Search NASA⌕ Search

SEARCH · Search NASA

Results for “Computational Load Balancing”

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 55 records · Page 3

Asynchronous distributed-memory task-parallel algorithm for compressible flows on unstructured 3D Eulerian grids

Here, we discuss the implementation of a finite element method, used to numerically solve the Euler equations of compressible flows, using an asynchronous runtime system (RTS). The algorithm is implemented for distributed-memory machines, using stationary unstructured 3D meshes, combining data-, and task-parallelism on top of the Charm++ RTS. Charm++’s execution model is asynchronous by default, allowing arbitrary overlap of computation and communication. Task-parallelism allows scheduling parts of an algorithm independently of, or dependent on, each other. Built-in automatic load balancing enables continuous redistribution of computational load by migration of work units based on real-time CPU load measurement. The RTS also features automatic checkpointing, fault tolerance, resilience against hardware failure, and supports power-, and energy-aware computation. We demonstrate scalability up to 25 x 10 9 cells at $\mathscr{O}$10 4 compute cores and the benefits of automatic load balancing for irregular workloads. The full source code with documentation is available at https://quinoacomputing.org.

42 ENGINEERING↗

Planning of distributed data production for High Energy and Nuclear Physics

Modern experiments in High Energy and Nuclear Physics heavily rely on distributed computations using multiple computational facilities across the world. One of the essential types of the computations is a distributed data production where petabytes of raw files from a single source has to be processed once (per production campaign) using thousands of CPUs at distant locations and the output has to be transferred back to that source. The data distribution over a large system does not necessary match the distribution of storage, network and CPU capacity. Therefore, bottlenecks may appear and lead to increased latency and degraded performance. In this paper we propose a new scheduling approach for distributed data production which is based on the network flow maximization model. In our approach a central planner defines how much input and output data should be transferred over each network link in order to maximize the computational throughput. Such plans are created periodically for a fixed planning time interval using up-to-date information on network, storage and CPU resources. The centrally created plans are executed in a distributed manner by dedicated services running at participating sites. In conclusion, our simulations based on the log records from the data production framework of the experiment STAR (Solenoid Tracker at RHIC) have shown that the proposed model systematically provides a better performance compared to the simulated traditional techniques.

73 NUCLEAR PHYSICS AND RADIATION PHYSICS↗

Scalable Implicit Solvers with Dynamic Mesh Adaptation for a Relativistic Drift-Kinetic Fokker–Planck–Boltzmann Model

In this work we consider a relativistic drift-kinetic model for runaway electrons along with a Fokker–Planck operator for small-angle Coulomb collisions, a radiation damping operator, and a secondary knock-on (Boltzmann) collision source. Here, we develop a new scalable fully implicit solver utilizing finite volume and conservative finite difference schemes and dynamic mesh adaptivity. A new data management framework in the PETSc library based on the p4est library is developed to enable simulations with dynamic adaptive mesh refinement (AMR), distributed memory parallelization, and dynamic load balancing of computational work. This framework and the runaway electron solver building on the framework are able to dynamically capture both bulk Maxwellian at the low-energy region and a runaway tail at the high-energy region. To effectively capture features via the AMR algorithm, a new AMR indicator prediction strategy is proposed that is performed alongside the implicit time evolution of the solution. This strategy is complemented by the introduction of computationally cheap feature-based AMR indicators that are analyzed theoretically. Numerical results quantify the advantages of the prediction strategy in better capturing features compared with nonpredictive strategies; and we demonstrate trade-offs regarding computational costs. The robustness with respect to model parameters, algorithmic scalability, and parallel scalability are demonstrated through several benchmark problems including manufactured solutions and solutions of different physics models. We focus on demonstrating the advantages of using implicit time stepping and AMR for runaway electron simulations.

70 PLASMA PHYSICS AND FUSION TECHNOLOGY↗

The development of laser-plasma interaction program LAP3D on thousands of processors

Modeling laser-plasma interaction (LPI) processes in real-size experiments scale is recognized as a challenging task. For explorering the influence of various instabilities in LPI processes, a three-dimensional laser and plasma code (LAP3D) has been developed, which includes filamentation, stimulated Brillouin backscattering (SBS), stimulated Raman backscattering (SRS), non-local heat transport and plasmas flow computation modules. In this program, a second-order upwind scheme is applied to solve the plasma equations which are represented by an Euler fluid model. Operator splitting method is used for solving the equations of the light wave propagation, where the Fast Fourier translation (FFT) is applied to compute the diffraction operator and the coordinate translations is used to solve the acoustic wave equation. The coupled terms of the different physics processes are computed by the second-order interpolations algorithm. In order to simulate the LPI processes in massively parallel computers well, several parallel techniques are used, such as the coupled parallel algorithm of FFT and fluid numerical computation, the load balance algorithm, and the data transfer algorithm. Now the phenomena of filamentation, SBS and SRS have been studied in low-density plasma successfully with LAP3D. Scalability of the program is demonstrated with a parallel efficiency above 50% on about ten thousand of processors.

70 PLASMA PHYSICS AND FUSION TECHNOLOGY↗

Web-based, GPU-accelerated, Monte Carlo simulation and visualization of indirect radiation imaging detector performance

Purpose: Monte Carlo simulations play a vital role in the understanding of the fundamental limitations, design, and optimization of existing and emerging medical imaging systems. Efforts in this area have resulted in the development of a wide variety of open-source software packages. One such package, hybridMANTIS, uses a novel hybrid concept to model indirect scintillator detectors by balancing the computational load using dual CPU and graphics processing unit (GPU) processors, obtaining computational efficiency with reasonable accuracy. In this work, the authors describe two open-source visualization interfaces, webMANTIS and visualMANTIS to facilitate the setup of computational experiments via hybridMANTIS. Methods: The visualization tools visualMANTIS and webMANTIS enable the user to control simulation properties through a user interface. In the case of webMANTIS, control via a web browser allows access through mobile devices such as smartphones or tablets. webMANTIS acts as a server back-end and communicates with an NVIDIA GPU computing cluster that can support multiuser environments where users can execute different experiments in parallel. Results: The output consists of point response and pulse-height spectrum, and optical transport statistics generated by hybridMANTIS. The users can download the output images and statistics through a zip file for future reference. In addition, webMANTIS provides a visualization window that displays a few selected optical photon path as they get transported through the detector columns and allows the user to trace the history of the optical photons. Conclusions: The visualization tools visualMANTIS and webMANTIS provide features such as on the fly generation of pulse-height spectra and response functions for microcolumnar x-ray imagers while allowing users to save simulation parameters and results from prior experiments. The graphical interfaces simplify the simulation setup and allow the user to go directly from specifying input parameters to receiving visual feedback for the model predictions.

46 INSTRUMENTATION RELATED TO NUCLEAR SCIENCE AND ↗

Concurrent processing simulation of the space station

The development of a new capability for the time-domain simulation of multibody dynamic systems and its application to the study of a large angle rotational maneuvers of the Space Station is described. The effort was divided into three sequential tasks, which required significant advancements of the state-of-the art to accomplish. These were: (1) the development of an explicit mathematical model via symbol manipulation of a flexible, multibody dynamic system; (2) the development of a methodology for balancing the computational load of an explicit mathematical model for concurrent processing; and (3) the implementation and successful simulation of the above on a prototype Custom Architectured Parallel Processing System (CAPPS) containing eight processors. The throughput rate achieved by the CAPPS operating at only 70 percent efficiency, was 3.9 times greater than that obtained sequentially by the IBM 3090 supercomputer simulating the same problem. More significantly, analysis of the results leads to the conclusion that the relative cost effectiveness of concurrent vs. sequential digital computation will grow substantially as the computational load is increased. This is a welcomed development in an era when very complex and cumbersome mathematical models of large space vehicles must be used as substitutes for full scale testing which has become impractical.

Gluck, R.↗

Nonrecursive formulations of multibody dynamics and concurrent multiprocessing

Since the late 1980's, research in recursive formulations of multibody dynamics has flourished. Historically, much of this research can be traced to applications of low dimensionality in mechanism and vehicle dynamics. Indeed, there is little doubt that recursive order N methods are the method of choice for this class of systems. This approach has the advantage that a minimal number of coordinates are utilized, parallelism can be induced for certain system topologies, and the method is of order N computational cost for systems of N rigid bodies. Despite the fact that many authors have dismissed redundant coordinate formulations as being of order N(exp 3), and hence less attractive than recursive formulations, we present recent research that demonstrates that at least three distinct classes of redundant, nonrecursive multibody formulations consistently achieve order N computational cost for systems of rigid and/or flexible bodies. These formulations are as follows: (1) the preconditioned range space formulation; (2) penalty methods; and (3) augmented Lagrangian methods for nonlinear multibody dynamics. The first method can be traced to its foundation in equality constrained quadratic optimization, while the last two methods have been studied extensively in the context of coercive variational boundary value problems in computational mechanics. Until recently, however, they have not been investigated in the context of multibody simulation, and present theoretical questions unique to nonlinear dynamics. All of these nonrecursive methods have additional advantages with respect to recursive order N methods: (1) the formalisms retain the highly desirable order N computational cost; (2) the techniques are amenable to concurrent simulation strategies; (3) the approaches do not depend upon system topology to induce concurrency; and (4) the methods can be derived to balance the computational load automatically on concurrent multiprocessors. In addition to the presentation of the fundamental formulations, this paper presents new theoretical results regarding the rate of convergence of order N constraint stabilization schemes associated with the newly introduced class of methods.

Kurdila, Andrew J.↗

Fast Multilevel Implementation of Recursive Spectral Bisection for Partitioning Unstructured Problems

If problems involving unstructured meshes are to be solved efficiently on distributed-memory parallel computers, the meshes must be partitioned and distributed across processors in a way that balances tile computational load and minimizes communication. The recursive spectral bisection method (RSB) has been shown to be very effective for such partitioning problems compared to alternative methods, but RSB in its simplest form is expensive. Here a multilevel version of RSB is introduced that attains about an order-of-magnitude improvement in run time on typical examples.

Barnard, Stephen T.↗

SSME alternate turbopump (pump section) axial load analysis

A flow balancing computer program constructed to calculate the axial loads on the Space Shuttle Main Engine (SSME) alternate turbopumps (ATs) pump sections are described. The loads are used in turn to determine load balancing piston design requirements. The application of the program to the inlet section, inducer/impeller/stage, bearings, seals, labyrinth, damper, piston, face and corner, and stationary/rotating surfaces is indicated. Design analysis results are reported which show that the balancing piston's designs are adequate and that performance and life will not be degraded by the turbopump's axial load characteristics.

Crease, G. A.↗

Dynamic Load Balancing for Finite Element Calculations on Parallel Computers

Computational requirements of full scale computational fluid dynamics change as computation progresses on a parallel machine. The change in computational intensity causes workload imbalance of processors, which in turn requires a large amount of data movement at runtime. If parallel CFD is to be successful on a parallel or massively parallel machine, balancing of the runtime load is indispensable. Here a frame work is presented for dynamic load balancing for CFD applications, called Jove. One processor is designated as a decision maker Jove while others are assigned to computational fluid dynamics. Processors running CFD send flags to Jove in a predetermined number of iterations to initiate load balancing. Jove starts working on load balancing while other processors continue working with the current data and load distribution. Jove goes through several steps to decide if the new data should be taken, including preliminary evaluate, partition, processor reassignment, cost evaluation, and decision. Jove running on a single SP2 node has been completely implemented. Preliminary experimental results show that the Jove approach to dynamic load balancing can be effective for full scale grid partitioning on the target machine SP2.

Pramono, Eddy↗

Performance Evaluation of Counter-Based Dynamic Load Balancing Schemes for Massive Contingency Analysis with Different Computing Environments

Contingency analysis is a key function in the Energy Management System (EMS) to assess the impact of various combinations of power system component failures based on state estimation. Contingency analysis is also extensively used in power market operation for feasibility test of market solutions. High performance computing holds the promise of faster analysis of more contingency cases for the purpose of safe and reliable operation of today’s power grids with less operating margin and more intermittent renewable energy sources. This paper evaluates the performance of counter-based dynamic load balancing schemes for massive contingency analysis under different computing environments. Insights from the performance evaluation can be used as guidance for users to select suitable schemes in the application of massive contingency analysis. Case studies, as well as MATLAB simulations, of massive contingency cases using the Western Electricity Coordinating Council power grid model are presented to illustrate the application of high performance computing with counter-based dynamic load balancing schemes.

24 POWER TRANSMISSION AND DISTRIBUTION↗

Adaptive Load-Balancing Algorithms Using Symmetric Broadcast Networks

In a distributed-computing environment, it is important to ensure that the processor workloads are adequately balanced. Among numerous load-balancing algorithms, a unique approach due to Dam and Prasad defines a symmetric broadcast network (SBN) that provides a robust communication pattern among the processors in a topology-independent manner. In this paper, we propose and analyze three novel SBN-based load-balancing algorithms, and implement them on an SP2. A thorough experimental study with Poisson-distributed synthetic loads demonstrates that these algorithms are very effective in balancing system load while minimizing processor idle time. They also compare favorably with several other existing load-balancing techniques. Additional experiments performed with real data demonstrate that the SBN approach is effective in adaptive computational science and engineering applications where dynamic load balancing is extremely crucial.

Das, Sajal K.↗

Unstructured Adaptive Grid Computations on an Array of SMPs

Dynamic load balancing is necessary for parallel adaptive methods to solve unsteady CFD problems on unstructured grids. We have presented such a dynamic load balancing framework called JOVE, in this paper. Results on a four-POWERnode POWER CHALLENGEarray demonstrated that load balancing gives significant performance improvements over no load balancing for such adaptive computations. The parallel speedup of JOVE, implemented using MPI on the POWER CHALLENCEarray, was significant, being as high as 31 for 32 processors. An implementation of JOVE that exploits 'an array of SMPS' architecture was also studied; this hybrid JOVE outperformed flat JOVE by up to 28% on the meshes and adaption models tested. With large, realistic meshes and actual flow-solver and adaption phases incorporated into JOVE, hybrid JOVE can be expected to yield significant advantage over flat JOVE, especially as the number of processors is increased, thus demonstrating the scalability of an array of SMPs architecture.

Biswas, Rupak↗

BALANCING THE LOAD: A VORONOI BASED SCHEME FOR PARALLEL COMPUTATIONS

One of the key issues when running a simulation on multiple CPUs is maintaining a proper load balance throughout the run and minimizing communications between CPUs. We propose a novel method of utilizing a Voronoi diagram to achieve a nearly perfect load balance without the need of any global redistributions of data. As a show case, we implement our method in RICH, a two-dimensional moving mesh hydrodynamical code, but it can be extended trivially to other codes in two or three dimensions. Our tests show that this method is indeed efficient and can be used in a large variety of existing hydrodynamical codes.

79 ASTRONOMY AND ASTROPHYSICS↗

Multitasking a three-dimensional Navier-Stokes algorithm on the Cray-2

A three-dimensional computational aerodynamics algorithm has been multitasked for efficient parallel execution on the Cray-2. It provides a means for examining the multitasking performance of a complete CFD application code. An embedded zonal multigrid scheme is used to solve the Reynolds-averaged Navier-Stokes equations for an internal flow model problem. The explicit nature of each component of the method allows a spatial partitioning of the computational domain to achieve a well-balanced task load for MIMD computers with vector-processing capability. Experiments have been conducted with both two- and three-dimensional multitasked cases. The best speedup attained by an individual task group was 3.54 on four processors of the Cray-2, while the entire solver yielded a speedup of 2.67 on four processors for the three-dimensional case. The multiprocessing efficiency of various types of computational tasks is examined, performance on two Cray-2s with different memory access speeds is compared, and extrapolation to larger problems is discussed.

Swisshelm, Julie M.↗

Dynamic Load Balancing on Single- and Multi-GPU Systems

The computational power provided by many-core graphics processing units (GPUs) has been exploited in many applications. The programming techniques supported and employed on these GPUs are not sufficient to address problems exhibiting irregular, and unbalanced workload. The problem is exacerbated when trying to effectively exploit multiple GPUs, which are commonly available in many modern systems. In this paper, we propose a task-based dynamic load-balancing solution for single- and multi-GPU systems. The solution allows load balancing at a finer granularity than what is supported in existing APIs such as NVIDIA’s CUDA. We evaluate our approach using both micro-benchmarks and a molecular dynamics application that exhibits significant load imbalance. Experimental results with a single-GPU configuration show that our fine-grained task solution can utilize the hardware more efficiently than the CUDA scheduler for unbalanced workload. On multi-GPU systems, our solution achieves near-linear speedup, load balance, and significant performance improvement over techniques based on standard CUDA APIs.

97 MATHEMATICS AND COMPUTING↗

Dynamic load balancing with enhanced shared-memory parallelism for particle-in-cell codes

Furthering our understanding of many of today’s interesting problems in plasma physics – including plasma based acceleration and magnetic reconnection with pair production due to quantum electrodynamic effects – requires large-scale kinetic simulations using particle-in-cell (PIC) codes. However, these simulations are extremely demanding, requiring that contemporary PIC codes be designed to efficiently use a new fleet of exascale computing architectures. To this end, the key issue of parallel load balance across computational nodes must be addressed. We discuss the implementation of dynamic load balancing by dividing the simulation space into many small, self-contained regions or ‘‘tiles,’’ along with shared-memory (e.g., OpenMP) parallelism both over many tiles and within single tiles. The load balancing algorithm can be used with three different topologies, including two space-filling curves. Here, we tested this implementation in the code Osiris and show low overhead and improved scalability with OpenMP thread number on simulations with both uniform load and severe load imbalance. Compared to other load-balancing techniques, our algorithm gives order-of-magnitude improvement in parallel scalability for simulations with severe load imbalance issues.

97 MATHEMATICS AND COMPUTING↗