Energy Efficiency Modeling of Parallel Applications
Explore the source record for details and available documents.
Computer science↗
SEARCH · Search NASA
Results for “Parallelization efficiency”
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.
Workload-adaptive Scheduling for Efficient Use of Parallel File Systems in High-performance Computing Clusters
Explore the source record for details and available documents.
Goponenko, Alexander V.↗
Revisiting Parallel Cyclic Reduction and Parallel Prefix-Based Algorithms for Block Tridiagonal System of Equations
Simulations that require solutions of block tridiagonal systems of equations rely on fast parallel solvers for runtime efficiency. Leading parallel solvers that are highly effective for general systems of equations, dense or sparse, are limited in scalability when applied to block tridiagonal systems. This paper presents scalability results as well as detailed analyses of two parallel solvers that exploit the special structure of block tridiagonal matrices to deliver superior performance, often by orders of magnitude. A rigorous analysis of their relative parallel runtimes is shown to reveal the existence of a critical block size that separates the parameter space spanned by the number of block rows, the block size and the processor count, into distinct regions that favor one or the other of the two solvers. Dependence of this critical block size on the above parameters as well as on machine-specific constants is established. These formal insights are supported by empirical results on up to 2,048 cores of a Cray XT4 system. To the best of our knowledge, this is the highest reported scalability for parallel block tridiagonal solvers to date.
Seal, Sudip K↗
A massively parallel and memory-efficient FEM toolbox with a hybrid total FETI solver with accelerator support
In this article, we present the ExaScale PaRallel finite element tearing and interconnecting SOlver (ESPRESO) finite element method (FEM) library, which includes an FEM toolbox with interfaces to professional and open-source simulation tools, and a massively parallel hybrid total finite element tearing and interconnecting (HTFETI) solver which can fully utilize the Oak Ridge Leadership Computing Facility Titan supercomputer and achieve superlinear scaling. This article presents several new techniques for finite element tearing and interconnecting (FETI) solvers designed for efficient utilization of supercomputers with a focus on (i) performance—we present a fivefold reduction of solver runtime for the Laplace equation by redesigning the FETI solver and offloading the key workload to the accelerator. We compare Intel Xeon Phi 7120p and Tesla K80 and P100 accelerators to Intel Xeon E5-2680v3 and Xeon Phi 7210 central processing units; and (ii) memory efficiency—we present two techniques which increase the efficiency of the HTFETI solver 1.8 times and push the limits of the largest possible problem ESPRESO that can solve from 124 to 223 billion unknowns for problems with unstructured meshes. Finally, we show that by dynamically tuning hardware parameters, we can reduce energy consumption by up to 33%.
Computer Science↗
Highly Efficient Compensation-Based Parallelism for Wavefront Loops on GPUs
Wavefront loops are widely used in many scientific applications, e.g., partial differential equation (PDE) solvers and sequence alignment tools. However, due to the data dependencies in wavefront loops, it is challenging to fully utilize the abundant compute units of GPUs and to reuse data through their memory hierarchy. Existing solutions can only optimize for these factors to a limited extent. For example, tiling-based methods optimize memory access but may result in load imbalance; while compensation-based methods, which change the original order of computation to expose more parallelism and then compensate for it, suffer from both global synchronization overhead and limited generality. In this paper, we first prove under which circumstances that breaking data dependencies and properly changing the sequence of computation operators in our compensation-based method does not affect the correctness of results. Based on this analysis, we design a highly efficient compensation-based parallelism on GPUs. Our method provides weighted scan-based GPU kernels to optimize the computation and combines with the tiling method to optimize memory access and synchronization. The performance results on the NVIDIA K80 and P100 GPU platforms demonstrate that our method can achieve significant improvements for four types of real-world application kernels over the state-of-the-art research.
Hou, Kaixi↗
Quantum Package 2.0: An Open-Source Determinant-Driven Suite of Programs
Quantum chemistry is a discipline which relies heavily on very expensive numerical computations. The scaling of correlated wave function methods lies, in their standard implementation, between O(N-5) and O(e(N)), where N is proportional to the system size. Therefore, performing accurate calculations on chemically meaningful systems requires (i) approximations that can lower the computational scaling and (ii) efficient implementations that take advantage of modern massively parallel architectures. QUANTUM PACKAGE is an open-source programming environment for quantum chemistry specially designed for wave function methods. Its main goal is the development of determinant-driven selected configuration interaction (sCI) methods and multireference second-order perturbation theory (PT2). The determinant-driven framework allows the programmer to include any arbitrary set of determinants in the reference space, hence providing greater methodological freedom. The sCI method implemented in QUANTUM PACKAGE is based on the CIPSI (Configuration Interaction using a Perturbative Selection made Iteratively) algorithm which complements the variational sCI energy with a PT2 correction. Additional external plugins have been recently added to perform calculations with multireference coupled cluster theory and range-separated density-functional theory. All the programs are developed with the IRPF90 code generator, which simplifies collaborative work and the development of new features. QUANTUM PACKAGE strives to allow easy implementation and experimentation of new methods, while making parallel computation as simple and efficient as possible on modern supercomputer architectures. Currently, the code enables, routinely, to realize runs on roughly 2 000 CPU cores, with tens of millions of determinants in the reference space. Moreover, we have been able to push up to 12 288 cores in order to test its parallel efficiency. In the present manuscript, we also introduce some key new developments: (i) a renormalized second-order perturbative correction for efficient extrapolation to the full CI limit and (ii) a stochastic version of the CIPSI selection performed simultaneously to the PT2 calculation at no extra cost.
Garniron, Yann↗
A Scalable Parallel Wideband MLFMA for Efficient Electromagnetic Simulations on Large Scale Clusters
The development of the multilevel fast multipole algorithm (MLFMA) and its multiscale variants have enabled the use of integral equation (IE) based solvers to compute scattering from complicated structures. Development of scalable parallel algorithms, to extend the reach of these solvers, has been a topic of intense research for about a decade. In this paper, we present a new algorithm for parallel implementation of IE solver that is augmented with a wideband MLFMA and scalable on large number of processors. The wideband MLFMA employed here, to handle multiscale problems, is a hybrid combination of the accelerated Cartesian expansion (ACE) and the classical MLFMA. The salient feature of the presented parallel algorithm is that it is implicitly load balanced and exhibits higher performance. This is achieved by developing a strategy to partition the MLFMA tree, and hence the associated computations, in a self-similar fashion among the parallel processors. As detailed in the paper, the algorithm employs both spatial and direction partitioning approaches in a flexible manner to ensure scalable performance. Plethora of results are presented here to exhibit the scalability of this algorithm on 512 and more processors.
99 GENERAL AND MISCELLANEOUS↗
A domain wall-magnetic tunnel junction artificial synapse with notched geometry for accurate and efficient training of deep neural networks
Inspired by the parallelism and efficiency of the brain, several candidates for artificial synapse devices have been developed for neuromorphic computing, yet a nonlinear and asymmetric synaptic response curve precludes their use for backpropagation, the foundation of modern supervised learning. Spintronic devices—which benefit from high endurance, low power consumption, low latency, and CMOS compatibility—are a promising technology for memory, and domain-wall magnetic tunnel junction (DW-MTJ) devices have been shown to implement synaptic functions such as long-term potentiation and spike-timing dependent plasticity. In this work, we propose a notched DW-MTJ synapse as a candidate for supervised learning. Using micromagnetic simulations at room temperature, we show that notched synapses ensure the non-volatility of the synaptic weight and allow for highly linear, symmetric, and reproducible weight updates using either spin transfer torque (STT) or spin–orbit torque (SOT) mechanisms of DW propagation. We use lookup tables constructed from micromagnetics simulations to model the training of neural networks built with DW-MTJ synapses on both the MNIST and Fashion-MNIST image classification tasks. Accounting for thermal noise and realistic process variations, the DW-MTJ devices achieve classification accuracy close to ideal floating-point updates using both STT and SOT devices at room temperature and at 400 K. Our work establishes the basis for a magnetic artificial synapse that can eventually lead to hardware neural networks with fully spintronic matrix operations implementing machine learning.
42 ENGINEERING↗
A Method to Alleviate the Long History Problem Encountered in Monte Carlo Simulations via Weight Window Variance Reduction
When the weight window is implemented in large and complex models, streaming path and deep penetration could lead to a long history problem that takes a disproportionate amount of time for the accomplishment of Monte Carlo simulations and exerts a detrimental effect on the efficiency of parallel computing. An approach set the limitation of weight window splitting was proposed for the alleviation of long history problem. Tests were conducted on a streaming dog-leg geometry model and a 3D model of the Chinese Fusion Engineering Testing Reactor. The results show that a suitable parameter in the alleviation method could optimize both the performance of variance reduction and the efficiency of parallel calculation, which makes long history problem tractable.
70 PLASMA PHYSICS AND FUSION TECHNOLOGY↗
Image Gradient Decomposition for Parallel and Memory-Efficient Ptychographic Reconstruction
Ptychography is a popular microscopic imaging modality for many scientific discoveries and sets the record for highest image resolution. Unfortunately, the high image resolution for ptychographic reconstruction requires significant amount of memory and computations, forcing many applications to compromise their image resolution in exchange for a smaller memory footprint and a shorter reconstruction time. In this paper, we propose a novel image gradient decomposition method that significantly reduces the memory footprint for ptychographic reconstruction by tessellating image gradients and diffraction measurements into tiles. In addition, we propose a parallel image gradient decomposition method that enables asynchronous point-to-point communications and parallel pipelining with minimal overhead on a large number of GPUs. Our experiments on a Titanate material dataset (PbTiO3) with 16632 probe locations show that our Gradient Decomposition algorithm reduces memory footprint by 51 times. In addition, it achieves time-to-solution within 2.2 minutes by scaling to 4158 GPUs with a super-linear strong scaling efficiency at 364% compared to runtimes at 6 GPUs. This performance is 2.7 times more memory efficient, 9 times more scalable and 86 times faster than the state-of-the-art algorithm.
Wang, Xiao↗
Parallel elastic elements improve energy efficiency on the STEPPR bipedal walking robot
This study describes how parallel elastic elements can be used to reduce energy consumption in the electric motor driven, fully-actuated, STEPPR bipedal walking robot without compromising or significantly limiting locomotive behaviors. A physically motivated approach is used to illustrate how selectively-engaging springs for hip adduction and ankle flexion predict benefits for three different flat ground walking gaits: human walking, human-like robot walking and crouched robot walking. Based on locomotion data, springs are designed and substantial reductions in power consumption are demonstrated using a bench dynamometer. These lessons are then applied to STEPPR (Sandia Transmission-Efficient Prototype Promoting Research), a fully actuated bipedal robot designed to explore the impact of tailored joint mechanisms on walking efficiency. Featuring high-torque brushless DC motors, efficient low-ratio transmissions, and high fidelity torque control, STEPPR provides the ability to incorporate novel joint-level mechanisms without dramatically altering high level control. Unique parallel elastic designs are incorporated into STEPPR, and walking data shows that hip adduction and ankle flexion springs significantly reduce the required actuator energy at those joints for several gaits. These results suggest that parallel joint springs offer a promising means of supporting quasi-static joint torques due to body mass during walking, relieving motors of the need to support these torques and substantially improving locomotive energy efficiency.
32 ENERGY CONSERVATION, CONSUMPTION, AND UTILIZATI↗
Massively parallel Monte Carlo for many-particle simulations on GPUs
Current trends in parallel processors call for the design of efficient massively parallel algorithms for scientific computing. Parallel algorithms for Monte Carlo simulations of thermodynamic ensembles of particles have received little attention because of the inherent serial nature of the statistical sampling. In this paper, we present a massively parallel method that obeys detailed balance and implement it for a system of hard disks on the GPU. We reproduce results of serial high-precision Monte Carlo runs to verify the method. This is a good test case because the hard disk equation of state over the range where the liquid transforms into the solid is particularly sensitive to small deviations away from the balance conditions. On a Tesla K20, our GPU implementation executes over one billion trial moves per second, which is 148 times faster than on a single Intel Xeon E5540 CPU core, enables 27 times better performance per dollar, and cuts energy usage by a factor of 13. With this improved performance we are able to calculate the equation of state for systems of up to one million hard disks. These large system sizes are required in order to probe the nature of the melting transition, which has been debated for the last forty years. In this paper we present the details of our computational method, and discuss the thermodynamics of hard disks separately in a companion paper.
97 MATHEMATICS AND COMPUTING↗
Phaser.MRage: automated molecular replacement
The functionality of the molecular-replacement pipeline phaser.MRage is introduced and illustrated with examples. Phaser.MRage is a molecular-replacement automation framework that implements a full model-generation workflow and provides several layers of model exploration to the user. It is designed to handle a large number of models and can distribute calculations efficiently onto parallel hardware. In addition, phaser.MRage can identify correct solutions and use this information to accelerate the search. Firstly, it can quickly score all alternative models of a component once a correct solution has been found. Secondly, it can perform extensive analysis of identified solutions to find protein assemblies and can employ assembled models for subsequent searches. Thirdly, it is able to use a priori assembly information (derived from, for example, homologues) to speculatively place and score molecules, thereby customizing the search procedure to a certain class of protein molecule (for example, antibodies) and incorporating additional biological information into molecular replacement.
75 CONDENSED MATTER PHYSICS, SUPERCONDUCTIVITY AND↗
A Parallel Alternative for Energy-Efficient Neural Network Training and Inferencing
Energy efficiency of training and inferencing with large neural network models is a critical challenge facing the future of sustainable large-scale machine learning workloads. This paper introduces an alternative strategy, called phantom parallelism, to minimize the net energy consumption of traditional tensor (model) parallelism, the most energy-inefficient component of large neural network training. The approach is presented in the context of feed-forward network architectures as a preliminary, but comprehensive, proof-of-principle study of the proposed methodology. We derive new forward and backward propagation operators for phantom parallelism, implement them as custom autograd operations within an end-to-end phantom parallel training pipeline and compare its parallel performance and energy-efficiency against those of conventional tensor parallel training pipelines. Formal analyses that predict lower bandwidth and FLOP counts are presented with supporting empirical results on up to 256 GPUs that corroborate these gains. Experiments are shown to deliver ∼50% reduction in the energy consumed to train FFNs using the proposed phantom parallel approach when compared with conventional tensor parallel methods. Additionally, the proposed approach is shown to train smaller phantom models to the same model loss on smaller GPU counts as larger tensor parallel models on larger GPU counts offering the possibility for even greater energy savings.
Seal, Sudip [ORNL] (ORCID:0000000332330656)↗
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↗
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.
Smith, Brian E.↗
Parallel deterministic transport sweeps of structured and unstructured meshes with overloaded mesh decompositions
Here, the efficiency of discrete ordinates transport sweeps depends on the scheduling algorithm, the domain decomposition, the problem to be solved, and the computational platform. Sweep scheduling algorithms may be categorized by their approach to several issues. In this paper we examine the strategy of domain overloading for mesh partitioning as one of the components of such algorithms. In particular, we extend the domain overloading strategy, previously defined and analyzed for structured meshes, to the general case of unstructured meshes. We also present computational results for both the structured and unstructured domain overloading cases. We find that an appropriate amount of domain overloading can greatly improve the efficiency of parallel sweeps for both structured and unstructured partitionings of the test problems examined on up to 10 5 processor cores.
97 MATHEMATICS AND COMPUTING↗
fftMPI, a library for performing 2d and 3d FFTs in parallel
The fftMPI library enables two and three dimensional Fast Fourier Transforms (FFTs) to be performed efficiently in parallel on a variety of platforms from desktop machines to supercomputers. It allows for flexibility in the initial distribution of the FFT grid across processors, as well as the final output. It handles the data communication and reordering necessary to perform multi-dimensional FFTs in parallel. It relies on open-source or vendor-provided libraries to perform one dimensional FFTs on a single processor. It uses MPI for inter-processor communication, which is a standard distributed memory message-passing library.
Plimpton, Steven↗