Many-core Graphics Processing Units Research Articles

Synchronous speculative simulation of tightly coupled agents in continuous time on CPUs and GPUs

Traditionally, parallel discrete-event simulations of agent-based models in continuous time are organized around logical processes exchanging time-stamped events, which clashes with the properties of models in which tightly coupled agents frequently and instantaneously access each other’s states. To illustrate the challenges of such models and to derive a solution, we consider the domain-specific modeling language ML3, which allows modelers to succinctly express transitions and interactions of linked agents based on a continuous-time Markov chain (CTMC) semantics. We propose synchronous optimistic synchronization algorithms tailored toward simulations of fine-grained interactions among tightly coupled agents in highly dynamic topologies and present implementations targeting multicore central processing units (CPUs) as well as many-core graphics processing units (GPUs). By dynamically restricting the temporal progress per round to ensure that at most one transition or state access per agent, the synchronization algorithms enable efficient direct agent interaction and limit the required agent state history to only a single current and projected state. To maintain concurrency given actions that depend on dynamically updated macro-level properties, we introduce a simple relaxation scheme with guaranteed error bounds. Using an extended variant of the classical susceptible-infected-recovered network model, we benchmark and profile the performance of the different algorithms running on CPUs and on a data center GPU.

Predictive modeling and analysis of runout distance of physical mudflows based on a discrete element method

ABSTRACT Both visualization simulation and numerical simulation play important roles in mudflow simulation. However, the computational bottlenecks and the correctness of mudflow simulation supporting a large-scale particle system are still key problems to be solved in modeling of mudflow dynamics. In this article, we propose a discrete element method that uses many-core graphics processing units to accelerate the simulation of such a depositional process and to predict runout distances for physical mudflows. A model constructed using this method was verified for accuracy and applicability by a comparative analysis of physical mudflow flume tests and corresponding simulation tests. Our results showed that the runout distance prediction model, constructed using the discrete element method, can simulate the depositional processes of physical mudflows (supporting the simulation of tens of millions of particles), and can simulate their runout distance and deposit morphologies. Because the heterogeneous CPU–GPU computing approach can effectively improve the simulation accuracy by using more mudflow particles, the error in the simulation of the runout distance of the physical mudflows was within a small and acceptable range. Thus, the model can be used for further studies on the prediction of risks associated with mudflows.

CUDAQuat: new parallel framework for fast computation of quaternion moments for color images applications

Quaternion moments are widely used in several applications, such as image classification, object recognition, and multimedia security. The computation of these moments requires a vast computational time, especially for big-size images. Several attempts to accelerate quaternion moments are not enough to process big-size color images with the desired speedup. In this work, we proposed a new parallel framework for fast computation of quaternion moments in Cartesian coordinates using multi-core CPUs and many-core graphics processing units (GPUs) with the Compute Unified Device Architecture (CUDA). We called the proposed unified computational framework “CUDAQuat.” This framework was tested by eleven sets of quaternion moments. Several applications executed using the proposed parallel framework where the CPU times, execution-time-improvement ratio, and speedup were reported. The evaluation outlined significant speedup over the single-core CPU implementation, where the proposed framework accelerated several sets of quaternion moments with speedup 600x.

Adaptive Heterogeneous Transient Analysis of Wind Farm Integrated Comprehensive AC/DC Grids

The increasingly complex AC/DC network as a result of the massive integration of wind farms manifests the significance of a comprehensive transient study. In this work, the wind turbine (WT) and the DC grid are modeled in detail for the electromagnetic transient (EMT) simulation to maximize its fidelity, whilst the AC grid transient stability is analyzed by dynamic simulation (DS). An interactive EMT-DS interface is thus introduced to enable their concurrency and subsequently form a co-simulation. The CPU which is dominant in system study faces a tremendous challenge in handling a great number of components albeit they exhibit homogeneity. The many-core graphics processing unit (GPU) featuring massive parallelism is therefore exploited and following the definition of an adaptive computing boundary, a flexible heterogeneous sequential-parallel processing architecture is proposed for efficient analysis of the wind-farm-integrated AC/DC grid. Topological reconfiguration of WTs is specifically carried out to reduce the numerical order whilst enhancing system homogeneity that enables the GPU to thoroughly utilize its peculiar property of single-instruction multiple-thread (SIMT) compute paradigm. Consequently, significant speedups can be attained by the proposed computing framework over pure CPU computation, while its accuracy is validated by the commercial EMT and dynamic security analysis tools PSCAD/EMTDC and DSATools, respectively.

High Performance GPU Tensor Completion With Tubal-Sampling Pattern

Data completion is a problem of filling missing or unobserved elements of partially observed datasets. Data completion algorithms have received wide attention and achievements in diverse domains including data mining, signal processing, and computer vision. We observe a ubiquitous tubal-sampling pattern in big data and Internet of Things (IoT) applications, which is introduced by many reasons such as high data acquisition cost, downsampling for data compression, sensor node failures, and packet losses in low-power wireless transmissions. To meet the time and accuracy requirements of applications, data completion methods are expected to be accurate as well as fast. However, the existing methods for data completion with the tubal-sampling pattern are either accurate or fast, but not both. In this article, we propose high-performance graphics processing unit (GPU) tensor completion for data completion with the tubal-sampling pattern. First, by exploiting the convolution theorem, we split a tensor least-squares minimization problem into multiple least-squares sub-problems in the frequency domain. In this way, massive parallelisms are exposed for many-core GPU architectures while still preserving high recovery accuracy. Second, we propose computing slice-level and tube-level tasks in batches to improve GPU utilization. Third, we reduce the data transfer cost by eliminating the accesses to the CPU memory inside algorithm loop structures. The experimental results show that the proposed tensor completion is both fast and accurate. Using synthetic data of varying sizes, the proposed GPU tensor completion achieves maximum 248.18×, 7, 403.27×, and 33.27× speedups over the CPU MATLAB implementation, GPU element-sampling tensor completion in the cuTensor-tubal library, and GPU high-performance matrix completion, respectively. With a 50 percent sampling rate, the proposed GPU tensor completion achieves a recovery error of 1.40e-5, which is comparable with that of the GPU element-sampling tensor completion and three orders of magnitude better than that of the GPU high-performance matrix completion. To utilize multiple GPUs in servers, we design a multi-GPU scheme for tubal-sampling tensor completion. The multi-GPU tensor completion achieves maximum 1.89× speedup on two GPUs versus on a single GPU for medium or big tensors. We further evaluate the performance of the proposed GPU tensor completion in three real applications, namely, video transmission in wireless camera networks, RF fingerprint-based indoor localization, and seismic data completion, and it achieves maximum speedups of 448.68×, 24.63×, and 311.54×, respectively. We integrate this high-performance GPU tensor completion implementation into the cuTensor-tubal library to support various applications.

Energy Efficiency of Machine Learning in Embedded Systems Using Neuromorphic Hardware

Recently, the application of machine learning on embedded systems has drawn interest in both the research community and industry because embedded systems located at the edge can produce a faster response and reduce network load. However, software implementation of neural networks on Central Processing Units (CPUs) is considered infeasible in embedded systems due to limited power supply. To accelerate AI processing, the many-core Graphics Processing Unit (GPU) has been a preferred device to the CPU. However, its energy efficiency is not still considered to be good enough for embedded systems. Among other approaches for machine learning on embedded systems, neuromorphic processing chips are expected to be less power-consuming and overcome the memory bottleneck. In this work, we implemented a pedestrian image detection system on an embedded device using a commercially available neuromorphic chip, NM500, which is based on NeuroMem technology. The NM500 processing time and the power consumption were measured as the number of chips was increased from one to seven, and they were compared to those of a multicore CPU system and a GPU-accelerated embedded system. The results show that NM500 is more efficient in terms of energy required to process data for both learning and classification than the GPU-accelerated system or the multicore CPU system. Additionally, limits and possible improvement of the current NM500 are identified based on the experimental results.

Parallelizing Multiple Flow Accumulation Algorithm using CUDA and OpenACC

Watershed analysis, as a fundamental component of digital terrain analysis, is based on the Digital Elevation Model (DEM), which is a grid (raster) model of the Earth surface and topography. Watershed analysis consists of computationally and data intensive computing algorithms that need to be implemented by leveraging parallel and high-performance computing methods and techniques. In this paper, the Multiple Flow Direction (MFD) algorithm for watershed analysis is implemented and evaluated on multi-core Central Processing Units (CPU) and many-core Graphics Processing Units (GPU), which provides significant improvements in performance and energy usage. The implementation is based on NVIDIA CUDA (Compute Unified Device Architecture) implementation for GPU, as well as on OpenACC (Open ACCelerators), a parallel programming model, and a standard for parallel computing. Both phases of the MFD algorithm (i) iterative DEM preprocessing and (ii) iterative MFD algorithm, are parallelized and run over multi-core CPU and GPU. The evaluation of the proposed solutions is performed with respect to the execution time, energy consumption, and programming effort for algorithm parallelization for different sizes of input data. An experimental evaluation has shown not only the advantage of using OpenACC programming over CUDA programming in implementing the watershed analysis on a GPU in terms of performance, energy consumption, and programming effort, but also significant benefits in implementing it on the multi-core CPU.

A suite of parallel algorithms for efficient band selection from hyperspectral images

The analysis of hyperspectral images is usually very heavy from the computational point-of-view, due to their high dimensionality. In order to avoid this problem, band selection (BS) has been widely used to reduce the dimensionality before the analysis. The aim is to extract a subset of the original bands of the hyperspectral image, preserving most of the information contained in the original data. The BS technique can be performed by prioritizing the bands on the basis of a score, assigned by specific criteria; in this case, BS turns out in the so-called band prioritization (BP). This paper focuses on BP algorithms based on the following parameters: signal-to-noise ratio, kurtosis, entropy, information divergence, variance and linearly constrained minimum variance. In particular, an optimized C serial version has been developed for each algorithm from which two parallel versions have been derived using OpenMP and NVIDIA’s compute unified device architecture. The former is designed for a multi-core CPU, while the latter is designed for a many-core graphics processing unit. For each version of these algorithms, several tests have been performed on a large database containing both synthetic and real hyperspectral images. In this way, scientists can integrate the proposed suite of efficient BP algorithms into existing frameworks, choosing the most suitable technique for their specific applications.

Study of infrared reflection characteristics of aerial target using MODIS data on GPU

To study of the infrared signature of an aerial target, it is required to precisely model the background radiation. Simple empirical models or standard atmospheric models in LOWTRAN/MODTRAN were used in earlier studies. To further precisely model the thermal radiation of earth’s surface and atmospheric radiance/transmittance, the atmospheric profile, land surface temperature, and emissivity, the sea surface temperature retrieved from moderate-resolution imaging spectroradiometer, and the sea surface emissivity model developed by Wu and Smith are utilized in this study. Meanwhile, considering that the reflection of background radiation incident from different directions in each spectral wavelength can be calculated in parallel, implementations using open multi-processing and compute unified device architecture on a multi-core CPU and many-core graphics processing unit (GPU) are presented and speedups of 9 $$\times$$ and 258 $$\times$$ are obtained on a platform with dual Xeon E5-2652 CPU and an NVIDIA Tesla K80 GPU card, respectively.

Parallel Motion Planning Using Poisson-Disk Sampling

We present a rapidly exploring-random-tree-based parallel motion planning algorithm that uses the maximal Poisson-disk sampling scheme. Our approach exploits the free-disk property of the maximal Poisson-disk samples to generate nodes and perform tree expansion. Furthermore, we use an adaptive scheme to generate more samples in challenging regions of the configuration space. The Poisson-disk sampling results in improved parallel performance and we highlight the performance benefits on multicore central processing units as well as manycore graphics processing units on different benchmarks.

A user mode CPU–GPU scheduling framework for hybrid workloads

Cloud platforms composed of multi-core CPU and many-core Graphics Processing Unit (GPU) have become powerful platforms to host incremental CPU–GPU workloads. In this paper, we study the problem of optimizing the CPU resource management while keeping the quality of service (QoS) of games. To this end, we propose vHybrid, a lightweight user mode runtime framework, in which we integrate a scheduling algorithm for GPU and two algorithms for CPU to efficiently utilize CPU resources with the control accuracy of QoS. vHybrid can maintain the desired QoS with low CPU utilization, while being able to guarantee better QoS performance with little overhead. Our evaluations show that vHybrid saves 37.29% of CPU utilization with satisfactory QoS for hybrid workloads, and reduces three orders of magnitude for QoS fluctuations, without any impact on GPU workloads.

Data processing performance analysis for ultrafast electron beam X-ray CT using parallel processing hardware architectures

The ultrafast electron beam X-ray computed tomography (CT) measuring system of the Helmholtz-Zentrum Dresden-Rossendorf (HZDR) is primarily operated for fundamental multiphase flow investigations, e.g. in various technical devices, and for validation of enhanced flow simulation models, e.g. developed for computational fluid dynamic codes (CFD). The CT scanner delivers cross-sectional material distributions by contactless measurements with a spatial resolution of approximately 1mm and a temporal resolution of maximal 8kHz. Currently, two central time-consuming processes have been identified limiting the efficient usage of that worldwide unique CT technique: a) the data transfer from the detector system to central data storages (e.g. computer or data base) and b) the data processing. Thus, data pre-processing and data reconstruction algorithms have been adapted for the use at multi-core central processing units (CPUs) and even many-core graphics processing units (GPUs). For optimal data processing results an advanced performance PC with two parallel operated high performance graphics processing units, a six-core processor, a high internal data bus speed and a large memory block has been assembled. The newly developed data processing algorithms induce a performance improvement of approximately 137 for the entire data processing sequence compared to the previous universally applicable single core CPU based data processing tool.

GPU-accelerated large-scale simulations of interfacial multiphase fluids for real-case applications

Abstract This paper presents a newly developed numerical model to simulate interfacial multi-phase flows in sanitary wares with high grid resolutions on Graphics Processing Unit (GPU)-equipped supercomputer. The numerical methods, including both interface capturing scheme and multi-phase fluid solver, are devised with high accuracy over compact computational stencil so as to be computationally efficient and well-suitable for hardwares of multi-core CPUs with many-core GPU accelerators. The code has been entirely ported to such a system using CUDA-driver API and MPI libraries to make fully possible use of the GPU acceleration capability on large-scale parallel computations. The high-resolution simulations with large number of mesh elements on the “TSUBAME” supercomputer, a GPU-rich system in Tokyo Institute of Technology, were conducted to simulate, for the first time, a whole sanitary ware with a mesh resolution of 1/10-mm level. Numerical results show that the present numerical model provides an innovative and practical CAE tool for the design process of water-saving sanitary ware products.

High-Efficiency Computing Technology for Thermohydrodynamic Lubrication Analysis

ABSTRACTNumerical analysis of thermal hydrodynamic (THD) lubrication is complex and time consuming because of the three-dimensional numerical domain of the solution of the energy equation. To reduce the execution time, many methods for determining solutions for bearing design problems are available; parallel computing technology is crucial for achieving high computational capability and involves using workstations equipped with multicore central processing units and many-core graphics processing units (GPUs). High-performance GPUs have emerged as powerful computing tools for modeling purposes in engineering and science. High-level GPUs depend on thousands of shaders, namely, processor cores. In this study, the traditional successive overrelaxation (SOR) method for solving the Reynolds equation and energy equation was replaced with the two- and three-dimensional red–black SOR methods for parallelizing the governing equations of a GPU system. The results show that a high-level GPU can be used to increase the parallel computing speed in the process of solving THD problems.

Accelerated Molecular Mechanical and Solvation Energetics on Multicore CPUs and Manycore GPUs.

Despite several reported acceleration successes of programmable GPUs (Graphics Processing Units) for molecular modeling and simulation tools, the general focus has been on fast computation with small molecules. This was primarily due to the limited memory size on the GPU. Moreover simultaneous use of CPU and GPU cores for a single kernel execution - a necessity for achieving high parallelism - has also not been fully considered. We present fast computation methods for molecular mechanical (Lennard-Jones and Coulombic) and generalized Born solvation energetics which run on commodity multicore CPUs and manycore GPUs. The key idea is to trade off accuracy of pairwise, long-range atomistic energetics for higher speed of execution. A simple yet efficient CUDA kernel for GPU acceleration is presented which ensures high arithmetic intensity and memory efficiency. Our CUDA kernel uses a cache-friendly, recursive and linear-space octree data structure to handle very large molecular structures with up to several million atoms. Based on this CUDA kernel, we present a hybrid method which simultaneously exploits both CPU and GPU cores to provide the best performance based on selected parameters of the approximation scheme. Our CUDA kernels achieve more than two orders of magnitude speedup over serial computation for many of the molecular energetics terms. The hybrid method is shown to be able to achieve the best performance for all values of the approximation parameter. The source code and binaries are freely available as PMEOPA (Parallel Molecular Energetic using Octree Pairwise Approximation) and downloadable from http://cvcweb.ices.utexas.edu/software.

A GPU based compressible multiphase hydrocode for modelling violent hydrodynamic impact problems

This paper presents a GPU based compressible multiphase hydrocode for modelling violent hydrodynamic impacts under harsh conditions such as slamming and underwater explosion. An effort is made to extend a one-dimensional five-equation reduced model (Kapila et al., 2001) to compute three-dimensional hydrodynamic impact problems on modern graphics hardware. In order to deal with free-surface problems such as water waves, gravitational terms, which are initially absent from the original model, are now considered and included in the governing equations. A third-order finite volume based MUSCL scheme is applied to discretise the integral form of the governing equations. The numerical flux across a mesh cell face is estimated by means of the HLLC approximate Riemann solver. The serial CPU program is firstly parallelised on multi-core CPUs with the OpenMP programming model and then further accelerated on many-core graphics processing units (GPUs) using the CUDA C programming language. To balance memory usage, computing efficiency and accuracy on multi- and many-core processors, a mixture of single and double precision floating-point operations is implemented. The most important data like conservative flow variables are handled with double-precision dynamic arrays, whilst all the other variables/arrays like fluxes, residual and source terms are treated in single precision. Several benchmark test cases including water-air shock tubes, one-dimensional liquid cavitation tube, dam break, 2D cylindrical underwater explosion near a planar rigid wall, 3D spherical explosion in a rigid cylindrical container and water entry of a 3D rigid flat plate have been calculated using the present approach. The obtained results agree well with experiments, exact solutions and other independent numerical computations. This demonstrates the capability of the present approach to deal with not only violent free-surface impact problems but also hull cavitation associated with underwater explosions. Performance analysis reveals that the running time cost of numerical simulations is dramatically reduced by use of GPUs with much less consumption of electrical energy than on the CPU.

Yet another Hybrid Strategy for Auto-tuning SpMV on GPUs

Sparse matrix-vector multiplication (SpMV) is a key linear algebra algorithm and is widely used in many application domains. Besides multi-core architecture, there is also extensive research focusing on accelerating SpMV on many-core Graphics Processing Units (GPUs). SpMV computations have many indirect and irregular memory accesses, and load imbalance could occur while mapping computations onto single-instruction, multiple-data (SIMD) GPUs. SpMV is highly memory bandwidth-bound, though GPUs have massive computational resources, the performance of SpMV on GPUs is still unsatisfying. In this paper, we present a new hybrid strategy for auto-tuning SpMV on GPUs. Our strategy combines the advantages of row-major storage and column-major storage. Like many other strategies, we reordered a given sparse matrix according to row lengths in decreasing order. In order to be more adaptive and efficient, we proposed a new hybrid Blocked CSR and JDS (BCJ) format based on original CSR and JDS. BCJ splits a sparse matrix into a denser part and a sparser part after reordering and uses different kernels to process the corresponding part. And we proposed corresponding auto-tuning framework to help transforming matrix and launching kernels according to the sparsity characteristics of the matrix. A CUDA implementation of BCJ outperforms the original formats significantly on a broad range of unstructured sparse matrices.

A parallel generalized relaxation method for high-performance image segmentation on GPUs

Fast and scalable software modules for image segmentation are needed for modern high-throughput screening platforms in Computational Biology. Indeed, accurate segmentation is one of the main steps to be applied in a basic software pipeline aimed to extract accurate measurements from a large amount of images. Image segmentation is often formulated through a variational principle, where the solution is the minimum of a suitable functional, as in the case of the Ambrosio–Tortorelli model. Euler–Lagrange equations associated with the above model are a system of two coupled elliptic partial differential equations whose finite-difference discretization can be efficiently solved by a generalized relaxation method, such as Jacobi or Gauss–Seidel, corresponding to a first-order alternating minimization scheme. In this work we present a parallel software module for image segmentation based on the Parallel Sparse Basic Linear Algebra Subprograms (PSBLAS), a general-purpose library for parallel sparse matrix computations, using its Graphics Processing Unit (GPU) extensions that allow us to exploit in a simple and transparent way the performance capabilities of both multi-core CPUs and of many-core GPUs. We discuss performance results in terms of execution times and speed-up of the segmentation module running on GPU as well as on multi-core CPUs, in the analysis of 2D gray-scale images of mouse embryonic stem cells colonies coming from biological experiments.

Bayesian State-Space Modelling on High-Performance Hardware UsingLibBi

LibBi is a software package for state space modelling and Bayesian inference on modern computer hardware, including multi-core central processing units, many-core graphics processing units, and distributed-memory clusters of such devices. The software parses a domain-specific language for model specification, then optimizes, generates, compiles and runs code for the given model, inference method and hardware platform. In presenting the software, this work serves as an introduction to state space models and the specialized methods developed for Bayesian inference with them. The focus is on sequential Monte Carlo (SMC) methods such as the particle filter for state estimation, and the particle Markov chain Monte Carlo and SMC2 methods for parameter estimation. All are well-suited to current computer hardware. Two examples are given and developed throughout, one a linear three-element windkessel model of the human arterial system, the other a nonlinear Lorenz '96 model. These are specified in the prescribed modelling language, and LibBi demonstrated by performing inference with them. Empirical results are presented, including a performance comparison of the software with different hardware configurations.

Methods to Load Balance a GCR Pressure Solver Using a Stencil Framework on Multi- and Many-Core Architectures

The recent advent of novel multi- and many-core architectures forces application programmers to deal with hardware-specific implementation details and to be familiar with software optimisation techniques to benefit from new high-performance computing machines. Extra care must be taken for communication-intensive algorithms, which may be a bottleneck for forthcoming era of exascale computing. This paper aims to present a high-level stencil framework implemented for the EULerian or LAGrangian model (EULAG) that efficiently utilises multi- and many-cores architectures. Only an efficient usage of both many-core processors (CPUs) and graphics processing units (GPUs) with the flexible data decomposition method can lead to the maximum performance that scales the communication-intensive Generalized Conjugate Residual (GCR) elliptic solver with preconditioner.