SEARCH · Engineering Papers
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.
This software implements a new algorithm and its shared-memory parallelization for computing maximum cardinality matchings in bipartite graphs. The algorithm searches for augmenting paths via specialized breadth-first searches (BFS) from multiple source vertices, hence creating more parallelism than single source algorithms.
Many iterative parallel-in-time algorithms have been shown to be highly efficient for diffusion-dominated partial differential equations (PDEs) but are inefficient or even divergent when applied to advection-dominated PDEs. We consider the application of the multigrid reduction-in-time (MGRIT) algorithm to linear advection PDEs. Here, the key to efficient time integration with this method is using a coarse-grid operator that provides a sufficiently accurate approximation to the so-called ideal coarse-grid operator. For certain classes of semi-Lagrangian discretizations, we present a novel semi-Lagrangian-based coarse-grid operator that leads to fast and scalable multilevel time integration of linear advection PDEs. The coarse-grid operator is composed of a semi-Lagrangian discretization followed by a correction term, with the correction designed so that the leading-order truncation error of the composite operator is approximately equal to that of the ideal coarse-grid operator. Parallel results show substantial speed-ups over sequential time integration for variable-wave-speed advection problems in one and two spatial dimensions, and using high-order discretizations up to order five. The proposed approach establishes the first practical method that provides small and scalable MGRIT iteration counts for advection problems.
‘‘Exact” quantum dynamics calculations of vibrational spectra are performed for two molecular systems of widely varying dimensionality (P{sub 2}O and CH{sub 2}NH), using a momentum-symmetrized Gaussian basis. This basis has been previously shown to defeat exponential scaling of computational cost with system dimensionality. The calculations were performed using the new “SWITCHBLADE” black-box code, which utilizes both dimensionally independent algorithms and massive parallelization to compute very large numbers of eigenstates for any fourth-order force field potential, in a single calculation. For both molecules considered here, many thousands of vibrationally excited states were computed, to at least an “intermediate” level of accuracy (tens of wavenumbers). Future modifications to increase the accuracy to “spectroscopic” levels, along with other potential future improvements of the new code, are also discussed.
Computational accelerator physics has changed and broadened over the last decade or so. Part of the change is due to the advent of multiple ways of parallel computing. Another part comes from algorithmic developments. The multiple ways of parallel computing include distributed memory parallelism and on-chip parallelism, with the latter coming from architectures (CPU and GPU) having multiple processing elements (cores or streaming multiprocessors) and wide vector (SIMD) instruction units. The basics of these new architectures and their application to computational accelerator physics are briefly reviewed. Algorithmic advances in the select areas of spin tracking, cavity calculations, plasma acceleration, and electron cooling are also reviewed. In some cases the algorithms provide increased fidelity improving the overall accuracy, while in other cases, such as controlled dispersion, the algorithms provide increased fidelity by better modeling the essential physical interaction. Finally, the use of computational frameworks, which provide the basic computational infrastructure, while allowing the capability developer to concentrate on the math and physics, is reviewed in the context of the Vorpal application, which has found use across accelerator physics and many other fields.
Scaling-up scientific data analysis and machine learning algorithms for data-driven discovery is a grand challenge that we face today. Despite the growing need for analysis from science domains that are generating `Big Data' from instruments and simulations, building high-performance analytical workflows of data-intensive algorithms have been daunting because: (i) the `Big Data' hardware and software architecture landscape is constantly evolving, (ii) newer architectures impose new programming models, and (iii) data-parallel kernels of analysis algorithms and their performance facets on different architectures are poorly understood. To address these problems, we have: (i) identified scalable data-parallel kernels of popular data analysis algorithms, (ii) implemented `Mini-Apps' of those kernels using different programming models (e.g. Map Reduce, MPI, etc.), (iii) benchmarked and validated the performance of the kernels in diverse architectures. In this paper, we discuss two of those Mini-Apps and show the execution of principal component analysis built as a workflow of the Mini-Apps. We show that Mini-Apps enable scientists to (i) write domain-specific data analysis code that scales on most HPC hardware and (ii) and offers the ability (most times with over a 10x speed-up) to analyze data sizes 100 times the size of what off-the-shelf desktop/workstations of today can handle.
Scaling-up scientific data analysis and machine learning algorithms for data-driven discovery is a grand challenge that we face today. Despite the growing need for analysis from science domains that are generating `Big Data' from instruments and simulations, building high-performance analytical workflows of data-intensive algorithms have been daunting because: (i) the `Big Data' hardware and software architecture landscape is constantly evolving, (ii) newer architectures impose new programming models, and (iii) data-parallel kernels of analysis algorithms and their performance facets on different architectures are poorly understood. To address these problems, we have: (i) identified scalable data-parallel kernels of popular data analysis algorithms, (ii) implemented `Mini-Apps' of those kernels using different programming models (e.g. Map Reduce, MPI, etc.), (iii) benchmarked and validated the performance of the kernels in diverse architectures. In this paper, we discuss two of those Mini-Apps and show the execution of principal component analysis built as a workflow of the Mini-Apps. We show that Mini-Apps enable scientists to (i) write domain-specific data analysis code that scales on most HPC hardware and (ii) and offers the ability (most times with over a 10x speed-up) to analyze data sizes 100 times the size of what off-the-shelf desktop/workstations of today can handle.
Stochastic mixed-integer programs (SMIPs) deal with optimization under uncertainty at many levels of the decision-making process. When solved as extensive formulation mixed- integer programs, problem instances can exceed available memory on a single workstation. In order to overcome this limitation, we present PIPS-SBB: a distributed-memory parallel stochastic MIP solver that takes advantage of parallelism at multiple levels of the optimization process. We also show promising results on the SIPLIB benchmark by combining methods known for accelerating Branch and Bound (B&B) methods with new ideas that leverage the structure of SMIPs. Finally, we expect the performance of PIPS-SBB to improve further as more functionality is added in the future.
Parallel-in-time methods for partial differential equations (PDEs) have been the subject of intense development over recent decades, particularly for diffusion-dominated problems. It has been widely reported in the literature, however, that many of these methods perform quite poorly for advection-dominated problems. In this report we analyze the particular iterative parallel-in-time algorithm of multigrid reduction-in-time (MGRIT) for discretizations of constant-wave-speed linear advection problems. We focus on common method-of-lines discretizations that employ upwind finite differences in space and Runge-Kutta methods in time. Using a convergence framework we developed in previous work, we prove for a subclass of these discretizations that, if using the standard approach of rediscretizing the fine-grid problem on the coarse grid, robust MGRIT convergence with respect to CFL number and coarsening factor is not possible. This poor convergence and non-robustness is caused, at least in part, by an inadequate coarse-grid correction for smooth Fourier modes in space-time known as characteristic components. We propose an alternative coarse-grid operator that provides a better correction of these modes. This coarse-grid operator is related to previous work and uses a semi-Lagrangian discretization combined with an implicitly treated truncation error correction. Theory and numerical experiments show the proposed coarse-grid operator yields fast MGRIT convergence for many of the method-of-lines discretizations considered, including for both implicit and explicit discretizations of high order. Parallel results demonstrate speed-up over sequential time-stepping.
Efficient parallel implementations of scientific applications on multi-core CPUs with accelerators such as GPUs and Xeon Phis is challenging. This requires - exploiting the data parallel architecture of the accelerator along with the vector pipelines of modern x86 CPU architectures, load balancing, and efficient memory transfer between different devices. It is relatively easy to meet these requirements for highly structured scientific applications. In contrast, a number of scientific and engineering applications are unstructured. Getting performance on accelerators for these applications is extremely challenging because many of these applications employ irregular algorithms which exhibit data-dependent control-ow and irregular memory accesses. Furthermore, these applications are often iterative with dependency between steps, and thus making it hard to parallelize across steps. As a result, parallelism in these applications is often limited to a single step. Numerical simulation of charged particles beam dynamics is one such application where the distribution of work and memory access pattern at each time step is irregular. Applications with these properties tend to present significant branch and memory divergence, load imbalance between different processor cores, and poor compute and memory utilization. Prior research on parallelizing such irregular applications have been focused around optimizing the irregular, data-dependent memory accesses and control-ow during a single step of the application independent of the other steps, with the assumption that these patterns are completely unpredictable. We observed that the structure of computation leading to control-ow divergence and irregular memory accesses in one step is similar to that in the next step. It is possible to predict this structure in the current step by observing the computation structure of previous steps. In this dissertation, we present novel machine learning based optimization techniques to address the parallel implementation challenges of such irregular applications on different HPC architectures. In particular, we use supervised learning to predict the computation structure and use it to address the control-ow and memory access irregularities in the parallel implementation of such applications on GPUs, Xeon Phis, and heterogeneous architectures composed of multi-core CPUs with GPUs or Xeon Phis. We use numerical simulation of charged particles beam dynamics simulation as a motivating example throughout the dissertation to present our new approach, though they should be equally applicable to a wide range of irregular applications. The machine learning approach presented here use predictive analytics and forecasting techniques to adaptively model and track the irregular memory access pattern at each time step of the simulation to anticipate the future memory access pattern. Access pattern forecasts can then be used to formulate optimization decisions during application execution which improves the performance of the application at a future time step based on the observations from earlier time steps. In heterogeneous architectures, forecasts can also be used to improve the memory performance and resource utilization of all the processing units to deliver a good aggregate performance. We used these optimization techniques and anticipation strategy to design a cache-aware, memory efficient parallel algorithm to address the irregularities in the parallel implementation of charged particles beam dynamics simulation on different HPC architectures. Experimental result using a diverse mix of HPC architectures shows that our approach in using anticipation strategy is effective in maximizing data reuse, ensuring workload balance, minimizing branch and memory divergence, and in improving resource utilization.
This work evaluated the use of OpenMP tasking with target GPU offloading as a potential solution for programming productivity and performance on heterogeneous systems. Also, it is proposed a new OpenMP specification to make the implementation of heterogeneous codes simpler by using OpenMP target task, which integrates both OpenMP tasking and target GPU offloading in a single OpenMP pragma. As a test case, the authors used one of the most popular and widely used Basic Linear Algebra Subprogram Level-3 routines: triangular solver (TRSM). To benefit from the heterogeneity of the current high-performance computing systems, the authors propose a different parallelization of the algorithm by using a nonuniform decomposition of the problem. This work used target GPU offloading inside OpenMP tasks to address the heterogeneity found in the hardware. This new approach can outperform the state-of-the-art algorithms, which use a uniform decomposition of the data, on both the CPU-only and hybrid CPU-GPU systems, reaching speedups of up to one order of magnitude. The performance that this approach achieves is faster than the IBM ESSL math library on CPU and competitive relative to a highly optimized heterogeneous CUDA version. One node of Oak Ridge National Laboratory’s supercomputer, Summit, was used for performance analysis.
Upon application of a sufficiently strong electric field, electrons break away from thermal equilibrium and approach relativistic speeds. These highly energetic ‘runaway’ electrons (~ MeV) play a significant role in tokamak disruption physics, and therefore their accurate understanding is essential to develop reliable mitigation strategies. As such, we have developed a fully implicit solver for the 0D-2P (i.e., including two momenta coordinates) relativistic nonlinear Fokker–Planck equation (rFP). As in earlier implicit rFP studies (NORSE, CQL3D), electron–ion interactions are modeled using the Lorentz operator, and synchrotron damping using the Abraham–Lorentz–Dirac reaction term. However, our implementation improves on these earlier studies by (1) ensuring exact conservation properties for electron collisions, (2) strictly preserving positivity, and (3) being scalable algorithmically and in parallel. Key to our proposed approach is an efficient multigrid preconditioner for the linearized rFP equation, a multigrid elliptic solver for the Braams–Karney potentials, and a novel adaptive technique to determine the associated boundary values. We verify the accuracy and efficiency of the proposed scheme with numerical results ranging from small electric-field electrical conductivity measurements to the accurate reproduction of runaway tail dynamics when strong electric fields are applied.
n this paper, we introduce a multigrid block-based preconditioner for solving linear systems arising from a Discontinuous Galerkin discretization of the all-speed Navier-Stokes equations with phase change. The equations are discretized in conservative form with a reconstructed Discontinuous Galerkin (rDG) method and integrated with fully-implicit time discretization schemes. To robustly converge the numerically stiff systems, we use the Newton-Krylov framework with a primitive-variable formulation (pressure, velocity, and temperature), which is better conditioned than the conservative-variable form at low-Mach number. In the limit of large acoustic CFL number and viscous Fourier number, there is a strong coupling between the velocity-pressure system and the linear systems become non-diagonally dominant. To effectively solve these ill-conditioned systems, an approximate block factorization preconditioner is developed, which uses the Schur complement to reduce a 3 x 3 block system into a sequence of two 2 x 2 block systems: velocity-pressure,v P , and velocity-temperature, v T . We compare the performance of the v P -v T Schur complement preconditioner to classic preconditioning strategies: monolithic algebraic multigrid (AMG), element-block SOR, and primitive variable block Gauss-Seidel. The performance of the preconditioned solver is investigated in the limit of large CFL and Fourier numbers for low-Mach lid-driven cavity flow, Rayleigh-Bénard melt convection, compressible internally heated convection, and 3D laser-induced melt pool flow. Here, numerical results demonstrate that the v P -v T Schur complement preconditioned solver scales well both algorithmically and in parallel, and is robust for highly ill-conditioned systems, for all tested rDG discretization schemes (up to 4th-order).
The nuclear industry aims to increase the cycle length of pressurized water reactors from 18 to 24 months to increase power plant capacity factors and economic viability. These cycle length extensions will inherently require fuel rods to exceed the current peak rod average burnup limit of 62 GWd/MTU. A chief concern of operating beyond the current burnup limit is the fuel fragmentation, relocation, and dispersal (FFRD) phenomenon in which pulverized fuel fragments can axially relocate and escape through a burst in the cladding formed during a loss-of-coolant accident. In this work, we demonstrate an approach for automating core design employing an optimization tool based on a penalty-free, parallel simulated annealing algorithm to produce pressurized water reactor core designs with two different optimization objectives. The two objectives were to produce core designs with (1) mitigated FFRD susceptibility while achieving 24-month cycle lengths (2) maximum cycle length with no regard for the likelihood of FFRD. Batch size was considered in tandem with both cases to maximize economic viability. The PARCS nodal model was the primary reactor physics tool used in the optimizations and used nuclear cross sections calculated with 2D Polaris lattice physics models. Reactor performance and safety characteristics of the optimized cores were verified using high-fidelity Virtual Environment for Reactor Applications models. The core designs produced by the optimization tool are compared with each other and to a high-burnup core design produced and analyzed in previous works to highlight the fuel management strategies that may enhance high-burnup reactor safety and economic viability. The optimized cores satisfied their respective objective functions, producing a maximum cycle length of 720 effective full-power days in one core design and one that may reduce FFRD susceptibility by up to 50% based on the first-order approximation to FFRD risk formulated in this work. The optimized cores met most constraints but exceeded the hot channel factor limit, especially in FFRD cases where fresh fuel carried more power. Furthermore, this highlights the need for future lattice-level optimizations and broader assembly options.
Real-time control of multiple plasma actuators is a requirement in advanced tokamaks; for example for burn control, plasma current profile control and MHD stabilization - EC wave absorption is ideally suited especially for the latter. For example, on ITER 24 EC sources can be switched between 54 inputs at the torus. In the torus, 5 launchers direct the power to various locations across the plasma profile via 11 steerable mirrors. For optimal usage of the available power, the aiming and polarization of the beams must be adapted to the plasma configuration and the needs of the scenario. Since the EC system performs many competing tasks, present day systems should demonstrate the ability of an EC plant to deal with several targets in parallel and/or to switch smoothly between goals to attain overall satisfaction. Recently TCV has taken a first step towards such a demonstration. Several EC launchers are used simultaneously to regulate the sawtooth period and to preempt m/n = 3/2 NTMs, by controlling the power levels. In parallel, a second algorithm stabilizes any NTM that saturates [1]. These and real-time MHD control experiments on ELMs [2] are presented.
Real-time lattice quantum electrodynamics (QED) provides a unique tool for simulating plasmas in the strong-field regime, where collective plasma scales are not well separated from relativistic-quantum scales. As a toy model, we study scalar QED, which describes self-consistent interactions between charged bosons and electromagnetic fields. To solve this model on a computer, we first discretize the scalar-QED action on a lattice, in a way that respects geometric structures of exterior calculus and U(1)-gauge symmetry. The lattice scalar QED can then be solved, in the classical-statistics regime, by advancing an ensemble of statistically equivalent initial conditions in time, using classical field equations obtained by extremizing the discrete action. To demonstrate the capability of our numerical scheme, we apply it to two example problems. The first example is the propagation of linear waves, where we recover analytic wave dispersion relations using numerical spectrum. The second example is an intense laser interacting with a one-dimensional plasma slab, where we demonstrate natural transition from wakefield acceleration to pair production when the wave amplitude exceeds the Schwinger threshold. Our real-time lattice scheme is fully explicit and respects local conservation laws, making it reliable for long-time dynamics. The algorithm is readily parallelized using domain decomposition, and the ensemble may also be computed using quantum parallelism in the future.
Finding the optimal implosion design on existing experimental facilities for inertial confinement fusion requires an exhaustive search of the vast design parameter space. This is infeasible both with experiments and with simulations. Consequently, a large fraction of the experimentally realizable design space remains unexplored, and new design schemes are challenging to optimize in a reasonable time frame. On the OMEGA laser facility, predictive machine learning models have been developed to accurately forecast the result of an experiment using only inexpensive simulations and the large dataset of prior experimental data. However, the full design space remains vast enough to be unassailable with simple optimization techniques. Here we develop an automated and optimally parallel Bayesian optimization algorithm that can entirely optimize the target and pulse shape of a direct-drive ICF implosion under a given design paradigm. We use this algorithm to find a markedly improved design for the performance implosions on OMEGA that is predicted to hydroequivalently scale to ignition at 2.15 MJ.
This paper presents a faster-than-real-time dynamic simulation software package that is designed for large-size power system dynamic simulation. It was developed on the GridPACKTM high-performance computing (HPC) framework. The key features of the developed software package include (1) faster-than-real-time dynamic simulation for a WECC system (17,000 buses) with different types of detailed generator, controller, and relay dynamic models, (2) a decoupled parallel dynamic simulation algorithm with optimized computation architecture to better leverage HPC resources and technologies, (3) options for HPC-based linear and iterative solvers, (4) hidden HPC details, such as data communication and distribution, to enable development centered on mathematical models and algorithms rather than on computational details for power system researchers, and (5) easy integration of new dynamic models and related algorithms into the software package.
An artificial intelligence system called MENNDL, which used 25,200 Nvidia Volta GPUs on Oak Ridge National Laboratory’s Summit machine, automatically designed an optimal deep learning network in order to extract structural information from raw atomic-resolution microscopy data. In a few hours, MENNDL creates and evaluates millions of networks using a scalable, parallel, asynchronous genetic algorithm augmented with a support vector machine to automatically find a superior deep learning network topology and hyper-parameter set than a human expert can find in months. For the application of electron microscopy, the system furthers the goal of improving our understanding of the electron-beam-matter interactions and real-time image-based feedback, which enables a huge step beyond human capacity toward nanofabricating materials automatically. Finally, MENNDL has been scaled to the 4,200 available nodes of Summit achieving a measured 152.5 PFlops, with an estimated sustained performance of 167 PFlops when the entire machine is available.