Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “parallel, scaling”

Search indexed NASA NTRS and DOE OSTI research on propulsion, heat transfer, battery materials and energy systems. Follow report and document links to the original sources.

Quote a phrase for an exact phrase match. Source license links do not imply unrestricted reuse.

At least 1,711 records · Page 95

Use Computer-Aided Tools to Parallelize Large CFD Applications

Porting applications to high performance parallel computers is always a challenging task. It is time consuming and costly. With rapid progressing in hardware architectures and increasing complexity of real applications in recent years, the problem becomes even more sever. Today, scalability and high performance are mostly involving handwritten parallel programs using message-passing libraries (e.g. MPI). However, this process is very difficult and often error-prone. The recent reemergence of shared memory parallel (SMP) architectures, such as the cache coherent Non-Uniform Memory Access (ccNUMA) architecture used in the SGI Origin 2000, show good prospects for scaling beyond hundreds of processors. Programming on an SMP is simplified by working in a globally accessible address space. The user can supply compiler directives, such as OpenMP, to parallelize the code. As an industry standard for portable implementation of parallel programs for SMPs, OpenMP is a set of compiler directives and callable runtime library routines that extend Fortran, C and C++ to express shared memory parallelism. It promises an incremental path for parallel conversion of existing software, as well as scalability and performance for a complete rewrite or an entirely new development. Perhaps the main disadvantage of programming with directives is that inserted directives may not necessarily enhance performance. In the worst cases, it can create erroneous results. While vendors have provided tools to perform error-checking and profiling, automation in directive insertion is very limited and often failed on large programs, primarily due to the lack of a thorough enough data dependence analysis. To overcome the deficiency, we have developed a toolkit, CAPO, to automatically insert OpenMP directives in Fortran programs and apply certain degrees of optimization. CAPO is aimed at taking advantage of detailed inter-procedural dependence analysis provided by CAPTools, developed by the University of Greenwich, to reduce potential errors made by users. Earlier tests on NAS Benchmarks and ARC3D have demonstrated good success of this tool. In this study, we have applied CAPO to parallelize three large applications in the area of computational fluid dynamics (CFD): OVERFLOW, TLNS3D and INS3D. These codes are widely used for solving Navier-Stokes equations with complicated boundary conditions and turbulence model in multiple zones. Each one comprises of from 50K to 1,00k lines of FORTRAN77. As an example, CAPO took 77 hours to complete the data dependence analysis of OVERFLOW on a workstation (SGI, 175MHz, R10K processor). A fair amount of effort was spent on correcting false dependencies due to lack of necessary knowledge during the analysis. Even so, CAPO provides an easy way for user to interact with the parallelization process. The OpenMP version was generated within a day after the analysis was completed. Due to sequential algorithms involved, code sections in TLNS3D and INS3D need to be restructured by hand to produce more efficient parallel codes. An included figure shows preliminary test results of the generated OVERFLOW with several test cases in single zone. The MPI data points for the small test case were taken from a handcoded MPI version. As we can see, CAPO's version has achieved 18 fold speed up on 32 nodes of the SGI O2K. For the small test case, it outperformed the MPI version. These results are very encouraging, but further work is needed. For example, although CAPO attempts to place directives on the outer- most parallel loops in an interprocedural framework, it does not insert directives based on the best manual strategy. In particular, it lacks the support of parallelization at the multi-zone level. Future work will emphasize on the development of methodology to work in a multi-zone level and with a hybrid approach. Development of tools to perform more complicated code transformation is also needed.

Jin, H.↗

Filament Extension Atomization for High Solids Loading in Energy Efficient Spray Drying Systems

We demonstrate that we could scale FEA to reach outputs needed by industrial production, while increasing solids loading of the sprayed product by at least 30% and maintaining equal or better spray powder. After testing a wide range of products, in collaboration with industry partners we decided on our primary spray products of dry whey and WPC-80, two common materials processed and sold by US manufacturers with different parameters. We sprayed these with FEA at solids loadings of 70% for dry whey and 45% for WPC-80 with a spray output with particle sizes similar to industrial particles sizes and reduced variation in particle size. We simultaneously scaled up FEA first with a multi-nip with 6 nips surrounding a central roller with parallel axis of rotation and eventually with a tapered design that solved problems we encountered with our initial design. We were able to achieve output from a single array from our first design of up to 4.7 liters per minute (L/min) and from an array of our second multi-nip of 8 L/min exceeding expectations. This demonstrates that FEA technology can indeed be scaled up to meet the needs of industrial production. More arrays can be added as necessary to meet a wide range of spray dryer designs. We also tested FEA to create dried powders from a small scale (10 L/hour of water removal) spray dryer. Though we were not able to produce large quantities of powder from FEA due to challenges in integration, the powder we produced was higher quality and produced from higher solids loading materials. From our technoeconomic analysis we for a typically sized spray dryer, we estimate a 27% cost reduction and 41% energy and carbon reduction for WPC-80 and a 39-57% cost reduction and 52-76% energy reduction for sweet dry whey (depending on the exact product).

32 ENERGY CONSERVATION, CONSUMPTION, AND UTILIZATI↗

Single Molecule Junctions: Probing Contact Chemistry and Fundamental Circuit Laws

By exploiting selective link chemistry, formation of single molecule junctions with reproducible conductance has become established. Systematic studies reveal the structure-conductance relationships for diverse molecules. I will draw on experiments from my collaborators at Columbia University, atomic-scale calculations and theory to describe progress in two areas. First, I will describe a novel route to form single molecule junctions, based on SnMe3 terminated molecules, in which gold directly bonds to carbon in the molecule backbone resulting in near ideal contact resistance [1]. Second, comparison of the conductance of junctions formed with molecular species containing either one backbone or two backbones in parallel allows demonstration of the role of quantum interference in the conductance superposition law at the molecular scale [2].

36 MATERIALS SCIENCE↗

Mirror force induced wave dispersion in Alfvén waves

Recent hybrid MHD-kinetic electron simulations of global scale standing shear Alfvén waves along the Earth's closed dipolar magnetic field lines show that the upward parallel current region within these waves saturates and broadens perpendicular to the ambient magnetic field and that this broadening increases with the electron temperature. Using resistive MHD simulations, with a parallel Ohm's law derived from the linear Knight relation (which expresses the current-voltage relationship along an auroral field line), we explore the nature of this broadening in the context of the increased perpendicular Poynting flux resulting from the increased parallel electric field associated with mirror force effects. This increased Poynting flux facilitates wave energy dispersion across field lines which in-turn allows for electron acceleration to carry the field aligned current on adjacent field lines. This mirror force driven dispersion can dominate over that associated with electron inertial effects for global scale waves.

70 PLASMA PHYSICS AND FUSION TECHNOLOGY↗

High Performance Distributed Computing in a Supercomputer Environment: Computational Services and Applications Issues

This tutorial proposes to be a practical guide for the uninitiated to the main topics and themes of high-performance computing (HPC), with particular emphasis to distributed computing. The intent is first to provide some guidance and directions in the rapidly increasing field of scientific computing using both massively parallel and traditional supercomputers. Because of their considerable potential computational power, loosely or tightly coupled clusters of workstations are increasingly considered as a third alternative to both the more conventional supercomputers based on a small number of powerful vector processors, as well as high massively parallel processors. Even though many research issues concerning the effective use of workstation clusters and their integration into a large scale production facility are still unresolved, such clusters are already used for production computing. In this tutorial we will utilize the unique experience made at the NAS facility at NASA Ames Research Center. Over the last five years at NAS massively parallel supercomputers such as the Connection Machines CM-2 and CM-5 from Thinking Machines Corporation and the iPSC/860 (Touchstone Gamma Machine) and Paragon Machines from Intel were used in a production supercomputer center alongside with traditional vector supercomputers such as the Cray Y-MP and C90.

Kramer, Williams T. C.↗

Characterizing Output Bottlenecks of a Production Supercomputer: Analysis and Implications

This article studies the I/O write behaviors of the Titan supercomputer and its Lustre parallel file stores under production load. The results can inform the design, deployment, and configuration of file systems along with the design of I/O software in the application, operating system, and adaptive I/O libraries.We propose a statistical benchmarking methodology to measure write performance across I/O configurations, hardware settings, and system conditions. Moreover, we introduce two relative measures to quantify the write-performance behaviors of hardware components under production load. In addition to designing experiments and benchmarking on Titan, we verify the experimental results on one real application and one real application I/O kernel, XGC and HACC IO, respectively. These two are representative and widely used to address the typical I/O behaviors of applications.In summary, we find that Titan’s I/O system is variable across the machine at fine time scales. This variability has two major implications. First, stragglers lessen the benefit of coupled I/O parallelism (striping). Peak median output bandwidths are obtained with parallel writes to many independent files, with no striping or write sharing of files across clients (compute nodes). I/O parallelism is most effective when the application—or its I/O libraries—distributes the I/O load so that each target stores files for multiple clients and each client writes files on multiple targets in a balanced way with minimal contention. Second, our results suggest that the potential benefit of dynamic adaptation is limited. In particular, it is not fruitful to attempt to identify “good locations” in the machine or in the file system: component performance is driven by transient load conditions and past performance is not a useful predictor of future performance. For example, we do not observe diurnal load patterns that are predictable.

97 MATHEMATICS AND COMPUTING↗

RingX: Scalable Parallel Attention for Long-Context Learning on HPC

The attention mechanism has become foundational for remarkable AI breakthroughs since the introduction of the Transformer, driving the demand for increasingly longer context to power frontier models such as large-scale reasoning language models and high-resolution image/video generators. However, its quadratic computational and memory complexities present substantial challenges. Current state-of-the-art parallel attention methods, such as ring attention, are widely adopted for long-context training but utilize a point-to-point communication strategy that fails to fully exploit the capabilities of modern HPC network architectures. In this work, we propose ringX, a scalable family of parallel attention methods optimized explicitly for HPC systems. By enhancing workload partitioning, refining communication patterns, and improving load balancing, ringX achieves up to 3.4 × speedup compared to conventional ring attention on the Frontier supercomputer. Optimized for both bi-directional and causal attention mechanisms, ringX demonstrates its effectiveness through training benchmarks of a Vision Transformer (ViT) on a climate dataset and a Generative Pre-Trained Transformer (GPT) model, Llama3 8B. Our method attains an end-to-end training speedup of approximately 1.5 × in both scenarios. To our knowledge, the achieved 38% model FLOPs utilization (MFU) for training Llama3 8B with a 1M-token sequence length on 4,096 GPUs represents one of the highest training efficiencies reported for long-context learning on HPC systems. Our code implementation is available at https://github.com/jqyin/ringX-attention.

Yin, Junqi [ORNL] (ORCID:0000000338435520)↗

Feasibility Study for a Hopi Utility-Scale Wind Project

The goal of this project was to investigate the feasibility for the generation of energy from wind and to parallel this work with the development of a tribal utility organization capable of undertaking potential joint ventures in utility businesses and projects on the Hopi reservation. The goal of this project was to investigate the feasibility for the generation of energy from wind and to parallel this work with the development of a tribal utility organization capable of undertaking potential joint ventures in utility businesses and projects on the Hopi reservation. Wind resource assessments were conducted at two study sites on Hopi fee simple lands located south of the city of Winslow. Reports from the study were recently completed and have not been compared to any existing historical wind data nor have they been processed under any wind assessment models to determine the output performance and the project economics of turbines at the wind study sites. Ongoing analysis of the wind data and project modeling will determine the feasibility of a tribal utility-scale wind energy generation.

17 WIND ENERGY↗

An Asynchronous Bundle-Trust-Region Method for Dual Decomposition of Stochastic Mixed-Integer Programming

Here, we present an asynchronous bundle-trust-region algorithm within the context of Lagrangian dual decomposition for stochastic mixed-integer programs. The approach solves the Lagrangian master problem by using a bundle method with a trust-region constraint. This scheme enables asynchronous computations and can thus help mitigate severe load imbalance issues (associated with the solution of scenario subproblems) and improve parallel efficiency. We provide a convergence analysis and an implementation of the proposed scheme. We also present extensive numerical results on eighty instances of a large-scale stochastic unit commitment problem, and demonstrate that the proposed approach provides significant reductions in solution time and achieves strong scaling.

97 MATHEMATICS AND COMPUTING↗

Introduction to Big Earth Data Applications

Climate and weather modeling generate enormous volumes that make iterative analysis challenging, spurring the development of new ways to work with the data. A theme going across applications is the need to identify and highlight "interesting" data for the scientist to focus on. Operational applications often scale up from small, local studies to larger spatial scales with more analysis targets.

parallel processing (computers)↗

High-Performance Computing Optimization for Aladyn – Adaptive Neural Network Molecular Dynamics Mini-Application

This report provides a description and performance evaluation of the optimization techniques for high performance computing (HPC) implementation of the open source Computational Materials mini-application Aladyn (https://github.com/nasa/aladyn). Aladyn is a basic molecular dynamics code written in FORTRAN 2003, which is designed to demonstrate the use of adaptive neural networks (ANNs) in atomistic simulations. The role of ANNs is to efficiently reproduce the very complex energy landscape resulting from the atomic interactions in materials with the accuracy of the more expensive quantum mechanics-based calculations. The ANN is trained on a large set of atomic structures calculated using the density functional theory (DFT) method. While achieving orders of magnitude faster computational performance than DFT, the ANN-based approach was still very computationally demanding compared to the conventional approach of using empirically fitted energy functions. After its initial development, Aladyn was evaluated and optimized by experts at the NASA Advanced Supercomputing (NAS) division to exploit modern supercomputer architectures. The code has been optimized for execution on multicore central processing units (CPUs), including Intel® Skylake microarchitecture, and on graphic accelerators, such as Nvidia® V100 graphic processing units (GPUs), using Open Multi-Processing (OpenMP) and Open Accelerators (OpenACC) programming interfaces. The optimization achieved a speedup of 4.7 times the baseline version on CPU performance and an additional 2.4 times on CPU+GPU performance. Atomistic computer simulations are a fundamental tool in materials research to model material properties form physics-based first principles. Atomic interaction, governed by Quantum Mechanics (QM) require sophisticated and highly computationally demanding mathematical models to calculate [1]. Classical methods use approximate functional forms, empirically fitted through a set of variable parameters to emulate atomic energies as direct functions of atomic coordinates [2]. While empirical potentials are computationally much simpler, allowing simulations of large-scale systems of up to a trillion (1012) atoms [3], they are substantially less accurate compared to quantum calculations and applicable only to very specific atomic configurations or predefined crystallographic phases. A recently suggested approach is to use heuristic machine learning methods [4], such as those based on Adaptive Neural Networks (ANNs) to predict atomic energies, after being trained on a sufficiently large database of QM-calculated structures [5,6]. This approach reduces significantly the computational complexity, allowing for simulations of orders of magnitude larger systems compared to QM-based methods without compromising accuracy. Still, compared to classical methods using empirical energy functions, ANN methods remain two- to three orders of magnitude more computationally demanding. Hence, the computational cost of simulations, together with the need for extensive training of ANNs, still makes the practical implementation of ANN-based methods quite challenging. The purpose of the Aladyn mini-application software [7], available as open source at https://github.com/nasa/aladyn, is to be a testbed for exploring possible optimization strategies to develop highly scalable parallel algorithms for ANN-based atomistic simulations. Aladyn is aimed at utilizing the architecture of the high-end modern highperformance computing (HPC) hardware based on multicore central processing units (CPUs) equipped with graphic processing unit (GPU) accelerators. Specifically, the goal is to optimize the performance on a single HPC compute node, before implementing scaling to multi-node parallelization using message passing interface (MPI). At the same time, the open source code of Aladyn can serve as a training model for students and professors in academia.

Yamakov, Vesselin I.↗

The NAS parallel benchmarks

A new set of benchmarks has been developed for the performance evaluation of highly parallel supercomputers in the framework of the NASA Ames Numerical Aerodynamic Simulation (NAS) Program. These consist of five 'parallel kernel' benchmarks and three 'simulated application' benchmarks. Together they mimic the computation and data movement characteristics of large-scale computational fluid dynamics applications. The principal distinguishing feature of these benchmarks is their 'pencil and paper' specification-all details of these benchmarks are specified only algorithmically. In this way many of the difficulties associated with conventional benchmarking approaches on highly parallel systems are avoided.

Bailey, D. H.↗

PIC simulations of the effect of velocity space instabilities on electron viscosity and thermal conduction

In low-collisionality plasmas, velocity-space instabilities are a key mechanism providing an effective collisionality for the plasma. We use particle-in-cell (PIC) simulations to study the interplay between electron- and ion-scale velocity-space instabilities and their effect on electron pressure anisotropy, viscous heating, and thermal conduction. The adiabatic invariance of the magnetic moment in low-collisionality plasmas leads to pressure anisotropy, Δp{sub j}≡p{sub ⊥,j}−p{sub ∥,j}>0, if the magnetic field B is amplified (p{sub ⊥,j} and p{sub ∥,j} denote the pressure of species j (electron, ion) perpendicular and parallel to B). If the resulting anisotropy is large enough, it can in turn trigger small-scale plasma instabilities. Our PIC simulations explore the nonlinear regime of the mirror, IC, and electron whistler instabilities, through continuous amplification of the magnetic field |B| by an imposed shear in the plasma. In the regime 1≲β{sub j}≲20 (β{sub j}≡8πp{sub j}/|B|{sup 2}), the saturated electron pressure anisotropy, Δp{sub e}/p{sub ∥,e}, is determined mainly by the (electron-lengthscale) whistler marginal stability condition, with a modest factor of ∼1.5–2 decrease due to the trapping of electrons into ion-lengthscale mirrors. We explicitly calculate the mean free path of the electrons and ions along the mean magnetic field and provide a simple physical prescription for the mean free path and thermal conductivity in low-collisionality β {sub j} ≳ 1 plasmas. Our results imply that velocity-space instabilities likely decrease the thermal conductivity of plasma in the outer parts of massive, hot, galaxy clusters. We also discuss the implications of our results for electron heating and thermal conduction in low-collisionality accretion flows onto black holes, including Sgr A* in the Galactic Center.

79 ASTRONOMY AND ASTROPHYSICS↗

Reducing Communication in Graph Neural Network Training

Graph Neural Networks (GNNs) are powerful and flexible neural networks that use the naturally sparse connectivity information of the data. GNNs represent this connectivity as sparse matrices, which have lower arithmetic intensity and thus higher communication costs compared to dense matrices, making GNNs harder to scale to high concurrencies than convolutional or fully-connected neural networks. Here, we introduce a family of parallel algorithms for training GNNs and show that they can asymptotically reduce communication compared to previous parallel GNN training methods. We implement these algorithms, which are based on 1D, 1. 5D, 2D, and 3D sparse-dense matrix multiplication, using torch.distributed on GPU-equipped clusters. Our algorithms optimize communication across the full GNN training pipeline. We train GNNs on over a hundred GPUs on multiple datasets, including a protein network with over a billion edges.

97 MATHEMATICS AND COMPUTING↗

The organic redox transistor for neuromorphic computing

Inspired by the in-memory computing architectures of biological systems, neuromorphic computing using crossbar arrays of artificial synapses based on non-volatile memory (NVM) devices with variable conductances has emerged as a new paradigm to enable massively parallel and ultra-low power computing hardware for data centric applications. Although inference has been demonstrated successfully using crossbars based on a variety of NMV technologies, efficient learning and scaling to large arrays (>10 6 elements) remains a challenge due to the synaptic elements' non-ideal electrical characteristics which degrades ANN accuracy. A further challenge is that in the conductive state memristors draw large currents >μA resulting in significant voltage drops in the interconnect wires and increased probability of failure in scaled arrays. We suggest the organic polymer redox transistor (RT) is an alternate approach that could solve many of these challenges, enabling both inference and parallel outer product updates, as recently demonstrated by Fuller et al. An RT consists of redox-active channel and gate electrodes in contact with a liquid or solid electrolyte. lon insertion through the electrolyte controls the channel electronic conductivity, while electron transfer through an external circuit maintains overall charge neutrality. Unlike a rechargeable battery, in the RT the voltage built-up across the electrolyte is kept to a minimum (typically <100 mV) by using the same material for the gate and channel. Elimination of the voltage offset simplifies integration of the RT into programmable arrays by enabling the use of various selectors. RTs based on inorganic and organic materials have been recently demonstrated with conductance tuning occurring at potentials of just a few mV and hundreds to thousands of linearly and symmetrically programmable conductance states, enabling near ideal accuracy in neural network simulations. Introduced in the 1980's, redox transistors with metallic gate electrodes and organic channel materials, also known as organic electrochemical transistors (OECTs), have been explored for a variety of applications such as chem- and bio-sensing, neural interfaces, and low cost printed circuits. A typical channel material for OECTs is the conducting polymer poly(3,4-ethylenedioxythiophene) doped with poly(styrene sulfonate) (PEDOT:PSS). PEDOT is a p-type semiconducting polymer with mobile positively charged polarons that hop chain-to-chain.

Talin, Albert Alec↗

A performance analysis of ensemble averaging for high fidelity turbulence simulations at the strong scaling limit

We analyze the potential performance benefits of estimating expected quantities in large eddy simulations of turbulent flows using true ensembles rather than ergodic time averaging. Multiple realizations of the same flow are simulated in parallel, using slightly perturbed initial conditions to create unique instantaneous evolutions of the flow field. Each realization is then used to calculate statistical quantities. Provided each instance is sufficiently de-correlated, this approach potentially allows considerable reduction in the time to solution beyond the strong scaling limit for a given accuracy. This study focuses on the theory and implementation of the methodology in Nek5000, a massively parallel open-source spectral element code.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗

A scalable framework for the global offline community land model ensemble simulation

Current earth system models have a large range of uncertainty, owing to differences in the simulation of feedbacks and insufficient information to constrain model parameters. Parameter disturbance experiment provides a straightforward method to quantify the variation (uncertainty) outputs caused by model inputs. Owing to the software complexity and computational intensity of earth system models, a large-scale simulation framework is needed to support ensemble simulation required by parameter disturbance experiment. This paper presents a parallel framework for the community land model ensemble simulation. After a software structure review of the community land model simulation, a single factor parameter disturbance experiment of a reference computational experiment design is used to demonstrate the software design principles, computational characteristics of individual application, parallel ensemble simulation implementation, as well as the weak scalability of this simulation framework on a high-end computer. Lastly, the paper discusses some preliminary diagnostic analysis results of the single factor parameter disturbance experiments. The framework design considerations and implementation details described in this paper can be beneficial to many other research programmes involving large scale, legacy modelling system.

58 GEOSCIENCES↗

Resource Management for Distributed Parallel Systems

Multiprocessor systems should exist in the the larger context of distributed systems, allowing multiprocessor resources to be shared by those that need them. Unfortunately, typical multiprocessor resource management techniques do not scale to large networks. The Prospero Resource Manager (PRM) is a scalable resource allocation system that supports the allocation of processing resources in large networks and multiprocessor systems. To manage resources in such distributed parallel systems, PRM employs three types of managers: system managers, job managers, and node managers. There exist multiple independent instances of each type of manager, reducing bottlenecks. The complexity of each manager is further reduced because each is designed to utilize information at an appropriate level of abstraction.

Neuman, B. Clifford↗