Search NASA⌕ Search

SEARCH · Search NASA

Results for “Parallel processors”

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 631 records · Page 35

Characterizing the impact of using spare-cores on application performance

Increased parallelism on a single processor is driving improvements in peak-performance at both the node and system levels. However achievable performance, in particular from production scientific applications, is not always directly proportional to the core count. Performance is often limited by constraints in the memory hierarchy and also by a node interconnectivity. Even on state-of-the-art processors, containing between four and eight cores, many applications cannot take full advantage of the compute-performance of all cores. This trend is expected to increase on future processors as the core count per processor increases. In this work we characterize the use of spare-cores, cores that do not provide any improvements in application performance, on current multi-core processors. By using a pulse-width modulation method, we examine the possible performance profile of using a spare-core and quantify under what situations its use will not impact application performance. We show that, for current AMD and Intel multi-core processors, spare-cores can be used for substantial computational tasks but can impact application performance when using shared caches or when significantly accessing main memory.

97 MATHEMATICS AND COMPUTING↗

Alternating Anderson–Richardson method: An efficient alternative to preconditioned Krylov methods for large, sparse linear systems

Here, we present the Alternating Anderson–Richardson (AAR) method: an efficient and scalable alternative to preconditioned Krylov solvers for the solution of large, sparse linear systems on high performance computing platforms. Specifically, we generalize the recently proposed Alternating Anderson–Jacobi (AAJ) method (Pratapa et al., 2016) to include preconditioning, discuss efficient parallel implementation, and provide serial MATLAB and parallel C/C++ implementations. In serial applications to nonsymmetric systems, we find that AAR is comparably robust to GMRES, using the same preconditioning, while often outperforming it in time to solution; and find AAR to be more robust than Bi-CGSTAB for the problems considered. In parallel applications to the Helmholtz and Poisson equations, we find that AAR shows superior strong and weak scaling to GMRES, Bi-CGSTAB, and Conjugate Gradient (CG) methods, using the same preconditioning, with consistently shorter times to solution at larger processor counts. Finally, in massively parallel applications to the Poisson equation, on up to 110,592 processors, we find that AAR shows superior strong and weak scaling to CG, with shorter minimum time to solution. We thus find that AAR offers a robust and efficient alternative to current state-of-the-art solvers, with increasing advantages as the number of processors grows.

97 MATHEMATICS AND COMPUTING↗

Production Level CFD Code Acceleration for Hybrid Many-Core Architectures

In this work, a novel graphics processing unit (GPU) distributed sharing model for hybrid many-core architectures is introduced and employed in the acceleration of a production-level computational fluid dynamics (CFD) code. The latest generation graphics hardware allows multiple processor cores to simultaneously share a single GPU through concurrent kernel execution. This feature has allowed the NASA FUN3D code to be accelerated in parallel with up to four processor cores sharing a single GPU. For codes to scale and fully use resources on these and the next generation machines, codes will need to employ some type of GPU sharing model, as presented in this work. Findings include the effects of GPU sharing on overall performance. A discussion of the inherent challenges that parallel unstructured CFD codes face in accelerator-based computing environments is included, with considerations for future generation architectures. This work was completed by the author in August 2010, and reflects the analysis and results of the time.

Duffy, Austen C.↗

Digital system for structural dynamics simulation

State-of-the-art digital hardware and software for the simulation of complex structural dynamic interactions, such as those which occur in rotating structures (engine systems). System were incorporated in a designed to use an array of processors in which the computation for each physical subelement or functional subsystem would be assigned to a single specific processor in the simulator. These node processors are microprogrammed bit-slice microcomputers which function autonomously and can communicate with each other and a central control minicomputer over parallel digital lines. Inter-processor nearest neighbor communications busses pass the constants which represent physical constraints and boundary conditions. The node processors are connected to the six nearest neighbor node processors to simulate the actual physical interface of real substructures. Computer generated finite element mesh and force models can be developed with the aid of the central control minicomputer. The control computer also oversees the animation of a graphics display system, disk-based mass storage along with the individual processing elements.

Krauter, A. I.↗

Summer Proceedings 2016: The Center for Computing Research at Sandia National Laboratories

Solving sparse linear systems from the discretization of elliptic partial differential equations (PDEs) is an important building block in many engineering applications. Sparse direct solvers can solve general linear systems, but are usually slower and use much more memory than effective iterative solvers. To overcome these two disadvantages, a hierarchical solver (LoRaSp) based on H2-matrices was introduced in [22]. Here, we have developed a parallel version of the algorithm in LoRaSp to solve large sparse matrices on distributed memory machines. On a single processor, the factorization time of our parallel solver scales almost linearly with the problem size for three-dimensional problems, as opposed to the quadratic scalability of many existing sparse direct solvers. Moreover, our solver leads to almost constant numbers of iterations, when used as a preconditioner for Poisson problems. On more than one processor, our algorithm has significant speedups compared to sequential runs. With this parallel algorithm, we are able to solve large problems much faster than many existing packages as demonstrated by the numerical experiments.

97 MATHEMATICS AND COMPUTING↗

In-Situ Statistical Analysis of Autotune Simulation Data using Graphical Processing Units

Developing accurate building energy simulation models to assist energy efficiency at speed and scale is one of the research goals of the Whole-Building and Community Integration group, which is a part of Building Technologies Research and Integration Center (BTRIC) at Oak Ridge National Laboratory (ORNL). The aim of the Autotune project is to speed up the automated calibration of building energy models to match measured utility or sensor data. The workflow of this project takes input parameters and runs EnergyPlus simulations on Oak Ridge Leadership Computing Facility’s (OLCF) computing resources such as Titan, the world’s second fastest supercomputer. Multiple simulations run in parallel on nodes having 16 processors each and a Graphics Processing Unit (GPU). Each node produces a 5.7 GB output file comprising 256 files from 64 simulations. Four types of output data covering monthly, daily, hourly, and 15- minute time steps for each annual simulation is produced. A total of 270TB+ of data has been produced. In this project, the simulation data is statistically analyzed in-situ using GPUs while annual simulations are being computed on the traditional processors. Titan, with its recent addition of 18,688 Compute Unified Device Architecture (CUDA) capable NVIDIA GPUs, has greatly extended its capability for massively parallel data processing. CUDA is used along with C/MPI to calculate statistical metrics such as sum, mean, variance, and standard deviation leveraging GPU acceleration. The workflow developed in this project produces statistical summaries of the data which reduces by multiple orders of magnitude the time and amount of data that needs to be stored. These statistical capabilities are anticipated to be useful for sensitivity analysis of EnergyPlus simulations

accelerators↗

Hierarchical fractional-step approximations and parallel kinetic Monte Carlo algorithms

We present a mathematical framework for constructing and analyzing parallel algorithms for lattice kinetic Monte Carlo (KMC) simulations. The resulting algorithms have the capacity to simulate a wide range of spatio-temporal scales in spatially distributed, non-equilibrium physiochemical processes with complex chemistry and transport micro-mechanisms. Rather than focusing on constructing exactly the stochastic trajectories, our approach relies on approximating the evolution of observables, such as density, coverage, correlations and so on. More specifically, we develop a spatial domain decomposition of the Markov operator (generator) that describes the evolution of all observables according to the kinetic Monte Carlo algorithm. This domain decomposition corresponds to a decomposition of the Markov generator into a hierarchy of operators and can be tailored to specific hierarchical parallel architectures such as multi-core processors or clusters of Graphical Processing Units (GPUs). Based on this operator decomposition, we formulate parallel Fractional step kinetic Monte Carlo algorithms by employing the Trotter Theorem and its randomized variants; these schemes, (a) are partially asynchronous on each fractional step time-window, and (b) are characterized by their communication schedule between processors. The proposed mathematical framework allows us to rigorously justify the numerical and statistical consistency of the proposed algorithms, showing the convergence of our approximating schemes to the original serial KMC. The approach also provides a systematic evaluation of different processor communicating schedules. We carry out a detailed benchmarking of the parallel KMC schemes using available exact solutions, for example, in Ising-type systems and we demonstrate the capabilities of the method to simulate complex spatially distributed reactions at very large scales on GPUs. Finally, we discuss work load balancing between processors and propose a re-balancing scheme based on probabilistic mass transport methods.

97 MATHEMATICS AND COMPUTING↗

Computational Particle Dynamic Simulations on Multicore Processors (CPDMu) Final Report Phase I

Statement of Problem - Department of Energy has many legacy codes for simulation of computational particle dynamics and computational fluid dynamics applications that are designed to run on sequential processors and are not easily parallelized. Emerging high-performance computing architectures employ massively parallel multicore architectures (e.g., graphics processing units) to increase throughput. Parallelization of legacy simulation codes is a high priority, to achieve compatibility, efficiency, accuracy, and extensibility. General Statement of Solution - A legacy simulation application designed for implementation on mainly-sequential processors has been represented as a graph G. Mathematical transformations, applied to G, produce a graph representation {und G} for a high-performance architecture. Key computational and data movement kernels of the application were analyzed/optimized for parallel execution using the mapping G {yields} {und G}, which can be performed semi-automatically. This approach is widely applicable to many types of high-performance computing systems, such as graphics processing units or clusters comprised of nodes that contain one or more such units. Phase I Accomplishments - Phase I research decomposed/profiled computational particle dynamics simulation code for rocket fuel combustion into low and high computational cost regions (respectively, mainly sequential and mainly parallel kernels), with analysis of space and time complexity. Using the research team's expertise in algorithm-to-architecture mappings, the high-cost kernels were transformed, parallelized, and implemented on Nvidia Fermi GPUs. Measured speedups (GPU with respect to single-core CPU) were approximately 20-32X for realistic model parameters, without final optimization. Error analysis showed no loss of computational accuracy. Commercial Applications and Other Benefits - The proposed research will constitute a breakthrough in solution of problems related to efficient parallel computation of particle and fluid dynamics simulations. These problems occur throughout DOE, military and commercial sectors: the potential payoff is high. We plan to license or sell the solution to contractors for military and domestic applications such as disaster simulation (aerodynamic and hydrodynamic), Government agencies (hydrological and environmental simulations), and medical applications (e.g., in tomographic image reconstruction). Keywords - High-performance Computing, Graphic Processing Unit, Fluid/Particle Simulation. Summary for Members of Congress - Department of Energy has many simulation codes that must compute faster, to be effective. The Phase I research parallelized particle/fluid simulations for rocket combustion, for high-performance computing systems.

97 MATHEMATICS AND COMPUTING↗

Load Balancing Scientific Applications

The largest supercomputers have millions of independent processors, and concurrency levels are rapidly increasing. For ideal efficiency, developers of the simulations that run on these machines must ensure that computational work is evenly balanced among processors. Assigning work evenly is challenging because many large modern parallel codes simulate behavior of physical systems that evolve over time, and their workloads change over time. Furthermore, the cost of imbalanced load increases with scale because most large-scale scientific simulations today use a Single Program Multiple Data (SPMD) parallel programming model, and an increasing number of processors will wait for the slowest one at the synchronization points. To address load imbalance, many large-scale parallel applications use dynamic load balance algorithms to redistribute work evenly. The research objective of this dissertation is to develop methods to decide when and how to load balance the application, and to balance it effectively and affordably. We measure and evaluate the computational load of the application, and develop strategies to decide when and how to correct the imbalance. Depending on the simulation, a fast, local load balance algorithm may be suitable, or a more sophisticated and expensive algorithm may be required. We developed a model for comparison of load balance algorithms for a specific state of the simulation that enables the selection of a balancing algorithm that will minimize overall runtime.

97 MATHEMATICS AND COMPUTING↗

On the parallel efficiency of the Frederickson-McBryan multigrid

To take full advantage of the parallelism in a standard multigrid algorithm requires as many processors as points. However, since coarse grids contain fewer points, most processors are idle during the coarse grid iterations. Frederickson and McBryan claim that retaining all points on all grid levels (using all processors) can lead to a superconvergent algorithm. The purpose of this work is to show that the parellel superconvergent multigrid (PSMG) algorithm of Frederickson and McBryan, though it achieves perfect processor utilization, is no more efficient than a parallel implementation of standard multigrid methods. PSMG is simply a new and perhaps simpler way of achieving the same results.

Decker, Naomi H.↗

A quantum processor based on coherent transport of entangled atom arrays

The ability to engineer parallel, programmable operations between desired qubits within a quantum processor is key for building scalable quantum information systems. In most state-of-the-art approaches, qubits interact locally, constrained by the connectivity associated with their fixed spatial layout. Here we demonstrate a quantum processor with dynamic, non-local connectivity, in which entangled qubits are coherently transported in a highly parallel manner across two spatial dimensions, between layers of single- and two-qubit operations. Our approach makes use of neutral atom arrays trapped and transported by optical tweezers; hyperfine states are used for robust quantum information storage, and excitation into Rydberg states is used for entanglement generation. We use this architecture to realize programmable generation of entangled graph states, such as cluster states and a seven-qubit Steane code state. Furthermore, we shuttle entangled ancilla arrays to realize a surface code state with thirteen data and six ancillary qubits and a toric code state on a torus with sixteen data and eight ancillary qubits. Finally, we use this architecture to realize a hybrid analogue–digital evolution and use it for measuring entanglement entropy in quantum simulations, experimentally observing non-monotonic entanglement dynamics associated with quantum many-body scars. Realizing a long-standing goal, these results provide a route towards scalable quantum processing and enable applications ranging from simulation to metrology.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗

Incentive Compatible Online Scheduling of Malleable Parallel Jobs with Individual Deadlines

We consider the online scheduling of malleable jobs on parallel systems, such as clusters, symmetric multiprocessing computers, and multi-core processor computers. Malleable jobs is a model of parallel processing in which jobs adapt to the number of processors assigned to them. This model permits the scheduler and resource manager to make more efficient use of the available resources. Each malleable job is characterized by arrival time, deadline, and value. If the job completes by its deadline, the user earns the payoff indicated by the value; otherwise, she earns a payoff of zero. The scheduling objective is to maximize the sum of the values of the jobs that complete by their associated deadlines. Complicating the matter is that users in the real world are rational and they will attempt to manipulate the scheduler by misreporting their jobs’ parameters if it benefits them to do so. To mitigate this behavior, we design an incentive compatible online scheduling mechanism. Incentive compatibility assures us that the users will obtain the maximum payoff only if they truthfully report their jobs’ parameters to the scheduler. Finally, we simulate and study the mechanism to show the effects of misreports on the cheaters and on the system.

99 GENERAL AND MISCELLANEOUS↗

Fluid dynamics parallel computer development at NASA Langley Research Center

To accomplish more detailed simulations of highly complex flows, such as the transition to turbulence, fluid dynamics research requires computers much more powerful than any available today. Only parallel processing on multiple-processor computers offers hope for achieving the required effective speeds. Looking ahead to the use of these machines, the fluid dynamicist faces three issues: algorithm development for near-term parallel computers, architecture development for future computer power increases, and assessment of possible advantages of special purpose designs. Two projects at NASA Langley address these issues. Software development and algorithm exploration is being done on the FLEX/32 Parallel Processing Research Computer. New architecture features are being explored in the special purpose hardware design of the Navier-Stokes Computer. These projects are complementary and are producing promising results.

Townsend, James C.↗

HEP - A semaphore-synchronized multiprocessor with central control

The paper describes the design concept of the Heterogeneous Element Processor (HEP), a system tailored to the special needs of scientific simulation. In order to achieve high-speed computation required by simulation, HEP features a hierarchy of processes executing in parallel on a number of processors, with synchronization being largely accomplished by hardware. A full-empty-reserve scheme of synchronization is realized by zero-one-valued hardware semaphores. A typical system has, besides the control computer and the scheduler, an algebraic module, a memory module, a first-in first-out (FIFO) module, an integrator module, and an I/O module. The architecture of the scheduler and the algebraic module is examined in detail.

Gilliland, M. C.↗

SCORPIO: A Scalable Two-Phase Parallel I/O Library With Application To A Large Scale Subsurface Simulator

Inefficient parallel I/O is known to be a major bottleneck among scientific applications employed on supercomputers as the number of processor cores grows into the thousands. Our prior experience indicated that parallel I/O libraries such as HDF5 that rely on MPI-IO do not scale well beyond 10K processor cores, especially on parallel file systems (like Lustre) with single point of resource contention. Our previous optimization efforts for a massively parallel multi-phase and multi-component subsurface simulator (PFLOTRAN) led to a two-phase I/O approach at the application level where a set of designated processes participate in the I/O process by splitting the I/O operation into a communication phase and a disk I/O phase. The designated I/O processes are created by splitting the MPI global communicator into multiple sub-communicators. The root process in each sub-communicator is responsible for performing the I/O operations for the entire group and then distributing the data to rest of the group. This approach resulted in over 25X speedup in HDF I/O read performance and 3X speedup in write performance for PFLOTRAN at over 100K processor cores on the ORNL Jaguar supercomputer. This research describes the design and development of a general purpose parallel I/O library, SCORPIO (SCalable block-ORiented Parallel I/O) that incorporates our optimized two-phase I/O approach. The library provides a simplified higher level abstraction to the user, sitting atop existing parallel I/O libraries (such as HDF5) and implements optimized I/O access patterns that can scale on larger number of processors. Performance results with standard benchmark problems and PFLOTRAN indicate that our library is able to maintain the same speedups as before with the added flexibility of being applicable to a wider range of I/O intensive applications.

High Performance Computing↗

SCORPIO: A scalable two-phase parallel I/O library with application to a large scale subsurface simulator

Inefficient parallel I/O is known to be a major bottleneck among scientific applications employed on supercomputers as the number of processor cores grows into the thousands. Our prior experience indicated that parallel I/O libraries such as HDF5 that rely on MPI-IO do not scale well beyond 10K processor cores, especially on parallel file systems (like Lustre) with single point of resource contention. Our previous optimization efforts for a massively parallel multi-phase and multi-component subsurface simulator (PFLOTRAN) led to a two-phase I/O approach at the application level where a set of designated processes participate in the I/O process by splitting the I/O operation into a communication phase and a disk I/O phase. The designated I/O processes are created by splitting the MPI global communicator into multiple sub-communicators. The root process in each sub-communicator is responsible for performing the I/O operations for the entire group and then distributing the data to rest of the group. This approach resulted in over 25X speedup in HDF I/O read performance and 3X speedup in write performance for PFLOTRAN at over 100K processor cores on the ORNL Jaguar supercomputer. This research describes the design and development of a general purpose parallel I/O library called Scorpio that incorporates our optimized two-phase I/O approach. The library provides a simplified higher level abstraction to the user, sitting atop existing parallel I/O libraries (such as HDF5) and implements optimized I/O access patterns that can scale on larger number of processors. Performance results with standard benchmark problems and PFLOTRAN indicate that our library is able to maintain the same speedups as before with the added flexibility of being applicable to a wider range of I/O intensive applications.

Sreepathi, Sarat↗

Concurrent processing adaptation of aeroplastic analysis of propfans

Discussed here is a study involving the adaptation of an advanced aeroelastic analysis program to run concurrently on a shared memory multiple processor computer. The program uses a three-dimensional compressible unsteady aerodynamic model and blade normal modes to calculate aeroelastic stability and response of propfan blades. The identification of the computational parallelism within the sequential code and the scheduling of the concurrent subtasks to minimize processor idle time are discussed. Processor idle time in the calculation of the unsteady aerodynamic coefficients was reduced by the simple strategy of appropriately ordering the computations. Speedup and efficiency results are presented for the calculation of the matched flutter point of an experimental propfan model. The results show that efficiencies above 70 percent can be obtained using the present implementation with 7 processors. The parallel computational strategy described here is also applicable to other aeroelastic analysis procedures based on panel methods.

Janetzke, David C.↗

Concurrent processing adaptation of aeroelastic analysis of propfans

This paper reports on a study involving the adaptation of an advanced aeroelastic analysis program to run concurrently on a shared memory multiple processor computer. The program uses a three-dimensional compressible unsteady aerodynamic model and blade normal modes to calculate aeroelastic stability and response of propfan blades. The identification of the computational parallelism within the sequential code and the scheduling of the concurrent subtasks to minimize processor idle time are discussed. Processor idle time in the calculation of the unsteady aerodynamic coefficients was reduced by the simple strategy of appropriately ordering the computations. Speedup and efficiency results are presented for the calculation of the matched flutter point of an experimental propfan model. The results show that efficiencies above 70 percent can be obtained using the present implementation with 7 processors. The parallel computational strategy described here is also applicable to other aeroelastic analysis procedures based on parallel methods.

Murthy, Durbha V.↗