Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “Scalable Algorithms”

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 775 records · Page 43

GraphReduce: Processing Large-Scale Graphs on Accelerator-Based Systems

Recent work on real-world graph analytics has sought to leverage the massive amount of parallelism offered by GPU devices, but challenges remain due to the inherent irregularity of graph algorithms and limitations in GPU-resident memory for storing large graphs. We present GraphReduce, a highly efficient and scalable GPU-based framework that operates on graphs that exceed the device’s internal memory capacity. GraphReduce adopts a combination of edge- and vertex-centric implementations of the Gather-Apply-Scatter programming model and operates on multiple asynchronous GPU streams to fully exploit the high degrees of parallelism in GPUs with efficient graph data movement between the host and device.

Architecture optimization, reuse, performance, ene↗

Determination of Experimental Fuel Rod Parameters using 3D Modelling of PCMI with MPS Defect

An in-reactor experiment is being designed in order to validate the pellet-cladding mechanical interaction (PCMI) behavior of the BISON fuel performance code. The experimental parameters for the test rod being placed in the Halden Research Reactor are being determined using BISON simulations. The 3D model includes a missing pellet surface (MPS) defect to generate large local cladding deformations, which should be measureable after typical burnup times. The BISON fuel performance code is being developed at Idaho National Laboratory (INL) and is built on the Multiphysics Object-Oriented Simulation Environment (MOOSE) framework. BISON supports both 2D and 3D finite elements and solves the fully coupled equations for solid mechanics, heat conduction and species diffusion. A number of fuel performance effects are included using models for swelling, densification, creep, relocation and fission gas production & release. In addition, the mechanical and thermal contact between the fuel and cladding is explicitly modelled using a master-slave based contact algorithm. In order to accurately predict PCMI effects, the BISON code includes the relevant physics involved and provides a scalable and robust solution procedure. The depth of the proposed MPS defect is being varied in the BISON model to establish an optimum value for the experiment. The experiment will be interrupted approximately every 6 months to measure cladding radial deformation and provide data to validate BISON. The complete rodlet (~20 discrete pellets) is being simulated using a 180° half symmetry 3D model with MPS defects at two axial locations. In addition, annular pellets will be used at the top and bottom of the pellet stack to allow thermocouples within the rod to measure the fuel centerline temperature. Simulation results will be presented to illustrate the expected PCMI behavior and support the chosen experimental design parameters.

BISON↗

Analysis of Experimental Fuel Rod Parameters using 3D Modelling of PCMI with MPS Defect

An in-reactor experiment is being designed in order to validate the pellet-cladding mechanical interaction (PCMI) behavior of the BISON fuel performance code. The experimental parameters for the test rod being placed in the Halden Research Reactor are being determined using BISON simulations. The 3D model includes a missing pellet surface (MPS) defect to generate large local cladding deformations, which should be measureable after typical burnup times. The BISON fuel performance code is being developed at Idaho National Laboratory (INL) and is built on the Multiphysics Object-Oriented Simulation Environment (MOOSE) framework. BISON supports both 2D and 3D finite elements and solves the fully coupled equations for solid mechanics, heat conduction and species diffusion. A number of fuel performance effects are included using models for swelling, densification, creep, relocation and fission gas production & release. In addition, the mechanical and thermal contact between the fuel and cladding is explicitly modelled using a master-slave based contact algorithm. In order to accurately predict PCMI effects, the BISON code includes the relevant physics involved and provides a scalable and robust solution procedure. The depth of the proposed MPS defect is being varied in the BISON model to establish an optimum value for the experiment. The experiment will be interrupted approximately every 6 months to measure cladding radial deformation and provide data to validate BISON. The complete rodlet (~20 discrete pellets) is being simulated using a 180° half symmetry 3D model with MPS defects at two axial locations. In addition, annular pellets will be used at the top and bottom of the pellet stack to allow thermocouples within the rod to measure the fuel centerline temperature. Simulation results will be presented to illustrate the expected PCMI behavior and support the chosen experimental design parameters.

BISON↗

APHiD: Hierarchical Task Placement to Enable a Tapered Fat Tree Topology for Lower Power and Cost in HPC Networks

The power and procurement cost of bandwidth in system-wide networks has forced a steady drop in the byte/flop ratio. This trend of computation becoming faster relative to the network is expected to hold. In this paper, we explore how cost-oriented task placement enables reducing the cost of system-wide networks by enabling high performance even on tapered topologies where more bandwidth is provisioned at lower levels. We describe APHiD, an efficient hierarchical placement algorithm that uses new techniques to improve the quality of heuristic solutions and reduces the demand on high-level, expensive bandwidth in hierarchical topologies. We apply APHiD to a tapered fat-tree, demonstrating that APHiD maintains application scalability even for severely tapered network configurations. Using simulation, we show that for tapered networks APHiD improves performance by more than 50% over random placement and even 15% in some cases over costlier, state-of-the-art placement algorithms.

Michelogiannakis, George↗

Hierarchical computation in the SPMD programming model

Large-scale parallel machines are programmed mainly with the single program, multiple data (SPMD) model of parallelism. While this model has advantages of scalability and simplicity, it does not fit well with divide-and-conquer parallelism or hierarchical machines that mix shared and distributed memory. In this paper, we define the recursive single program, multiple data model (RSPMD) that extends SPMD with a hierarchical team mechanism to support hierarchical algorithms and machines. We implement this model in the Titanium language and describe how to eliminate a class of deadlocks by ensuring alignment of collective operations. We present application case studies evaluating the RSPMD model, showing that it enables divide-and-conquer algorithms such as sorting to be elegantly expressed and that team collective operations increase performance of conjugate gradient by up to a factor of two. The model also facilitates optimizations for hierarchical machines, improving scalability of particle in cell by 8x and performance of sorting and a stencil code by up to 40% and 14%, respectively.

Kamil, A↗

EAGLE: 'EAGLE'Is an' Algorithmic Graph Library for Exploration

The Resource Description Framework (RDF) and SPARQL Protocol and RDF Query Language (SPARQL) were introduced about a decade ago to enable flexible schema-free data interchange on the Semantic Web. Today data scientists use the framework as a scalable graph representation for integrating, querying, exploring and analyzing data sets hosted at different sources. With increasing adoption, the need for graph mining capabilities for the Semantic Web has emerged. Today there is no tools to conduct "graph mining" on RDF standard data sets. We address that need through implementation of popular iterative Graph Mining algorithms (Triangle count, Connected component analysis, degree distribution, diversity degree, PageRank, etc.). We implement these algorithms as SPARQL queries, wrapped within Python scripts and call our software tool as EAGLE. In RDF style, EAGLE stands for "EAGLE 'Is an' algorithmic graph library for exploration. EAGLE is like 'MATLAB' for 'Linked Data.'

Source record↗

Linux Kernel Co-Scheduling For Bulk Synchronous Parallel Applications

This paper describes a kernel scheduling algorithm that is based on co-scheduling principles and that is intended for parallel applications running on 1000 cores or more where inter-node scalability is key. Experimental results for a Linux implementation on a Cray XT5 machine are presented.1 The results indicate that Linux is a suitable operating system for this new scheduling scheme, and that this design provides a dramatic improvement in scaling performance for synchronizing collective operations at scale.

99 GENERAL AND MISCELLANEOUS↗

Aspherical particle models for molecular dynamics simulation

In traditional molecular dynamics (MD) simulations, atoms and coarse-grained particles are modeled as point masses interacting via isotropic potentials. For studies where particle shape plays a vital role, more complex models are required. Here, we describe a spectrum of approaches for modeling aspherical particles, all of which are now available (some recently) as options within the LAMMPS MD package. Broadly these include two classes of models. In the first, individual particles are aspherical, either via a pairwise anisotropic potential which implicitly assigns a simple geometric shape to each particle, or in a more general way where particles store internal state which can explicitly define a complex geometric shape. In the second class of models, individual particles are simple points or spheres, but rigid body constraints are used to create composite aspherical particles in a variety of complex shapes. This work discusses parallel algorithms and associated data structures for both kinds of models, which enable dynamics simulations of aspherical particle systems across a wide range of length and time scales. We also highlight parallel performance and scalability and give a few illustrative examples of aspherical models in different contexts.

74 ATOMIC AND MOLECULAR PHYSICS↗

A fast parallel Poisson solver on irregular domains applied to beam dynamics simulations

We discuss the scalable parallel solution of the Poisson equation within a Particle-In-Cell (PIC) code for the simulation of electron beams in particle accelerators of irregular shape. The problem is discretized by Finite Differences. Depending on the treatment of the Dirichlet boundary the resulting system of equations is symmetric or 'mildly' nonsymmetric positive definite. In all cases, the system is solved by the preconditioned conjugate gradient algorithm with smoothed aggregation (SA) based algebraic multigrid (AMG) preconditioning. We investigate variants of the implementation of SA-AMG that lead to considerable improvements in the execution times. We demonstrate good scalability of the solver on distributed memory parallel processor with up to 2048 processors. We also compare our iterative solver with an FFT-based solver that is more commonly used for applications in beam dynamics.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗

Multiresolution molecular mechanics: Implementation and efficiency

Atomistic/continuum coupling methods combine accurate atomistic methods and efficient continuum methods to simulate the behavior of highly ordered crystalline systems. Coupled methods utilize the advantages of both approaches to simulate systems at a lower computational cost, while retaining the accuracy associated with atomistic methods. Many concurrent atomistic/continuum coupling methods have been proposed in the past; however, their true computational efficiency has not been demonstrated. The present work presents an efficient implementation of a concurrent coupling method called the Multiresolution Molecular Mechanics (MMM) for serial, parallel, and adaptive analysis. First, we present the features of the software implemented along with the associated technologies. The scalability of the software implementation is demonstrated, and the competing effects of multiscale modeling and parallelization are discussed. Then, the algorithms contributing to the efficiency of the software are presented. These include algorithms for eliminating latent ghost atoms from calculations and measurement-based dynamic balancing of parallel workload. The efficiency improvements made by these algorithms are demonstrated by benchmark tests. The efficiency of the software is found to be on par with LAMMPS, a state-of-the-art Molecular Dynamics (MD) simulation code, when performing full atomistic simulations. Speed-up of the MMM method is shown to be directly proportional to the reduction of the number of the atoms visited in force computation. Finally, an adaptive MMM analysis on a nanoindentation problem, containing over a million atoms, is performed, yielding an improvement of 6.3–8.5 times in efficiency, over the full atomistic MD method. For the first time, the efficiency of a concurrent atomistic/continuum coupling method is comprehensively investigated and demonstrated.

75 CONDENSED MATTER PHYSICS, SUPERCONDUCTIVITY AND↗

Scaling and Quality of Modularity Optimization Methods for Graph Clustering

Real-world graphs exhibit structures known as “communities” or “clusters” consisting of a group of vertices with relatively high connectivity between them, as compared to the rest of the vertices in the network. Graph clustering or community detection is a fundamental graph operation used to analyze real-world graphs occurring in the areas of computational biology, cybersecurity, electrical grids, etc. Similar to other graph algorithms, owing to irregular memory accesses and inherently sequential nature, current algorithms for community detection are challenging to parallelize. However, in order to analyze large networks, it is important to develop scalable parallel implementations of graph clustering that are capable of exploiting the architectural features of modern supercomputers. In response to the 2019 Streaming Graph Challenge, we present quality and performance analysis of our distributed- memory community detection using Vite, which is our distributed memory implementation of the popular Louvain method, on the ALCF Theta supercomputer. Clustering methods such as Louvain that rely on modularity maximization are known to suffer from the resolution limit problem, preventing identification of clusters of certain sizes. Hence, we also include quality analysis of our shared-memory implementation of the Fast-tracking Resistance method, in comparison with Louvain on the challenge datasets. Furthermore, we introduce an edge-balanced graph distribution for our distributed memory implementation, that significantly reduces communication, offering up to 80% improvement in the overall execution time. In addition to performance/quality analysis, we also include details on the power/energy consumption, and memory traffic of the distributed-memory clustering implementation using real-world graphs with over a billion edges.

Ghosh, Sayan↗

Two-Stage Reinforcement Learning Policy Search for Grid-Interactive Building Control

This paper develops an intelligent grid-interactive building controller, which optimizes building operation during both normal hours and demand response (DR) events. To avoid costly on-demand computation and to adapt to non-linear building models, the controller utilizes reinforcement learning (RL) and makes real-time decisions based on a near-optimal control policy. Learning such a policy typically amounts to solving a hard non-convex optimization problem. We propose to address this problem with a novel global-local policy search method. In the first stage, an RL algorithm based on zero-order gradient estimation is leveraged to search for the optimal policy globally, due to its scalability and the potential to escape some poor performing local optima. The obtained policy is then fine-tuned locally to bring the first-stage solution closer to that of the original unsmoothed problem. Experiments on a simulated five-zone commercial building demonstrate the advantages of the proposed method over existing learning approaches. They also show that the learned control policy outperforms a pragmatic linear model predictive controller (MPC) and approaches the performance of an oracle MPC in testing scenarios. Using a state-of-the-art advanced computing system, we demonstrate that the controller can be learned and deployed within hours of training.

32 ENERGY CONSERVATION, CONSUMPTION, AND UTILIZATI↗

Validation and quantification of uncertainty in coupled climate models using network analysis

We developed a fast, robust and scalable methodology to examine, quantify, and visualize climate patterns and their relationships. It is based on a set of notions, algorithms and metrics used in the study of graphs, referred to as complex network analysis. This approach can be applied to explain known climate phenomena in terms of an underlying network structure and to uncover regional and global linkages in the climate system, while comparing general circulation models outputs with observations. The proposed method is based on a two-layer network representation, and is substantially new within the available network methodologies developed for climate studies. At the first layer, gridded climate data are used to identify ‘‘areas’’, i.e., geographical regions that are highly homogeneous in terms of the given climate variable. At the second layer, the identified areas are interconnected with links of varying strength, forming a global climate network. The robustness of the method (i.e. the ability to separate between topological distinct fields, while identifying correctly similarities) has been extensively tested. It has been proved that it provides a reliable, fast framework for comparing and ranking the ability of climate models of reproducing observed climate patterns and their connectivity. We further developed the methodology to account for lags in the connectivity between climate patterns and refined our area identification algorithm to account for autocorrelation in the data. The new methodology based on complex network analysis has been applied to state-of-the-art climate model simulations that participated to the last IPCC (International Panel for Climate Change) assessment to verify their performances, quantify uncertainties, and uncover changes in global linkages between past and future projections. Network properties of modeled sea surface temperature and rainfall over 1956–2005 have been constrained towards observations or reanalysis data sets, and their differences quantified using two metrics. Projected changes from 2051 to 2300 under the scenario with the highest representative and extended concentration pathways (RCP8.5 and ECP8.5) have then been determined. The network of models capable of reproducing well major climate modes in the recent past, changes little during this century. In contrast, among those models the uncertainties in the projections after 2100 remain substantial, and primarily associated with divergences in the representation of the modes of variability, particularly of the El Niño Southern Oscillation (ENSO), and their connectivity, and therefore with their intrinsic predictability, more so than with differences in the mean state evolution. Additionally, we evaluated the relation between the size and the ‘strength’ of the area identified by the network analysis as corresponding to ENSO noting that only a small subset of models can reproduce realistically the observations.

54 ENVIRONMENTAL SCIENCES↗

Noise-Induced Transition in Optimal Solutions of Variational Quantum Algorithms

Variational quantum algorithms are widely considered a promising class of near-team algorithms that may demonstrate a practical quantum advantage in problems of scientific interest. One of the major obstacles to a scalable realization is the difficulty in optimizing the noisy cost function.

72 PHYSICS OF ELEMENTARY PARTICLES AND FIELDS↗

CRIC and Rucio integration for CMS Data Management

The Large Hadron Collider (LHC) at CERN smashes protons together at close to the speed of light with seven times the energy of the most powerful accelerators built up to now. Some of the collision energy is turned into mass, creating new particles which are observed in the Compact Muon Solenoid (CMS) particle detector. CMS data is analyzed by scientists around the world to build up a picture of what happened at the heart of the collision. This will help us answer questions such as: ”what is the Universe really made of and what forces act within it?” and ”what gives everything substance?”. All the data produced by the experiment are collected and stored in a distributed system, since it must provide an high throughput and performing access to these data to scientists and physicists spread all over the world. In order to answer the previous questions and many others, CMS users must first obtain a copy of a portion of these data, and then to analyze it, therefore the first challenge is to decide in which storage site to save this duplicate so they can access it as quickly as possible. The second issue is each storage site shares a limited memory space with many CMS users, and each one needs to be assigned to a well defined personal amount of storage to pin data he needs. The main goal of my Summer Internship is to make scalable and automatic this user-site mapping which is based on a preexisting policy, proving a structured and organized algorithm to accomplish this task. Hence we are going to define, and solve, the main challenges encountered, and the technical solution used to generalize the algorithm to handle different policies for different countries.

43 PARTICLE ACCELERATORS↗

Demonstrating UPC++/Kokkos Interoperability in a Heat Conduction Simulation (Extended Abstract)

We describe the replacement of MPI with UPC++ in an existing Kokkos code that simulates heat conduction within a rectangular 3D object, as well as an analysis of the new code’s performance on CUDA accelerators. The key challenges were packing the halos in Kokkos data structures in a way that allowed for UPC++ remote memory access, and streamlining synchronization costs. Additional UPC++ abstractions used included global pointers, distributed objects, remote procedure calls, and futures. We also make use of the device allocator concept to facilitate data management in memory with unique properties, such as GPUs. Our results demonstrate that despite the algorithm’s good semantic match to message passing abstractions, straightforward modifications to use UPC++ communication deliver vastly improved performance and scalability in the common case. We find the one-sided UPC++ version written in a natural way exhibits good performance, whereas the message-passing version written in a straightforward way exhibits performance anomalies. We argue this represents a productivity benefit for one-sided communication models.

Waters, Daniel↗

Scalable Combinatorial Tools for Health Disparities Research

Despite staggering investments made in unraveling the human genome, current estimates suggest that as much as 90% of the variance in cancer and chronic diseases can be attributed to factors outside an individual’s genetic endowment, particularly to environmental exposures experienced across his or her life course. New analytical approaches are clearly required as investigators turn to complicated systems theory and ecological, place-based and life-history perspectives in order to understand more clearly the relationships between social determinants, environmental exposures and health disparities. While traditional data analysis techniques remain foundational to health disparities research, they are easily overwhelmed by the ever-increasing size and heterogeneity of available data needed to illuminate latent gene x environment interactions. This has prompted the adaptation and application of scalable combinatorial methods, many from genome science research, to the study of population health. Most of these powerful tools are algorithmically sophisticated, highly automated and mathematically abstract. Their utility motivates the main theme of this paper, which is to describe real applications of innovative transdisciplinary models and analyses in an effort to help move the research community closer toward identifying the causal mechanisms and associated environmental contexts underlying health disparities. The public health exposome is used as a contemporary focus for addressing the complex nature of this subject.

Environmental Sciences & Ecology↗

Metrics for Intercomparison of Remapping Algorithms (MIRA) protocol applied to Earth system models

Abstract. Strongly coupled nonlinear phenomena such as those described by Earth system models (ESMs) are composed of multiple component models with independent mesh topologies and scalable numerical solvers. A common operation in ESMs is to remap or interpolate component solution fields defined on their computational mesh to another mesh with a different combinatorial structure and decomposition, e.g., from the atmosphere to the ocean, during the temporal integration of the coupled system. Several remapping schemes are currently in use or available for ESMs. However, a unified approach to compare the properties of these different schemes has not been attempted previously. We present a rigorous methodology for the evaluation and intercomparison of remapping methods through an independently implemented suite of metrics that measure the ability of a method to adhere to constraints such as grid independence, monotonicity, global conservation, and local extrema or feature preservation. A comprehensive set of numerical evaluations is conducted based on a progression of scalar fields from idealized and smooth to more general climate data with strong discontinuities and strict bounds. We examine four remapping algorithms with distinct design approaches, namely ESMF Regrid (Hill et al., 2004), TempestRemap (Ullrich and Taylor, 2015), generalized moving least squares (GMLS) (Trask and Kuberry, 2020) with post-processing filters, and WLS-ENOR (Li et al., 2020). By repeated iterative application of the high-order remapping methods to the test fields, we verify the accuracy of each scheme in terms of their observed convergence order for smooth data and determine the bounded error propagation using challenging, realistic field data on both uniform and regionally refined mesh cases. In addition to retaining high-order accuracy under idealized conditions, the methods also demonstrate robust remapping performance when dealing with non-smooth data. There is a failure to maintain monotonicity in the traditional L2-minimization approaches used in ESMF and TempestRemap, in contrast to stable recovery through nonlinear filters used in both meshless GMLS and hybrid mesh-based WLS-ENOR schemes. Local feature preservation analysis indicates that high-order methods perform better than low-order dissipative schemes for all test cases. The behavior of these remappers remains consistent when applied on regionally refined meshes, indicating mesh-invariant implementations. The MIRA intercomparison protocol proposed in this paper and the detailed comparison of the four algorithms demonstrate that the new schemes, namely GMLS and WLS-ENOR, are competitive compared to standard conservative minimization methods requiring computation of mesh intersections. The work presented in this paper provides a foundation that can be extended to include complex field definitions, realistic mesh topologies, and spectral element discretizations, thereby allowing for a more complete analysis of production-ready remapping packages.

58 GEOSCIENCES↗