Search NASA⌕ Search

SEARCH · Search NASA

Results for “DATA INTENSIVE COMPUTING”

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 145 records · Page 8

Large-scale compute-intensive analysis via a combined in-situ and co-scheduling workflow approach, In: SC '15 Proceedings of the International Conference for High Performance Computing, Networking, Storage and Analysis

Large-scale simulations can produce hundreds of terabytes to petabytes of data, complicating and limiting the efficiency of workflows. Traditionally, outputs are stored on the file system and analyzed in post-processing. With the rapidly increasing size and complexity of simulations, this approach faces an uncertain future. Trending techniques consist of performing the analysis in-situ, utilizing the same resources as the simulation, and/or off-loading subsets of the data to a compute-intensive analysis system. We introduce an analysis framework developed for HACC, a cosmological N-body code, that uses both in-situ and co-scheduling approaches for handling petabyte-scale outputs. We compare different analysis set-ups ranging from purely off-line, to purely in-situ to in-situ/co-scheduling. The analysis routines are implemented using the PISTON/VTK-m framework, allowing a single implementation of an algorithm that simultaneously targets a variety of GPU, multi-core, and many-core architectures.

Computer Science↗

Defect identification in simulated Bragg coherent diffraction imaging by automated AI

X-ray Bragg coherent diffraction imaging is a powerful technique for operando and in situ materials characterization and provides a unique means of quantifying the influence of one-dimensional (1D) and two-dimensional (2D) material defects on material response. However, obtaining full images from raw x-ray diffraction data is nontrivial and computationally intensive, precluding real-time experimental feedback. Here, we present a machine learning approach to identify the presence of crystalline line defects (edge and screw) in samples from the raw, 2D, coherent diffraction data without the need for image reconstruction through iterative phase retrieval. Further, we compare different approaches to designing neural networks for this application and demonstrate the potential of automated ML (autoML) approaches.

36 MATERIALS SCIENCE↗

Domain wall-magnetic tunnel junction spin–orbit torque devices and circuits for in-memory computing

There are pressing problems with traditional computing, especially for accomplishing data-intensive and real-time tasks, that motivate the development of in-memory computing devices to both store information and perform computation. Magnetic tunnel junction memory elements can be used for computation by manipulating a domain wall, a transition region between magnetic domains, but the experimental study of such devices has been limited by high current densities and low tunnel magnetoresistance. Here, we study prototypes of three-terminal domain wall-magnetic tunnel junction in-memory computing devices that can address data processing bottlenecks and resolve these challenges by using perpendicular magnetic anisotropy, spin–orbit torque switching, and an optimized lithography process to produce average device tunnel magnetoresistance TMR = 171% and average resistance-area product RA = 29 Ω μm2, close to the RA of the unpatterned film. Device initialization variation in switching voltage is shown to be curtailed to 7%–10% by controlling the domain wall initial position, which we show corresponds to 90%–96% accuracy in a domain wall-magnetic tunnel junction full adder simulation. Repeatability of writing and resetting the device is shown. A circuit shows an inverter operation between two devices, showing that a voltage window is large enough, compared to the variation noise, to repeatably operate a domain wall-magnetic tunnel junction circuit. These results make strides in using magnetic tunnel junctions and domain walls for in-memory and neuromorphic computing applications.

Alamdar, Mahshid (ORCID:0000000221732935)↗

Software-Hardware Co-design of Heterogeneous SmartNIC System for Recommendation Models Inference and Training

Deep Learning Recommendation Models (DLRMs) are critical applications in various domains and have evolved as one of the single largest machine learning applications. Trillions of DLRM parameters exceed the on-chip memory capacity of GPUs. Large-scale multi-node systems are required for distributed DLRM inference and training, which suffer from the all-to-all communication bottleneck, mainly limiting the scalability of ever-growing DLRMs. In recent years, SmartNICs have evolved with coupled computation and communication capabilities providing opportunities for a powerful heterogeneous device in the system. However, there isn't such a distributed system that fully leverages the abundant smartNIC resources that resolve the scalability issue of DLRMs. In this work, we proposed a software-hardware co-design of a heterogeneous smartNIC system that resolves the communication bottleneck of distributed DLRMs, mitigates the memory bandwidth pressure, and improves computation efficiency. We provide a set of smartNIC designs of cache systems (including local cache and remote cache) and smartNIC computation kernels which reduce data movement, relieve memory lookup intensity, and improve the GPU's computation efficiency. In addition, we propose a graph algorithm that improves the data locality of queries within batches which optimizes the overall system performance with higher data reuse. Our evaluation shows that our system achieves 2.1x latency speedup for inference and 1.6x throughput speedup for training.

Guo, Anqi↗

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

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

97 MATHEMATICS AND COMPUTING↗

LOGAN: High-Performance X-Drop Pairwise Alignment on GPU (LOGAN) v1.0

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

Guidi, Giulia↗

Special issue: Advanced stencil-code engineering

Here, stencil codes are compute-intensive algorithms, in which data points arranged in a large grid are being recomputed repeatedly from the values of data points in a predefined neighborhood. This fixed neighborhood pattern is called a stencil. Stencils codes see wide-spread use in computing the discrete solutions of partial differential equations and systems composed of such equations.

97 MATHEMATICS AND COMPUTING↗

Integrating Applied Energy and BER Smart Data Capabilities to Develop a DOE Data Fabric for Energy-Water R&D

Focal Area(s): 1) Data acquisition and assimilation enabled by machine learning, AI, and advanced methods including experimental/network design/optimization, unsupervised learning (including deep learning), and hardware-related efforts involving AI (e.g., edge computing). Science Challenge: DOE R&D, including DOE’s Basic Energy Research (BER)’s Environmental Systems Science Division (EESSD) program and DOE’s applied energy research (AER) programs (EERE, FE, and NE) are producers and consumers of Earth systems datasets. This white paper focuses on the first topic area from the call in relation to how crosscutting resources and innovations from DOE’s EESSD and AER can be brought to bear to mutual benefit and more efficient energy-water, Earth system data resources through improved. The overarching challenge posed by this call focuses on how DOE can directly leverage artificial intelligence (AI) to engineer a substantial (paradigm-changing) improvement in Earth System Predictability? While stemming from DOE BER’s EESSD program, this is a challenge that is faced and also being addressed by DOE’s AER programs. Over the past decade plus, FE, EERE, and NE programs have made important strides towards addressing this need. These strides are in many ways highly complementary to EESSD’s MODEX efforts. Energy water systems spanning metocean to groundwater to surface water systems all are data driven whether for basic energy or applied energy. These are remote, multi-variate, complex natural, and in many cases engineered, systems. Key needs and challenges of both EESSD and AER include developing data-focused tools to enhance data search and discovery to fill in knowledge gaps (address sparse data challenge), and rapidly transform datasets, including disparate and multi-source data. Leveraging DOE on-premise computing (HPC, exascale) infrastructure supports the computing-intensive algorithms required to execute these data acquisition and transformation processes to derive enriched knowledge and data, driving AI/ML and big data analytics for these systems. The opportunity lies in combining BER and AER efforts to provide a more robust, advanced, efficient and complete computing data fabric to address energy-water data acquisition and assimilation needs which currently pose significant impediments to AI/ML predictions and research.

54 ENVIRONMENTAL SCIENCES↗

GLIDE: a grid-based light-weight infrastructure for data-intensive environments

The promise of the grid is that it will enable public access and sharing of immense amounts of computational and data resources among dynamic coalitions of individuals and institutions. However, the current grid solutions make several limiting assumptions that curtail their widespread adoption. To address these limitations, we present GLIDE, a prototype light-weight, data-intensive middleware infrastructure that enables access to the robust data and computational power of the grid on DREAM platforms.

PRISM-MW↗

DeepSAT: A Deep Learning Approach to Tree-Cover Delineation in 1-m NAIP Imagery for the Continental United States

High resolution tree cover classification maps are needed to increase the accuracy of current land ecosystem and climate model outputs. Limited studies are in place that demonstrates the state-of-the-art in deriving very high resolution (VHR) tree cover products. In addition, most methods heavily rely on commercial softwares that are difficult to scale given the region of study (e.g. continents to globe). Complexities in present approaches relate to (a) scalability of the algorithm, (b) large image data processing (compute and memory intensive), (c) computational cost, (d) massively parallel architecture, and (e) machine learning automation. In addition, VHR satellite datasets are of the order of terabytes and features extracted from these datasets are of the order of petabytes. In our present study, we have acquired the National Agriculture Imagery Program (NAIP) dataset for the Continental United States at a spatial resolution of 1-m. This data comes as image tiles (a total of quarter million image scenes with ~60 million pixels) and has a total size of ~65 terabytes for a single acquisition. Features extracted from the entire dataset would amount to ~8-10 petabytes. In our proposed approach, we have implemented a novel semi-automated machine learning algorithm rooted on the principles of "deep learning" to delineate the percentage of tree cover. Using the NASA Earth Exchange (NEX) initiative, we have developed an end-to-end architecture by integrating a segmentation module based on Statistical Region Merging, a classification algorithm using Deep Belief Network and a structured prediction algorithm using Conditional Random Fields to integrate the results from the segmentation and classification modules to create per-pixel class labels. The training process is scaled up using the power of GPUs and the prediction is scaled to quarter million NAIP tiles spanning the whole of Continental United States using the NEX HPC supercomputing cluster. An initial pilot over the state of California spanning a total of 11,095 NAIP tiles covering a total geographical area of 163,696 sq. miles has produced true positive rates of around 88 percent for fragmented forests and 74 percent for urban tree cover areas, with false positive rates lower than 2 percent for both landscapes.

Imagery↗

Enabling Scalable and Extensible Memory-mapped Datastores in Userspace

Exascale workloads are expected to incorporate data-intensive processing in close coordination with traditional physics simulations. These emerging scientific, data-analytics and machine learning applications need to access a wide variety of datastores in flat files and structured databases. Programmer productivity is greatly enhanced by mapping datastores into the application process's virtual memory space to provide a unified “in-memory” interface. Currently, memory mapping is provided by system software primarily designed for generality and reliability. However, scalability at high concurrency is a formidable challenge on exascale systems. Also, there is a need for extensibility to support new datastores potentially requiring HPC data transfer services. In this article, we present UMap , a scalable and extensible userspace service for memory-mapping datastores. Furthermore, through decoupled queue management, concurrency aware adaptation, and dynamic load balancing, UMap enables application performance to scale even at high concurrency. We evaluate UMap in data-intensive applications, including sorting, graph traversal, database operations, and metagenomic analytics. Our results show that UMap as a userspace service outperforms an optimized kernel-based service across a wide range of intra-node concurrency by 1.22-1.9 × . We performed two case studies to demonstrate UMap 's extensibility. First, a new datastore residing in remote memory is incorporated into UMap as an application-specific plugin. Second, we present a persistent memory allocator Metall built atop UMap for unified storage/memory.

97 MATHEMATICS AND COMPUTING↗

BReW: Blackbox Resource Selection for e-Science Workflows

Workflows are commonly used to model data intensive scientific analysis. As computational resource needs increase for eScience, emerging platforms like clouds present additional resource choices for scientists and policy makers. We introduce BReW, a tool enables users to make rapid, highlevel platform selection for their workflows using limited workflow knowledge. This helps make informed decisions on whether to port a workflow to a new platform. Our analysis of synthetic and real eScience workflows shows that using just total runtime length, maximum task fanout, and total data used and produced by the workflow, BReW can provide platform predictions comparable to whitebox models with detailed workflow knowledge.

97 MATHEMATICS AND COMPUTING↗

Navigating Uncertainty: Challenges in Visualizing Ensemble Data and Surrogate Models for Decision Systems

Uncertainty visualization plays a critical role in transforming ensemble simulation data into actionable insights by effectively communicating various dimensions of uncertainty within a system. The emergence of artificial intelligence-driven surrogate models trained on multirun ensemble data offers a transformative opportunity to replace computationally intensive simulations with fast estimates, enabling users to explore data spaces with unprecedented depth and interactivity. However, integrating ensemble data and surrogate models into decision-making workflows and tools introduces novel challenges for uncertainty visualization. These include reconciling and clearly communicating the unique uncertainties associated with ensembles and their surrogate model estimates, and leveraging these approximations to inform actionable decisions. This work explores these challenges in the context of high-dimensional data visualization, bridging discrete datasets with their continuous representations and addressing the complexities of systems that support iterative navigation between input and output spaces. We evaluate the role of uncertainty visualization in fostering intuitive, actionable interactions and identify critical hurdles in advancing this frontier of computational simulation.

97 MATHEMATICS AND COMPUTING↗

Development of an Efficient Binaural Simulation for the Analysis of Structural Acoustic Data

Applying binaural simulation techniques to structural acoustic data can be very computationally intensive as the number of discrete noise sources can be very large. Typically, Head Related Transfer Functions (HRTFs) are used to individually filter the signals from each of the sources in the acoustic field. Therefore, creating a binaural simulation implies the use of potentially hundreds of real time filters. This paper details two methods of reducing the number of real-time computations required by: (i) using the singular value decomposition (SVD) to reduce the complexity of the HRTFs by breaking them into dominant singular values and vectors and (ii) by using equivalent source reduction (ESR) to reduce the number of sources to be analyzed in real-time by replacing sources on the scale of a structural wavelength with sources on the scale of an acoustic wavelength. The ESR and SVD reduction methods can be combined to provide an estimated computation time reduction of 99.4% for the structural acoustic data tested. In addition, preliminary tests have shown that there is a 97% correlation between the results of the combined reduction methods and the results found with the current binaural simulation techniques

Johnson, Marty E.↗

Accelerating nuclear configuration interaction calculations through a preconditioned block iterative eigensolver

We describe a number of recently developed techniques for improving the performance of large-scalenuclear configuration interaction calculations on high performance parallel computers. We show thebenefit of using a preconditioned block iterative method to replace the Lanczos algorithm that hastraditionally been used to perform this type of computation. The rapid convergence of the block iterativemethod is achieved by a proper choice of starting guesses of the eigenvectors and the construction of aneffective preconditioner. These acceleration techniques take advantage of special structure of the nuclearconfiguration interaction problem which we discuss in detail. The use of a block method also allows usto improve the concurrency of the computation, and take advantage of the memory hierarchy of modernmicroprocessors to increase the arithmetic intensity of the computation relative to data movement. Wealso discuss the implementation details that are critical to achieving high performance on massivelyparallel multi-core supercomputers, and demonstrate that the new block iterative solver is two to threetimes faster than the Lanczos based algorithm for problems of moderate sizes on a Cray XC30 system.

Shao, Meiyue↗

Virtual machine provisioning, code management, and data movement design for the Fermilab HEPCloud Facility

The Fermilab HEPCloud Facility Project has as its goal to extend the current Fermilab facility interface to provide transparent access to disparate resources including commercial and community clouds, grid federations, and HPC centers. This facility enables experiments to perform the full spectrum of computing tasks, including data-intensive simulation and reconstruction. We have evaluated the use of the commercial cloud to provide elasticity to respond to peaks of demand without overprovisioning local resources. Full scale data-intensive workflows have been successfully completed on Amazon Web Services for two High Energy Physics Experiments, CMS and NOνA, at the scale of 58000 simultaneous cores. This paper describes the significant improvements that were made to the virtual machine provisioning system, code caching system, and data movement system to accomplish this work. The virtual image provisioning and contextualization service was extended to multiple AWS regions, and to support experiment-specific data configurations. A prototype Decision Engine was written to determine the optimal availability zone and instance type to run on, minimizing cost and job interruptions. We have deployed a scalable on-demand caching service to deliver code and database information to jobs running on the commercial cloud. It uses the frontiersquid server and CERN VM File System (CVMFS) clients on EC2 instances and utilizes various services provided by AWS to build the infrastructure (stack). We discuss the architecture and load testing benchmarks on the squid servers. We also describe various approaches that were evaluated to transport experimental data to and from the cloud, and the optimal solutions that were used for the bulk of the data transport. Finally, we summarize lessons learned from this scale test, and our future plans to expand and improve the Fermilab HEP Cloud Facility.

72 PHYSICS OF ELEMENTARY PARTICLES AND FIELDS↗

Parallel Breadth-First Search on Distributed Memory Systems

Data-intensive, graph-based computations are pervasive in several scientific applications, and are known to be quite challenging to implement on distributed memory systems. In this work, we explore the design space of parallel algorithms for Breadth-First Search (BFS), a key subroutine in several graph algorithms. We present two highly-tuned parallel approaches for BFS on large parallel systems: a level-synchronous strategy that relies on a simple vertex-based partitioning of the graph, and a two-dimensional sparse matrix-partitioning-based approach that mitigates parallel communication overhead. For both approaches, we also present hybrid versions with intra-node multithreading. Our novel hybrid two-dimensional algorithm reduces communication times by up to a factor of 3.5, relative to a common vertex based approach. Our experimental study identifies execution regimes in which these approaches will be competitive, and we demonstrate extremely high performance on leading distributed-memory parallel systems. For instance, for a 40,000-core parallel execution on Hopper, an AMD Magny-Cours based system, we achieve a BFS performance rate of 17.8 billion edge visits per second on an undirected graph of 4.3 billion vertices and 68.7 billion edges with skewed degree distribution.

97 MATHEMATICS AND COMPUTING↗

Validating and Comparing Energy Estimation Methods at Water Resource Recovery Facilities

Water resource recovery facilities play a crucial role in the water-energy nexus, consuming a substantial amount of energy in the United States. Growing treatment volumes and more stringent water quality standards are expected to increase the amount of energy needed to treat wastewater, but accurately estimating energy consumption and potential remains challenging due to variability in scale, treatment methods, and effluent treatment standards. In this study, we used publicly available data to evaluate the accuracy of methods for estimating energy consumption and generation, then quantified uncertainty based on key factors like flow rate, treatment level, and geographic location. To validate methods, we estimated energy consumption and generation at the facility-level, then compared estimates to self-reported data from utilities in major U.S. cities. We found that process models of treatment trains under best practice configurations were accurate relative to other methods for estimating electricity use, total energy use, and electricity generation from biogas utilization, and less complex methods based on effluent treatment level and prime movers also performed well for estimating electricity consumption and generation, respectively. Applying the evaluated methods to a national inventory of treatment facilities, we estimate that annual energy consumption ranged from 56.3 x 10^3 to 82.5 x 10^3 TJ in 2012 and 83.6 x 10^3 to 127 x 10^3 TJ in 2042. Our results indicate that not all estimation methods are suited for every use case, so we recommend that researchers and practitioners select an estimation method based on data availability and desired computational intensity.

Hodson, Abigayle↗