Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “Supercomputer”

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 1,225 records · Page 68

A matrix-algebraic formulation of distributed-memory maximal cardinality matching algorithms in bipartite graphs

We describe parallel algorithms for computing maximal cardinality matching in a bipartite graph on distributed-memory systems. Unlike traditional algorithms that match one vertex at a time, our algorithms process many unmatched vertices simultaneously using a matrix-algebraic formulation of maximal matching. This generic matrix-algebraic framework is used to develop three efficient maximal matching algorithms with minimal changes. The newly developed algorithms have two benefits over existing graph-based algorithms. First, unlike existing parallel algorithms, cardinality of matching obtained by the new algorithms stays constant with increasing processor counts, which is important for predictable and reproducible performance. Second, relying on bulk-synchronous matrix operations, these algorithms expose a higher degree of parallelism on distributed-memory platforms than existing graph-based algorithms. We report high-performance implementations of three maximal matching algorithms using hybrid OpenMP-MPI and evaluate the performance of these algorithm using more than 35 real and randomly generated graphs. On real instances, our algorithms achieve up to 200 × speedup on 2048 cores of a Cray XC30 supercomputer. Even higher speedups are obtained on larger synthetically generated graphs where our algorithms show good scaling on up to 16,384 cores.

97 MATHEMATICS AND COMPUTING↗

HPC formulations of optimization algorithms for tensor completion

Tensor completion is a powerful tool used to estimate or recover missing values in multi-way data. It has seen great success in domains such as product recommendation and healthcare. Tensor completion is most often accomplished via low-rank sparse tensor factorization, a computationally expensive non-convex optimization problem which has only recently been studied in the context of parallel computing. In this work, we study three optimization algorithms that have been successfully applied to tensor completion: alternating least squares (ALS), stochastic gradient descent (SGD), and coordinate descent (CCD++). We explore opportunities for parallelism on shared- and distributed-memory systems and address challenges such as memory- and operation-efficiency, load balance, cache locality, and communication. Among our advancements are a communication-efficient CCD++ algorithm, an ALS algorithm rich in level-3 BLAS routines, and an SGD algorithm which combines stratification with asynchronous communication. Furthermore, we show that introducing randomization during ALS and CCD++ can accelerate convergence. We evaluate our parallel formulations on a variety of real datasets on a modern supercomputer and demonstrate speedups through 16384 cores. Further, these improvements reduce time-to-solution from hours to seconds on real-world datasets. We show that after our optimizations, ALS is advantageous on parallel systems of small-to-moderate scale, while both ALS and CCD++ provide the lowest time-to-solution on large-scale distributed systems.

97 MATHEMATICS AND COMPUTING↗

Scalable Algorithms for MPI Intergroup Allgather and Allgatherv

MPI intergroup collective communication defines message transfer patterns between two disjoint groups of MPI processes. Such patterns occur in coupled applications, and in modern scientific application workflows, mostly with large data sizes. However, current implementations in MPI production libraries adopt the “root gathering algorithm”, which does not achieve optimal communication transfer time. In this paper, we propose algorithms for the intergroup Allgather and Allgatherv communication operations under single-port communication constraints. Here we implement the new algorithms using MPI point-to-point and standard intra-communicator collective communication functions. We evaluate their performance on the Cori supercomputer at NERSC. Using message sizes per compute node ranging from 64KBytes to 8MBytes, our experiments show significant performance improvements of up to 23.67 times on 256 compute nodes compared with the implementations of production MPI libraries.

97 MATHEMATICS AND COMPUTING↗

Exploring utilization options of heterogeneous architectures for multi-physics simulations

Top supercomputers in the TOP500 list have transitioned from homogeneous node architectures toward heterogeneous manycore nodes with accelerators and CPUs. These new architectures present significant challenges to developers of large-scale multiphysics applications, especially at DOE laboratories that have invested heavily in scalable MPI codes over decades. Much of these scientific application porting efforts for the new heterogeneous architectures are focused on running the computation on the accelerators, which usually comprise > 90% of the FLOPS of the system, typically using one core per GPU to offload work. Furthermore, we describe an approach to utilizing the remaining FLOPS on a heterogeneous machine by running a portion of the computation on the CPUs cooperatively with the GPU computation. Orthogonally, we evaluate using the additional cores to also offload work to GPUs, effectively oversubscribing the GPUs. We present a proof-of-concept implementation in ARES, a multiphysics ALE-AMR code at LLNL. ARES uses a portability layer, RAJA, which enables us to utilize the same source code for both the CPU and the GPU. We develop an approach to utilize both types of processors cooperatively in a heterogeneous system. Our implementation divides the work between the computing resources via domain decomposition, and utilizes all cores of the CPU and all of the GPUs on the node for computation. Load balancing is necessary to use the heterogeneous resources effectively. We present preliminary results on early delivery pre-Sierra machines at LLNL, showing up to 40% performance benefit of oversubscribing the GPUs (with caveats), and up to an 18% performance benefit of using the CPUs on the heterogeneous nodes for computing in addition to using the GPUs.

97 MATHEMATICS AND COMPUTING↗

Performance of asynchronous optimized Schwarz with one-sided communication

In asynchronous iterative methods on distributed-memory computers, processes update their local solutions using data from other processes without an implicit or explicit global synchronization that corresponds to advancing the global iteration counter. Here in this work, we test the asynchronous optimized Schwarz domain-decomposition iterative method using various one-sided (remote direct memory access) communication schemes with passive target completion. The results show that when one-sided communication is well-supported, the asynchronous version of optimized Schwarz can outperform the synchronous version even for perfectly balanced partitionings of the problem on a supercomputer with uniform nodes.

97 MATHEMATICS AND COMPUTING↗

Targeting GPUs with OpenMP directives on Summit: A simple and effective Fortran experience

We use OpenMP to target hardware accelerators (GPUs) on Summit, a newly deployed supercomputer at the Oak Ridge Leadership Computing Facility (OLCF), demonstrating simplified access to GPU devices for users of our astrophysics code GenASiS and useful speedup on a sample fluid dynamics problem. We modify our workhorse class for data storage to include members and methods that significantly streamline the persistent allocation of and association to GPU memory. Users offload computational kernels with OpenMP target directives that are rather similar to constructs already familiar from multi-core parallelization. In this initial example we ask, “With a given number of Summit nodes, how fast can we compute with and without GPUs?”, and find total wall time speedups of ~ 12X. We also find reasonable weak scaling up to 8000 GPUs (1334 Summit nodes). We make available the source code from this work at https://github.com/GenASiS/GenASiS_Basics.

97 MATHEMATICS AND COMPUTING↗

A parallel strategy for density functional theory computations on accelerated nodes

Using the Löwdin orthonormalization of tall-skinny matrices as a proxy-app for wavefunction-based Density Functional Theory solvers, we investigate a distributed memory parallel strategy focusing on Graphics Processing Unit (GPU)-accelerated nodes as available on some of the top ranked supercomputers at the present time. Here we present numerical results in the strong limit regime, as it is particularly relevant for First-Principles Molecular Dynamics. We also examine how matrix product-based iterative solvers provide a competitive alternative to dense eigensolvers on GPUs, allowing to push the strong scaling limit of these computations to a larger number of distributed tasks. Our strategy, which relies on replicated Gram matrices and efficient collective communications using the NCCL library, leads to a time-to-solution under 0.5 s for the Löwdin orthonormalization of a tall-skinny matrix of 3000 columns on Summit at Oak Ridge Leadership Facility (OLCF). Given the similarity in computational operations between one iteration of a DFT solver and this proxy-app, this shows the possibility of solving accurately the DFT equations well under a minute for 3000 electronic wave functions, and thus perform First-Principles molecular dynamics of physical systems much larger than traditionally solved on CPU systems.

97 MATHEMATICS AND COMPUTING↗

Immortal rays: Rethinking random ray neutron transport on GPU architectures

The Random Ray Method (TRRM) is a recently developed adaptation of the Method of Characteristics for neutral particle transport simulations. TRRM has demonstrated excellent performance on 3D nuclear reactor benchmark problems using CPU-based compute systems. When porting to GPU-based systems, however, new performance challenges arise that are unique to processors targeting massive fine-grained parallelism. For smaller problems, or for large problems that are domain decomposed across many computational nodes, the problem size per node has insufficient parallelism to saturate GPU node resources, thus greatly limiting speedup. In this study, we report on a newly developed “immortal ray” variant of TRRM. Here, the immortal ray technique exposes significantly more fine-grained parallelism by fundamentally reformulating the numerical details of ray discretization, resulting in performance tradeoffs with significant overall benefit on GPUs. For very small 2D simulation problems we found the new immortal ray variant allowed for up to a 4.4x speedup when run on a single GPU. For larger 3D simulation problems we found the new variant improved strong scaling by 3x when run on the Summit supercomputer.

97 MATHEMATICS AND COMPUTING↗

Porting WarpX to GPU-accelerated platforms

WarpX is a general purpose electromagnetic particle-in-cell code that was originally designed to run on many-core CPU architectures. We describe the strategy, based on the AMReX library, followed to allow WarpX to use the GPU-accelerated nodes on OLCF’s Summit supercomputer, a strategy we believe will extend to the upcoming machines Frontier and Aurora. We summarize the challenges encountered, lessons learned, and give current performance results on a series of relevant benchmark problems.

97 MATHEMATICS AND COMPUTING↗

Towards scaling community detection on distributed-memory heterogeneous systems

Distributed multi-GPU systems pose significant challenges and opportunities for efficient execution of parallel applications. Graph algorithms are generally characterized by irregular memory accesses, low computation to communication ratios, and load balancing problems that are especially hard to address on multi-GPU systems. Graph community detection is an important problem in the emerging domain of graph analytics with numerous applications. In this paper, we present our ongoing work on distributed-memory multi-GPU implementation for graph community detection. Our work parallelizes the widely used (albeit serial) Louvain method on distributed multi-GPU platforms. Supported by an extensive set of experiments on a multi-GPU enabled supercomputer (OLCF Summit) and a single compute node (Nvidia DGX-2®), we demonstrate competitive performance to existing distributed-memory CPU-based implementation, and up to 6.5 better results than Nvidia RAPIDS® CUGRAPH. To the best of our knowledge, this work represents the first effort for community detection on distributed multi-GPU systems. Our approach and related findings can be extended to numerous other iterative graph algorithms on multi-GPU systems.

97 MATHEMATICS AND COMPUTING↗

Low-synch Gram–Schmidt with delayed reorthogonalization for Krylov solvers

The parallel strong-scaling of iterative methods is often determined by the number of global reductions at each iteration. Low-synch Gram-Schmidt algorithms are applied here to the Arnoldi algorithm to reduce the number of global reductions and therefore to improve the parallel strong-scaling of iterative solvers for nonsymmetric matrices such as the GMRES and the Krylov-Schur iterative methods. In the Arnoldi context, the factorization is "left-looking" and processes one column at a time. Among the methods for generating an orthogonal basis for the Arnoldi algorithm, the classical Gram-Schmidt algorithm, with reorthogonalization (CGS2) requires three global reductions per iteration. A new variant of CGS2 that requires only one reduction per iteration is presented and applied to the Arnoldi algorithm. Delayed CGS2 (DCGS2) employs the minimum number of global reductions per iteration (one) for a one-column at-a-time algorithm. The main idea behind the new algorithm is to group global reductions by rearranging the order of operations. DCGS2 must be carefully integrated into an Arnoldi expansion or a GMRES solver. Numerical stability experiments assess robustness for Krylov-Schur eigenvalue computations. Performance experiments on the ORNL Summit supercomputer then establish the superiority of DCGS2 over CGS2.

97 MATHEMATICS AND COMPUTING↗

AI and extreme scale computing to learn and infer the physics of higher order gravitational wave modes of quasi-circular, spinning, non-precessing black hole mergers

We use artificial intelligence (AI) to learn and infer the physics of higher order gravitational wave modes of quasi-circular, spinning, non precessing binary black hole mergers. We trained AI models using 14 million waveforms, produced with the surrogate model NRHybSur3dq8, that include modes up to $\ell$ ≤ 4 and (5,5), except for (4,0) and (4,1), that describe binaries with mass-ratios $\textit{q}$ ≤ 8, individual spins $s^z_{\{1,2\}} \in$[–0.8,0.8], and inclination angle $θ \in$ [0,π]. Our probabilistic AI surrogates can accurately constrain the mass-ratio, individual spins, effective spin, and inclination angle of numerical relativity waveforms that describe such signal manifold. We compared the predictions of our AI models with Gaussian process regression, random forest, k-nearest neighbors, and linear regression, and with traditional Bayesian inference methods through the PyCBC Inference toolkit, finding that AI outperforms all these approaches in terms of accuracy, and are between three to four orders of magnitude faster than traditional Bayesian inference methods. Our AI surrogates were trained within 3.4 hours using distributed training on 1,536 NVIDIA V100 GPUs in the Summit supercomputer.

79 ASTRONOMY AND ASTROPHYSICS↗

Turbulence theories and statistical closure approaches

When discussing research in physics and in science more generally, it is common to ascribe equal importance to the three components of the scientific trinity: theoretical, experimental, and computational studies. This review will explore the future of modern turbulence theory by tracing its history, which began in earnest with Kolmogorov’s 1941 analysis of turbulence cascade and inertial range [A.N. Kolmogorov, Dokl. Akad. Nauk SSSR, 30, 299, (1941); 32, 19, (1941)]. The 80th Anniversary of Kolmogorov’s landmark study is a welcome opportunity to survey the achievements and evaluate the future of the theoretical approach of turbulence research. Over the years, turbulence theories have been critically important in laying the foundation of our understanding of the nature of turbulent flows. In particular, the Direct Interaction Approximation (DIA) [R.H. Kraichnan, J. Fluid Mech., 5, 497 (1959)] and its subsequent development, known as the statistical closure approach, can be identified as perhaps the most profound single advancement. The remarkable success of the statistical closure has furnished a platform to study such essential concepts as the energy transfer process and interacting scales, and the roles of the straining and sweeping motions. More recently, the quasi-Lagrangian formulation of V. L’vov & I. Procaccia and Kraichnan’s solvable passive scalar model provided powerful ways to explore another fundamental aspect of turbulent flows, the phenomena of intermittency, and the associated anomalous scaling exponents. In the meantime, the theory of fluid equilibria has been developed to describe the large-scale structures that can emerge from turbulent cascades of two-dimensional and geophysical flows at a later time. And yet, despite all these successes, analytical treatments suffer from mathematical complexities. As a result, the utility of theoretical approaches has been limited to relatively idealized flows. On the other hand, in recent decades, computational abilities and experimental facilities have reached an unprecedented scale. Looking beyond the horizon, the imminent deployment of exascale supercomputers will generate complete datasets of the entire flow field of key benchmark flows, allowing researchers to extract additional measurements concerning fully developed, complex turbulent flow fields far beyond those available from the statistical closure theories. Some other developments that could potentially influence the future course of turbulence theories include the advancement of machine learning, artificial intelligence, and data science; likely disruptions arising from the advent of quantum computation; and the increasingly prominent role of turbulence research in providing more accurate climate scientific data. Finally, turbulence theorists can leverage these developments by asking the right questions and developing advanced, sophisticated frameworks that will be able to predict and correlate vast amounts of data from the other two components of the trinity.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗

ParCAT: Parallel Climate Analysis Toolkit

Climate science is employing increasingly complex models and simulations to analyze the past and predict the future of Earth's climate. This growth in complexity is creating a widening gap between the data being produced and the ability to analyze the datasets. Parallel computing tools are necessary to analyze, compare, and interpret the simulation data. The Parallel Climate Analysis Toolkit (ParCAT) provides basic tools to efficiently use parallel computing techniques to make analysis of these datasets manageable. The toolkit provides the ability to compute spatio-temporal means, differences between runs or differences between averages of runs, and histograms of the values in a data set. ParCAT is implemented as a command-line utility written in C. This allows for easy integration in other tools and allows for use in scripts. This also makes it possible to run ParCAT on many platforms – from laptops to supercomputers. ParCAT outputs NetCDF files so it is compatible with existing utilities such as Panoply and UV-CDAT. This paper describes ParCAT and presents results from some example runs on the Titan system at ORNL.

97 MATHEMATICS AND COMPUTING↗

MapReduce SVM Game

Despite technological advances making computing devices faster, smaller, and more prevalent in today's age, data generation and collection has outpaced data processing capabilities. Simply having more compute platforms does not provide a means of addressing challenging problems in the big data era. Rather, alternative processing approaches are needed and the application of machine learning to big data is hugely important. The MapReduce programming paradigm is an alternative to conventional supercomputing approaches, and requires less stringent data passing constrained problem decompositions. Rather, MapReduce relies upon defining a means of partitioning the desired problem so that subsets may be computed independently and recom- bined to yield the net desired result. However, not all machine learning algorithms are amenable to such an approach. Game-theoretic algorithms are often innately distributed, consisting of local interactions between players without requiring a central authority and are iterative by nature rather than requiring extensive retraining. Effectively, a game-theoretic approach to machine learning is well suited for the MapReduce paradigm and provides a novel, alternative new perspective to addressing the big data problem. In this paper we present a variant of our Support Vector Machine (SVM) Game classifier which may be used in a distributed manner, and show an illustrative example of applying this algorithm.

97 MATHEMATICS AND COMPUTING↗

Asynchronous Two-Level Checkpointing Scheme for Large-Scale Adjoints in the Spectral-Element Solver Nek5000

Adjoints are an important computational tool for large-scale sensitivity evaluation, uncertainty quantification, and derivative-based optimization. An essential component of their performance is the storage/recomputation balance in which efficient checkpointing methods play a key role. We introduce a novel asynchronous two-level adjoint checkpointing scheme for multistep numerical time discretizations targeted at large-scale numerical simulations. The checkpointing scheme combines bandwidth-limited disk checkpointing and binomial memory checkpointing. Based on assumptions about the target petascale systems, which we later demonstrate to be realistic on the IBM Blue Gene/Q system Mira, we create a model of the expected performance of our checkpointing approach and validate it using the highly scalable Navier-Stokes spectralelement solver Nek5000 on small to moderate subsystems of the Mira supercomputer. In turn, this allows us to predict optimal algorithmic choices when using all of Mira. We also demonstrate that two-level checkpointing is significantly superior to single-level checkpointing when adjoining a large number of time integration steps. To our knowledge, this is the first time two-level checkpointing had been designed, implemented, tuned, and demonstrated on fluid dynamics codes at large scale of 50k+ cores.

Adjoints↗

Generating unstructured nuclear reactor core meshes in parallel

Recent advances in supercomputers and parallel solver techniques have enabled users to run large simulations problems using millions of processors. Techniques for multiphysics nuclear reactor core simulations are under active development in several countries. Most of these techniques require large unstructured meshes that can be hard to generate in a standalone desktop computers because of high memory requirements, limited processing power, and other complexities. We have previously reported on a hierarchical lattice-based approach for generating reactor core meshes. Here, we describe efforts to exploit coarse-grained parallelism during reactor assembly and reactor core mesh generation processes. We highlight several reactor core examples including a very high temperature reactor, a full-core model of the Korean MONJU reactor, a ¼ pressurized water reactor core, the fast reactor Experimental Breeder Reactor-II core with a XX09 assembly, and an advanced breeder test reactor core. The times required to generate large mesh models, along with speedups obtained from running these problems in parallel, are reported. A graphical user interface to the tools described here has also been developed.

22 GENERAL STUDIES OF NUCLEAR REACTORS↗

G-protein coupled receptors: advances in simulation and drug discovery

G-protein coupled receptors (GPCRs), the largest family of human membrane proteins, mediate cellular signaling and represent primary targets of about one third of currently marketed drugs. GPCRs undergo highly dynamic structural transitions during signal transduction, from binding of extracellular ligands to coupling with intracellular effector proteins. Molecular dynamics (MD) simulations have been utilized to investigate GPCR signaling mechanisms (such as pathways of ligand binding and receptor activation/deactivation) and to design novel small-molecule drug candidates. Future research directions point towards modeling cooperative binding of multiple orthosteric and allosteric ligands to GPCRs, GPCR oligomerization and interactions of GPCRs with different intracellular signaling proteins. Lastly, through methodological and supercomputing advances, MD simulations will continue to provide important insights into GPCR signaling mechanisms and further facilitate structure-based drug design.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗