Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “HPC performance modeling”

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 559 records · Page 31

High Performance Computing Metrics to Enable Application-Platform Communication

Sandia has invested heavily in scientific/engineering application development and in the research, development, and deployment of large scale HPC platforms to support the computational needs of these applications. As application developers continually expand the capabilities of their software and spend more time on performance tuning of applications for these platforms, HPC platform resources are at a premium as they are a heavily shared resource serving the varied needs of many users. To ensure that the HPC platform resources are being used effciently and perform as designed, it is necessary to obtain reliable data on resource utilization that will allow us to investigate the occurrence, severity, and causes of performance-affecting contention between applications. The work presented in this paper was an initial step to determine if resource contention can be understood and minimized through monitoring, modeling, planning and infrastructure. This paper describes the set of metric definitions, identified in this research, that can be used as meaningful and potentially actionable indicators of performance-affecting contention between applications. These metrics were verified using the observed slowdown of IOR, IMB, and CTH in operating scenarios that forced contention. This paper also describes system/application monitoring activities that are critical to distilling vast amounts of data into quantities that hold the key to understanding for an application's performance under production conditions and that will ultimately aid in Sandia's efforts to succeed in extreme-scale computing.

97 MATHEMATICS AND COMPUTING↗

Adiabatic Quantum Computation Applied to Deep Learning Networks

Training deep learning networks is a difficult task due to computational complexity, and this is traditionally handled by simplifying network topology to enable parallel computation on graphical processing units (GPUs). However, the emergence of quantum devices allows reconsideration of complex topologies. We illustrate a particular network topology that can be trained to classify MNIST data (an image dataset of handwritten digits) and neutrino detection data using a restricted form of adiabatic quantum computation known as quantum annealing performed by a D-Wave processor. We provide a brief description of the hardware and how it solves Ising models, how we translate our data into the corresponding Ising models, and how we use available expanded topology options to explore potential performance improvements. Although we focus on the application of quantum annealing in this article, the work discussed here is just one of three approaches we explored as part of a larger project that considers alternative means for training deep learning networks. The other approaches involve using a high performance computing (HPC) environment to automatically find network topologies with good performance and using neuromorphic computing to find a low-power solution for training deep learning networks. Our results show that our quantum approach can find good network parameters in a reasonable time despite increased network topology complexity; that HPC can find good parameters for traditional, simplified network topologies; and that neuromorphic computers can use low power memristive hardware to represent complex topologies and parameters derived from other architecture choices.

97 MATHEMATICS AND COMPUTING↗

Enabling discovery data science through cross-facility workflows

Experimental and observational instruments for scientific research (such as light sources, genome sequencers, accelerators, telescopes and electron microscopes) increasingly require High Performance Computing (HPC) scale capabilities for data analysis and workflow processing. Next-generation instruments are being deployed with higher resolutions and faster data capture rates, creating a big data crunch that cannot be handled by modest institutional computing resources. Often these big data analysis pipelines also require near real-time computing and have higher resilience requirements than the simulation and modeling workloads more traditionally seen at HPC centers. While some facilities have enabled workflows to run at a single HPC facility, there is a growing need to integrate capabilities across HPC facilities to enable cross-facility workflows, either to provide resilience to an experiment, increase analysis throughput capabilities, or to better match a workflow to a particular architecture. In this paper we describe the barriers to executing complex data analysis workflows across HPC facilities and propose an architectural design pattern for enabling scientific discovery using cross-facility workflows that includes orchestration services, application programming interfaces (APIs), data access and co-scheduling.

Antypas, Katerina B.↗

MPI-FAUN: An MPI-Based Framework for Alternating-Updating Nonnegative Matrix Factorization

Non-negative matrix factorization (NMF) is the problem of determining two non-negative low rank factors W and H, for the given input matrix A, such that A≈WH. NMF is a useful tool for many applications in different domains such as topic modeling in text mining, background separation in video analysis, and community detection in social networks. Despite its popularity in the data mining community, there is a lack of efficient parallel algorithms to solve the problem for big data sets. The main contribution of this work is a new, high-performance parallel computational framework for a broad class of NMF algorithms that iteratively solves alternating non-negative least squares (NLS) subproblems for W and H. It maintains the data and factor matrices in memory (distributed across processors), uses MPI for interprocessor communication, and, in the dense case, provably minimizes communication costs (under mild assumptions). The framework is flexible and able to leverage a variety of NMF and NLS algorithms, including Multiplicative Update, Hierarchical Alternating Least Squares, and Block Principal Pivoting. Our implementation allows us to benchmark and compare different algorithms on massive dense and sparse data matrices of size that spans from few hundreds of millions to billions. We demonstrate the scalability of our algorithm and compare it with baseline implementations, showing significant performance improvements. The code and the datasets used for conducting the experiments are available online.

2D↗

Simulation-based characterization of the variability of earthquake risk to buildings in the near-field

Recent advancements in high performance computing platforms and computational workflow for regional-scale simulations are enabling unprecedented modeling of fault-to-structure earthquake processes. Regional simulations resolving ground motions at frequencies relevant to engineered systems are becoming computationally viable and provide a new capability to improve understanding of the geographical distribution and intensity of risk to buildings and critical infrastructure. As computational capabilities advance, it is essential to move beyond illustrative single rupture realizations for scenario earthquake events towards the development of a full suite of rupture realizations that appropriately characterize the range of risk to building systems. The work described in this article investigates the application of a suite of fault rupture realizations with the objective of assessing near-fault, site-specific seismic demand variability for building structures. A representative high-performance regional-scale computational model is utilized to execute ground motion and building response simulations based on 18 kinematic rupture realizations of an M7 strike-slip scenario earthquake. The fault rupture models for the scenario earthquake are created by systematically perturbing the hypocenter location and stochastically generating rupture parameters (slip, rise time, rake angle) to represent a breadth of ground motion intensities resulting from the spatial and temporal variabilities of an earthquake rupture process. The resulting seismic demand variability for three-story (short period) and forty-story (long period) steel moment-resisting frame buildings is characterized in terms of the median and distribution of peak inter-story drift ratio for a range of near-fault sites. The full suite of 18 fault rupture realizations and approximately 280,000 nonlinear dynamic building simulations indicate that the three-story building undergoes higher median seismic demand and significantly greater variability of demand at a given site than the forty-story building, which has important implications for the level of certainty in predicting building performance during an earthquake. The simulations performed provide deeper insight into the relationship between fault rupture parameterization and building response, which is essential information for developing a representative suite of rupture realizations for specific earthquake scenarios.

58 GEOSCIENCES↗

Compilation of Abstracts for SC12 Conference Proceedings

1 A Breakthrough in Rotorcraft Prediction Accuracy Using Detached Eddy Simulation; 2 Adjoint-Based Design for Complex Aerospace Configurations; 3 Simulating Hypersonic Turbulent Combustion for Future Aircraft; 4 From a Roar to a Whisper: Making Modern Aircraft Quieter; 5 Modeling of Extended Formation Flight on High-Performance Computers; 6 Supersonic Retropropulsion for Mars Entry; 7 Validating Water Spray Simulation Models for the SLS Launch Environment; 8 Simulating Moving Valves for Space Launch System Liquid Engines; 9 Innovative Simulations for Modeling the SLS Solid Rocket Booster Ignition; 10 Solid Rocket Booster Ignition Overpressure Simulations for the Space Launch System; 11 CFD Simulations to Support the Next Generation of Launch Pads; 12 Modeling and Simulation Support for NASA's Next-Generation Space Launch System; 13 Simulating Planetary Entry Environments for Space Exploration Vehicles; 14 NASA Center for Climate Simulation Highlights; 15 Ultrascale Climate Data Visualization and Analysis; 16 NASA Climate Simulations and Observations for the IPCC and Beyond; 17 Next-Generation Climate Data Services: MERRA Analytics; 18 Recent Advances in High-Resolution Global Atmospheric Modeling; 19 Causes and Consequences of Turbulence in the Earths Protective Shield; 20 NASA Earth Exchange (NEX): A Collaborative Supercomputing Platform; 21 Powering Deep Space Missions: Thermoelectric Properties of Complex Materials; 22 Meeting NASA's High-End Computing Goals Through Innovation; 23 Continuous Enhancements to the Pleiades Supercomputer for Maximum Uptime; 24 Live Demonstrations of 100-Gbps File Transfers Across LANs and WANs; 25 Untangling the Computing Landscape for Climate Simulations; 26 Simulating Galaxies and the Universe; 27 The Mysterious Origin of Stellar Masses; 28 Hot-Plasma Geysers on the Sun; 29 Turbulent Life of Kepler Stars; 30 Modeling Weather on the Sun; 31 Weather on Mars: The Meteorology of Gale Crater; 32 Enhancing Performance of NASAs High-End Computing Applications; 33 Designing Curiosity's Perfect Landing on Mars; 34 The Search Continues: Kepler's Quest for Habitable Earth-Sized Planets.

HPC↗

2008 ALCF annual report.

The word 'breakthrough' aptly describes the transformational science and milestones achieved at the Argonne Leadership Computing Facility (ALCF) throughout 2008. The number of research endeavors undertaken at the ALCF through the U.S. Department of Energy's (DOE) Innovative and Novel Computational Impact on Theory and Experiment (INCITE) program grew from 9 in 2007 to 20 in 2008. The allocation of computer time awarded to researchers on the Blue Gene/P also spiked significantly - from nearly 10 million processor hours in 2007 to 111 million in 2008. To support this research, we expanded the capabilities of Intrepid, an IBM Blue Gene/P system at the ALCF, to 557 teraflops (TF) for production use. Furthermore, we enabled breakthrough levels of productivity and capability in visualization and data analysis with Eureka, a powerful installation of NVIDIA Quadro Plex S4 external graphics processing units. Eureka delivered a quantum leap in visual compute density, providing more than 111 TF and more than 3.2 terabytes of RAM. On April 21, 2008, the dedication of the ALCF realized DOE's vision to bring the power of the Department's high performance computing to open scientific research. In June, the IBM Blue Gene/P supercomputer at the ALCF debuted as the world's fastest for open science and third fastest overall. No question that the science benefited from this growth and system improvement. Four research projects spearheaded by Argonne National Laboratory computer scientists and ALCF users were named to the list of top ten scientific accomplishments supported by DOE's Advanced Scientific Computing Research (ASCR) program. Three of the top ten projects used extensive grants of computing time on the ALCF's Blue Gene/P to model the molecular basis of Parkinson's disease, design proteins at atomic scale, and create enzymes. As the year came to a close, the ALCF was recognized with several prestigious awards at SC08 in November. We provided resources for Linear Scaling Divide-and-Conquer Electronic Structure Calculations for Thousand Atom Nanostructures, a collaborative effort between Argonne, Lawrence Berkeley National Laboratory, and Oak Ridge National Laboratory that received the ACM Gordon Bell Prize Special Award for Algorithmic Innovation. The ALCF also was named a winner in two of the four categories in the HPC Challenge best performance benchmark competition.

59 BASIC BIOLOGICAL SCIENCES↗

A High-level Design for Bidirectional Data Streaming to High-Performance Computing Systems from External Science Facilities

Cutting-edge science is increasingly data-driven due to the emergence of scientific machine learning models that can guide scientists toward fruitful areas of exploration. Experimental science facilities such as light and neutron sources, particle colliders, and radio astronomy telescopes are also producing raw measurement data at rates that exceed available data storage and computing capacity at those facilities. As a result, scientific workflows are being developed that concurrently couple experiments at science facilities with high-performance computing (HPC) facilities to enable analysis of experimental data while the experiment is ongoing, and where analysis results are potentially fed back to the experiment for use in experimental control and/or steering in a time-sensitive manner. Our goal is to design, prototype, and deploy a new capability for the Oak Ridge Leadership Computing Facility (OLCF) that enables such workflows through support for bidirectional, memory-based streaming of data from external experiments into and out of OLCF HPC systems. This high-level design document describes the related work and motivating use cases that inform our understanding of the technical requirements for this capability, and describes a proposed architectural solution that meets these requirements and our plans for demonstrating the capability.

97 MATHEMATICS AND COMPUTING↗

SharP

The pre-exascale systems are expected to have a significant amount of hierarchical and heterogeneous on-node memory, and this trend of system architecture in extreme-scale systems is expected to continue into the exascale era. along with hierarchical-heterogeneous memory, the system typically has a high-performing network ad a compute accelerator. This system architecture is not only effective for running traditional High Performance Computing (HPC) applications (Big-Compute), but also for running data-intensive HPC applications and Big-Data applications. As a consequence, there is a growing desire to have a single system serve the needs of both Big-Compute and Big-Data applications. Though the system architecture supports the convergence of the Big-Compute and Big-Data, the programming models and software layer have yet to evolve to support either hierarchical-heterogeneous memory systems or the convergence. A programming abstraction to address this problem. The programming abstraction is implemented as a software library and runs on pre-exascale and exascale systems supporting current and emerging system architecture. Using distributed data-structures as a central concept, it provides (1) a simple, usable, and portable abstraction for hierarchical-heterogeneous memory and (2) a unified programming abstraction for Big-Compute and Big-Data applications.

Venkata, Manjunath Gorentla↗

Improving Performance of M-to-N Processing and Data Redistribution in In Transit Analysis and Visualization

In an in transit setting, a parallel data producer, such as a numerical simulation, runs on one set of ranks M, while a data consumer, such as a parallel visualization application, runs on a different set of ranks N. One of the central challenges in this in transit setting is to determine the mapping of data from the set of M producer ranks to the set of N consumer ranks. This is a challenging problem for several reasons, such as the producer and consumer codes potentially having different scaling characteristics and different data models. The resulting mapping from M to N ranks can have a significant impact on aggregate application performance. In this work, we present an approach for performing this M-to-N mapping in a way that has broad applicability across a diversity of data producer and consumer applications. We evaluate its design and performance with a study that runs at high concurrency on a modern HPC platform. By leveraging design characteristics, which facilitate an “intelligent” mapping from M-to-N, we observe significant performance gains are possible in terms of several different metrics, including time-to-solution and amount of data moved.

Loring, Burlen↗

Geomancy: Automated Performance Enhancement through Data Layout Optimization

Large distributed storage systems such as high- performance computing (HPC) systems used by national or international laboratories require sufficient performance and scale for demanding scientific workloads and must handle shifting workloads with ease. Ideally, data is placed in locations to optimize performance, but the size and complexity of large storage systems inhibit rapid effective restructuring of data layouts to maintain performance as workloads shift. To address these issues, we have developed Geomancy, a tool that models the placement of data within a distributed storage system and reacts to drops in performance. Using a combination of machine learning techniques suitable for temporal modeling, Geomancy determines when and where a bottleneck may happen due to changing workloads and suggests changes in the layout that mitigate or prevent them. Our approach to optimizing throughput offers benefits for storage systems such as avoiding potential bottlenecks and increasing overall I/O throughput from 11% to 30%.

Bel, Oceane M.↗

Improving Performance of M-to-N Processing and Data Redistribution in In Transit Analysis and Visualization

In an in transit setting, a parallel data producer, such as a numerical simulation, runs on one set of ranks M, while a data consumer, such as a parallel visualization application, runs on a different set of ranks N: One of the central challenges in this in transit setting is to determine the mapping of data from the set of M producer ranks to the set of N consumer ranks. This is a challenging problem for several reasons, such as the producer and consumer codes potentially having different scaling characteristics and different data models. The resulting mapping from M to N ranks can have a significant impact on aggregate application performance. In this work, we present an approach for performing this M-to-N mapping in a way that has broad applicability across a diversity of data producer and consumer applications. We evaluate its design and performance with a study that runs at high concurrency on a modern HPC platform. By leveraging design characteristics, which facilitate an ''intelligent'' mapping from M-to-N, we observe significant performance gains are possible in terms of several different metrics, including time-to-solution and amount of data moved.

Loring, Burlen↗

Landrush: Rethinking In-Situ Analysis for GPGPU Workflows

In-situ analysis on the output data of scientific simulations has been made necessary by ever-growing output data volumes and increasing costs of data movement as supercomputing is moving towards exascale. With hardware accelerators like GPUs becoming increasingly common in high end machines, new opportunities arise to co-locate scientific simulations and online analysis performed on the scientific data generated by the simulations. However, the asynchronous nature of GPGPU programming models and the limited context-switching capabilities on the GPU pose challenges to co-locating the scientific simulation and analysis on the same GPU. This paper dives deeper into these challenges to understand how best to co-locate analysis with scientific simulations on the GPUs in HPC clusters. Specifically, our 'Landrush' approach to GPU sharing proposes a solution that utilizes idle cycles on the GPU to provide an improved time-to-answer, that is, the total time to run the scientific simulation and analysis of the generated data. Landrush is demonstrated with experimental results obtained from leadership high-end applications on ORNL's Titan supercomputer, which show that (i) GPU-based scientific simulations have varying degrees of idle cycles to afford useful analysis task co-location, and (ii) the inability to context switch on the GPU at instruction granularity can be overcome by careful control of the analysis kernel launches and software-controlled early completion of analysis kernel executions. Results show that Landrush is superior in terms of time-to-answer compared to serially running simulations followed by analysis or by relying on the GPU driver and hardwired thread dispatcher to run analysis concurrently on a single GPU.

Computer Science↗

Acceleration of the IMplicit–EXplicit nonhydrostatic unified model of the atmosphere on manycore processors

We present the acceleration of an IMplicit-EXplicit (IMEX) nonhydrostatic atmospheric model on manycore processors such as graphic processing units (GPUs) and Intel's Many Integrated Core (MIC) architecture. IMEX time integration methods sidestep the constraint imposed by the Courant-Friedrichs-Lewy condition on explicit methods through corrective implicit solves within each time step. In this work, we implement and evaluate the performance of IMEX on manycore processors relative to explicit methods. Using 3D-IMEX at Courant number C = 15, we obtained a speedup of about 4x relative to an explicit time stepping method run with the maximum allowable C = 1. Moreover, the unconditional stability of IMEX with respect to the fast waves means the speedup can increase significantly with the Courant number as long as the accuracy of the resulting solution is acceptable. We show a speedup of 100x at C = 150 using 1D-IMEX to demonstrate this point. Several improvements on the IMEX procedure were necessary in order to outperform our results with explicit methods: (a) reducing the number of degrees of freedom of the IMEX formulation by forming the Schur complement, (b) formulating a horizontally explicit vertically implicit 1D-IMEX scheme that has a lower workload and better scalability than 3D-IMEX, (c) using high-order polynomial preconditioners to reduce the condition number of the resulting system, and (d) using a direct solver for the 1D-IMEX method by performing and storing LU factorizations once to obtain a constant cost for any Courant number. Without all of these improvements, explicit time integration methods turned out to be difficult to beat. We discuss in detail the IMEX infrastructure required for formulating and implementing efficient methods on manycore processors. Several parametric studies are conducted to demonstrate the gain from each of the abovementioned improvements. Lastly, we validate our results with standard benchmark problems in numerical weather prediction and evaluate the performance and scalability of the IMEX method using up to 4192 GPUs and 16 Knights Landing processors.

97 MATHEMATICS AND COMPUTING↗

RAJA - LLNL HPC Architecture Portability Encapsulation Layer Version 1.0

RAJA is a collection of C++ software abstractions designed to make HPC applications portable across a range of hardware architectures and programming models. The main conceptual abstraction in FAJA is a loop. RAJA provides abstractions for types, loop execution and scheduling, and loop iteration space portioning and ordering cooperate to insulate algorithms in application code that domain experts write from implementations details of those algorithms that may need to depend on the underlying computer hardware for high performance.

ECP↗

RAJA - LLNL HPC Architecture Portability Encapsulation Layer Version 1.0

RAJA is a collection of C++ software abstractions designed to make HPC applications portable across a range of hardware architectures and programming models. The main conceptual abstraction in FAJA is a loop. RAJA provides abstractions for types, loop execution and scheduling, and loop iteration space portioning and ordering cooperate to insulate algorithms in application code that domain experts write from implementations details of those algorithms that may need to depend on the underlying computer hardware for high performance.

Hornung, Richard↗

Hyperspectral Aquatic Radiative Transfer Modeling Using a High-Performance Cluster Computing-Based Approach

Abstract For aquatic studies, radiative transfer (RT) modeling can be used to compute hyperspectral above-surface remote sensing reflectance that can be utilized for inverse model development. Inverse models can provide bathymetry and inherent-and bottom-optical property estimation. Because measured oceanic field/organic datasets are often spatio-temporally sparse, synthetic data generation is useful in yielding sufficiently large datasets for inversion model development; however, these forward-modeled data are computationally expensive and time-consuming to generate. This study establishes the magnitude of wall-clock-time savings achieved for performing large, aquatic RT batch-runs using parallel computing versus a sequential approach. Given 2,600 simulations and identical compute-node characteristics, sequential architecture required ~100 hours until termination, whereas a parallel approach required only ~2.5 hours (42 compute nodes) a 40x speed-up. Tools developed for this parallel execution are discussed.

58 GEOSCIENCES↗