOE OSTI2020

Adaptive multi-level checkpointing

In some examples, with respect to adaptive multi-level checkpointing, a transfer parameter associated with transfer of checkpoint data from a node-local storage to a parallel file system may be ascertained for the checkpoint data stored in the node-local storage. The transfer parameter may be compared to a specified transfer parameter threshold. A determination may be made, based on the comparison of the transfer parameter to the specified transfer parameter threshold, as to whether to transfer the checkpoint data from the node-local storage to the parallel file system.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2012

Petascale Data Storage Institute (PDSI) Final Report

Petascale computing infrastructures for scientific discovery make petascale demands on information storage capacity, performance, concurrency, reliability, availability, and manageability. The Petascale Data Storage Institute focuses on the data storage problems found in petascale scientific computing environments, with special attention to community issues such as interoperability, community buy-in, and shared tools. The Petascale Data Storage Institute is a collaboration between researchers at Carnegie Mellon University, National Energy Research Scientific Computing Center, Pacific Northwest National Laboratory, Oak Ridge National Laboratory, Sandia National Laboratory, Los Alamos National Laboratory, University of Michigan, and the University of California at Santa Cruz. Because the Institute focuses on low level files systems and storage systems, its role in improving SciDAC systems was one of supporting application middleware such as data management and system-level performance tuning.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2012

GlusterFS One Storage Server to Rule Them All

GlusterFS is a Linux based distributed file system, designed to be highly scalable and serve many clients. Some reasons to use GlusterFS are: No centralized metadata server, Scalability, Open Source, Dynamic and live service modifications, Can be used over Infiniband or Ethernet, Can be tuned for speed and/or resilience and Flexible administration. It's useful for enterprise environments - virtualization; high performance computing (HPC) and it works with Mac, Linux and Windows clients. Conclusions are: (1) GlusterFS proved to have widespread capabilities as a virtual file system; (2) Scalability is very dependent upon the underlying hardware; (3) Lack of built-in encryption and security paradigm; and (4) Best suited in a general purpose computing environment.

99 GENERAL AND MISCELLANEOUS↗
DOE OSTI2021

Online data analysis and reduction: An important co-design motif for extreme-scale computers

A growing disparity between supercomputer computation speeds and I/O rates means that it is rapidly becoming infeasible to analyze supercomputer application output only after that output has been written to a file system. Instead, data-generating applications must run concurrently with data reduction and/or analysis operations, with which they exchange information via high-speed methods such as interprocess communications. The resulting parallel computing motif, online data analysis and reduction (ODAR), has important implications for both application and HPC systems design. Here we introduce the ODAR motif and its co-design concerns, describe a co-design process for identifying and addressing those concerns, present tools that assist in the co-design process, and present case studies to illustrate the use of the process and tools in practical settings.

Data Analysis↗
DOE OSTI2019

Throughput Analytics of Data Transfer Infrastructures

To support increasingly distributed scientific and big-data applications, powerful data transfer infrastructures are being built with dedicated networks and software frameworks customized to distributed file systems and data transfer nodes. The data transfer performance of such infrastructures critically depends on the combined choices of file, disk, and host systems as well as network protocols and file transfer software, all of which may vary across sites. The randomness of throughput measurements makes it challenging to assess the impact of these choices on the performance of infrastructure or its parts. We propose regression-based throughput profiles by aggregating measurements from sites of the infrastructure, with RTT as the independent variable. The peak values and convex-concave shape of a profile together determine the overall throughput performance of memory and file transfers, and its variations show the performance differences among the sites. We then present projection and difference operators, and coefficients of throughput profiles to characterize the performance of infrastructure and its parts, including sites and file transfer tools. In particular, the utilization-concavity coefficient provides a value in the range [0, 1] that reflects overall transfer effectiveness. We present results of measurements collected using (i) testbed experiments over dedicated 0–366 ms 10 Gbps connections with combinations of TCP versions, file systems, host systems and transfer tools, and (ii) Globus GridFTP transfers over production infrastructure with varying site configurations.

Rao, Nageswara S.↗
DOE OSTI2019

Big Data Framework with Machine Learning for D and D Applications - 19108

The nuclear industry is experiencing a steady increase in maintenance costs even though plants are maintained under high levels of safety, capability and reliability. Nuclear power plants are expected to run every unit at maximum capacity at all times, efficiently utilizing assets with minimal downtime. Surveillance and maintenance of nuclear decommissioning infrastructure provides lot challenges with respect to maintenance or decommissioning of the buildings. There is a need for a framework to analyze the huge amount of data generated by the sensors on the nuclear reactor components as well as structures, to monitor the conditions of these building over a period of time. Emerging technologies such as big data analytics have become a requirement in the nuclear industry to improve structural health monitoring and diagnostics. FIU will make use of existing mature technologies in the areas of imaging, robotics, big data, and machine learning/deep learning to assess the structural integrity of aging facilities at DOE sites. As these facilities await decommissioning, there is a need to understand the structural health of these structures. Many of these facilities were built over 50 years ago and in some cases these facilities have gone beyond the operational life expectancy. In other cases, the facilities have been placed in a state of 'cold and dark' and they are sitting unused, awaiting decommissioning. Finally, some older facilities are one-of-a-kind operational/production facilities supporting critical DOE missions and cannot be shut down. In any of these scenarios, the structural integrity of these facilities may be compromised, so it is imperative that adequate inspections and data collection/analysis be performed on a continuous and ongoing basis. The primary goals of the research include collecting various formats of data such as structured, unstructured and streaming data from the various sensors deployed in buildings, collect videos/pictures from various imaging sources, ingest them using a Hadoop distributed file system and process the data using Spark to perform batch processing and real time analytics. Research and development on various machine learning/deep learning algorithms will also be performed to analyze the heterogeneous data collected from nuclear decommission infrastructure. FIU will design the big data framework to ingest, store and process huge amounts of heterogeneous data collected from many sources and optimize the algorithms to provide insights into the data and predict anomalies observed when compared against baseline conditions. Various modules of this framework will include heterogeneous data sources, message broker, Hadoop distributed file system, Spark for stream and batch processing, machine learning/ deep learning, Cassandra - persistent data store and visualization. (authors)

12 MANAGEMENT OF RADIOACTIVE AND NON-RADIOACTIVE W↗
DOE OSTI2023

LCIO

Performance of file systems shift during their life cycles. Evaluating this performance change over time is not trivial. Complexity arises in the interplay between external (i.e. application I/O workloads) and internal (i.e. the filesystem state) factors. Many benchmarks can test how a filesystem performs at the current snapshot state, but to observe the change over time necessitates that the filesystem state mutate (age) between benchmark runs. For a large-scale HPC parallel filesystem, the sheer scale and amount of interacting components during I/O operations magnify these challenges. LCIO addresses the question - how will the filesystem perform at different stages of its life cycle? LCIO is a synthetic benchmark, which provides the file system aging process to increase the amount of information that existing benchmarks like IOR and MDTest yield, as well as provide additional points of data that will be useful to system architects and engineers.

Bachstein, Matthew↗
DOE OSTI2022

Large Scale Caching and Streaming of Training Data for Online Deep Learning

The training of deep neural network models on large data remains a difficult problem, despite progress towards scalable techniques. In particular, there is a mismatch between the random but predetermined order in which AI flows select training samples and the streaming I/O patterns for which traditional HPC data storage (e.g., parallel file systems) are designed. In addition, as more data are obtained, it is feasible neither simply to train learning models incrementally, due to catastrophic forgetting (i.e., bias towards new samples), nor to train frequently from scratch, due to prohibitive time and/or resource constraints. In this paper, we study data management techniques that combine caching and streaming with rehearsal support in order to enable efficient access to training samples in both offline training and continual learning. We revisit state-of-art streaming approaches based on data pipelines that transparently handle prefetching, caching, shuffling, and data augmentation, and discuss the challenges and opportunities that arise when combining these methods with data-parallel training techniques. We also report on preliminary experiments that evaluate the I/O overheads involved in accessing the training samples from a parallel file system (PFS) under several concurrency scenarios, highlighting the impact of the PFS on the design of the data pipelines.

data pipelines↗
DOE OSTI2018

ARCHIE: Data Analysis Acceleration with Array Caching in Hierarchical Storage

© 2018 IEEE. Scientific data analysis typically involves reading massive amounts of data that was generated by simulations, experiments, and observations. Performance of reading such large volumes of data from disk-based file systems is often poor because of the slow and mechanical components in the disks. Recent supercomputing systems are adding non-volatile storage layers in a hierarchy to handle the performance gap between fast main memory and slow disk-based storage. Software libraries for managing this hierarchy not only need efficient reading of data but also reduce user-involvement for cross-layer data movement. Furthermore, these libraries need to support array data access patterns into hierarchical storage management as scientific data is often organized in array-based data structures. Existing software typically manage individual storage layers requiring significant manual process in moving data among them. In this paper, we introduce a new array caching in hierarchical storage (ARCHIE) to accelerate array data analysis in a seamless fashion. ARCHIE evaluates array access patterns and prefetches data with array semantics between storage layers. Our evaluation shows that ARCHIE outperforms state-of-the-art file systems, i.e., Lustre and DataWarp, on a production supercomputing system by up to 5.8× in accessing data by scientific analysis applications.

Dong, Bin↗
DOE OSTI2026

FitCache: A Transparent Drop-In Framework for Multi-Tier Caching to Accelerate Distributed Deep Learning Workloads

Training in Deep learning (DL) remains highly compute- and data-intensive, with I/O becoming a critical bottleneck as models and datasets scale. Recent studies report that data loading can dominate training time, especially on large-scale HPC systems with shared parallel file systems (PFS). Existing caching approaches either rely on single-tier designs or require intrusive modifications to training pipelines, limiting their portability and effectiveness. In this work, we present FitCache, a transparent drop-in framework for multi-tier caching to accelerate distributed DL training by coordinating fast local memory (e.g., DRAM, Persistent Memory (PMem)) and NVMe as hierarchical caches atop PFS. Our design adapts to hardware diversity, i.e., if NVMe is missing, memory transparently acts as a caching tier, ensuring stable performance. FitCache transparently intercepts I/O requests and issues concurrent fetches across all tiers, returning data from the fastest responder without centralized metadata or static redirection paths. FitCache adapts to dynamic workloads and heterogeneous clusters while maintaining POSIX compatibility. Experiments on Frontier (2048 GPUs) and smaller research clusters show that FitCache reduces training time by up to 40% and per-batch I/O latency by up to 71.6% compared to Lustre Orion PFS, offering a drop-in solution for scalable DL training.

Hu, Guangxing [ORNL] (ORCID:0009000283203614)↗
DOE OSTI2012

FY12 Quarter 3 Computing Utilization Report – LANL

DSW continues to dominate the capacity workload, with a focus in Q3 on common model baselining runs in preparation for the Annual Assessment Review (AAR) of the weapon systems. There remains unmet demand for higher fidelity simulations, and for increased throughput of simulations. Common model baselining activities would benefit from doubling the resolution of the models and running twice as many simulations. Capacity systems were also utilized during the quarter to prepare for upcoming Level 2 milestones. Other notable DSW activities include validation of new physics models and safety studies. The safety team used the capacity resources extensively for projects involving 3D computer simulations for the Furrow series of experiments at DARHT (a Level 2 milestone), fragment impact, surety theme, PANTEX assessments, and the 120-day study. With the more than tripling of classified capacity computing resources with the addition of the Luna system and the safety team's imminent access to the Cielo system, demand has been met for current needs. The safety team has performed successful scaling studies on Luna up to 16K PE size-jobs with linear scaling, running the large 3D simulations required for the analysis of Furrow. They will be investigating scaling studies on the Cielo system with the Lustre file system in Q4. Overall average capacity utilization was impacted by negative effects of the LANL Voluntary Separation Program (VSP) at the beginning of Q3, in which programmatic staffing was reduced by 6%, with further losses due to management backfills and attrition, resulting in about 10% fewer users. All classified systems were impacted in April by a planned 2 day red network outage. ASC capacity workload continues to focus on code development, regression testing, and verification and validation (V&V) studies. Significant capacity cycles were used in preparation for a JOWOG in May and several upcoming L2 milestones due in Q4. A network transition has been underway on the unclassified networks to increase access of all ASC users to the unclassified systems through the Yellow Turquoise Integration (YeTI) project. This will help to alleviate the longstanding shortage of resources for ASC unclassified code development and regression testing, and also make a broader palette of machines available to unclassified ASC users, including PSAAP Alliance users. The Moonlight system will be the first capacity resource to be made available through the YETI project, and will make available a significant increase in cycles, as well as GPGPU accelerator technology. The Turing and Lobo machines will be decommissioned in the next quarter. ASC projects running on Cielo as part of the CCC-3 include turbulence, hydrodynamics, burn, asteroids, polycrystals, capability and runtime performance improvements, and materials including carbon and silicone.

43 PARTICLE ACCELERATORS↗
DOE OSTI2023

Analyzing File Access Patterns on Large-Scale HPC Systems: Opportunities for File Prefetching

This paper explores the potential opportunities for implementing file prefetching techniques on large-scale high-performance computing (HPC) systems. Specifically, we investigate the file access patterns of various applications across multiple scientific domains using two years' worth of Darshan I/O traces obtained from the Summit supercomputer. We identify recurring trends and patterns which indicate that prefetching can be effectively leveraged to improve data access performance on HPC systems. This study serves as a valuable reference for system architects and developers in the HPC community, providing insights into the opportunities and challenges associated with enabling file prefetching on large-scale HPC systems.

Karimi, Ahmad Maroof↗
DOE OSTI2016

Measurements of file transfer rates over dedicated long-haul connections

Wide-area file transfers are an integral part of several High-Performance Computing (HPC) scenarios. Dedicated network connections with high capacity, low loss rate and low competing traffic, are increasingly being provisioned over current HPC infrastructures to support such transfers. To gain insights into these file transfers, we collected transfer rate measurements for Lustre and xfs file systems between dedicated multi-core servers over emulated 10 Gbps connections with round trip times (rtt) in 0-366 ms range. Memory transfer throughput over these connections is measured using iperf, and file IO throughput on host systems is measured using xddprof. We consider two file system configurations: Lustre over IB network and xfs over SSD connected to PCI bus. Files are transferred using xdd across these connections, and the transfer rates are measured, which indicate the need to jointly optimize the connection and host file IO parameters to achieve peak transfer rates. In particular, these measurements indicate that (i) peak file transfer rate is lower than peak connection and host IO throughput, in some cases by as much as 50% or lower, (ii) xdd request sizes that achieve peak throughput for host file IO do not necessarily lead to peak file transfer rates, and (iii) parallelism in host IO and TCP transport does not always improve the file transfer rates.

log-haul connections↗
DOE OSTI2016

Measurements of file transfer rates over dedicated long-haul connections

Wide-area file transfers are an integral part of several High-Performance Computing (HPC) scenarios. Dedicated network connections with high capacity, low loss rate and low competing traffic, are increasingly being provisioned over current HPC infrastructures to support such transfers. To gain insights into these file transfers, we collected transfer rate measurements for Lustre and xfs file systems between dedicated multi-core servers over emulated 10 Gbps connections with round trip times (rtt) in 0-366 ms range. Memory transfer throughput over these connections is measured using iperf, and file IO throughput on host systems is measured using xddprof. We consider two file system configurations: Lustre over IB network and xfs over SSD connected to PCI bus. Files are transferred using xdd across these connections, and the transfer rates are measured, which indicate the need to jointly optimize the connection and host file IO parameters to achieve peak transfer rates. In particular, these measurements indicate that (i) peak file transfer rate is lower than peak connection and host IO throughput, in some cases by as much as 50% or lower, (ii) xdd request sizes that achieve peak throughput for host file IO do not necessarily lead to peak file transfer rates, and (iii) parallelism in host IO and TCP transport does not always improve the file transfer rates.

Rao, Nageswara S↗
NASA NTRS1998

The mass storage testing laboratory at GSFC

Industry-wide benchmarks exist for measuring the performance of processors (SPECmarks), and of database systems (Transaction Processing Council). Despite storage having become the dominant item in computing and IT (Information Technology) budgets, no such common benchmark is available in the mass storage field. Vendors and consultants provide services and tools for capacity planning and sizing, but these do not account for the complete set of metrics needed in today's archives. The availability of automated tape libraries, high-capacity RAID systems, and high- bandwidth interconnectivity between processor and peripherals has led to demands for services which traditional file systems cannot provide. File Storage and Management Systems (FSMS), which began to be marketed in the late 80's, have helped to some extent with large tape libraries, but their use has introduced additional parameters affecting performance. The aim of the Mass Storage Test Laboratory (MSTL) at Goddard Space Flight Center is to develop a test suite that includes not only a comprehensive check list to document a mass storage environment but also benchmark code. Benchmark code is being tested which will provide measurements for both baseline systems, i.e. applications interacting with peripherals through the operating system services, and for combinations involving an FSMS. The benchmarks are written in C, and are easily portable. They are initially being aimed at the UNIX Open Systems world. Measurements are being made using a Sun Ultra 170 Sparc with 256MB memory running Solaris 2.5.1 with the following configuration: 4mm tape stacker on SCSI 2 Fast/Wide; 4GB disk device on SCSI 2 Fast/Wide; and Sony Petaserve on Fast/Wide differential SCSI 2.

Venkataraman, Ravi↗
NASA NTRS2016

A Spatiotemporal Indexing Approach for Efficient Processing of Big Array-Based Climate Data with MapReduce

Climate observations and model simulations are producing vast amounts of array-based spatiotemporal data. Efficient processing of these data is essential for assessing global challenges such as climate change, natural disasters, and diseases. This is challenging not only because of the large data volume, but also because of the intrinsic high-dimensional nature of geoscience data. To tackle this challenge, we propose a spatiotemporal indexing approach to efficiently manage and process big climate data with MapReduce in a highly scalable environment. Using this approach, big climate data are directly stored in a Hadoop Distributed File System in its original, native file format. A spatiotemporal index is built to bridge the logical array-based data model and the physical data layout, which enables fast data retrieval when performing spatiotemporal queries. Based on the index, a data-partitioning algorithm is applied to enable MapReduce to achieve high data locality, as well as balancing the workload. The proposed indexing approach is evaluated using the National Aeronautics and Space Administration (NASA) Modern-Era Retrospective Analysis for Research and Applications (MERRA) climate reanalysis dataset. The experimental results show that the index can significantly accelerate querying and processing (10 speedup compared to the baseline test using the same computing cluster), while keeping the index-to-data ratio small (0.0328). The applicability of the indexing approach is demonstrated by a climate anomaly detection deployed on a NASA Hadoop cluster. This approach is also able to support efficient processing of general array-based spatiotemporal data in various geoscience domains without special configuration on a Hadoop cluster.

big data↗
DOE OSTI2016

Collectively loading programs in a multiple program multiple data environment

Techniques are disclosed for loading programs efficiently in a parallel computing system. In one embodiment, nodes of the parallel computing system receive a load description file which indicates, for each program of a multiple program multiple data (MPMD) job, nodes which are to load the program. The nodes determine, using collective operations, a total number of programs to load and a number of programs to load in parallel. The nodes further generate a class route for each program to be loaded in parallel, where the class route generated for a particular program includes only those nodes on which the program needs to be loaded. For each class route, a node is selected using a collective operation to be a load leader which accesses a file system to load the program associated with a class route and broadcasts the program via the class route to other nodes which require the program.

97 MATHEMATICS AND COMPUTING↗