Proposed Parallel Approach Research Articles

HGRBOL2: Human gait recognition for biometric application using Bayesian optimization and extreme learning machine

The goal of gait recognition is to identify a person from a distance based on their walking style using a visual camera. However, the covariates such as a walk with carrying a bag and a change in clothes impact the recognition accuracy. This paper proposed a framework for human gait recognition based on deep learning and Bayesian optimization. The proposed framework includes both sequential and parallel steps. In the first step, optical flow-based motion regions are extracted and utilized to train the fine-tuned EfficentNet-B0 deep model. Simultaneously, we proposed a video frame enhancement technique. The same model is also separately trained on enhanced video frames. Instead of selecting static hyperparameters during the training process, Bayesian Optimization is employed in this work. Following that, two models have obtained: the motion frames model and the original frames model. Features are extracted from both models and fused using a proposed parallel approach named-Sq-Parallel Fusion (SqPF). The Tiger optimization algorithm was then improved and named Entropy controlled Tiger optimization (EVcTO). The best features are finally classified using an extreme learning machine (ELM) classifier. The experiment was carried out on two publicly available datasets, CASIA B and CASIA C, and yielded average accuracy of 92.04 and 94.97%, respectively. The proposed framework outperforms other deep learning-based networks in terms of accuracy.

Massively parallelized interpolated factored Green function method

This paper presents the first parallel implementation of the novel “Interpolated Factored Green Function” (IFGF) method introduced recently for the accelerated evaluation of discrete integral operators arising in wave scattering and other areas (Bauinger and Bruno, Jour. Computat. Phys., 2021). On the basis of the hierarchical IFGF interpolation strategy, the proposed (hybrid MPI-OpenMP) parallel implementation results in efficient data communication, and it scales up to large numbers of cores—without any hard limitations on the number of cores efficiently employed. Moreover, on any given number of cores, the proposed parallel approach preserves the O(Nlog⁡N) computing cost inherent in the sequential version of the IFGF algorithm. Unlike other approaches, the IFGF method does not utilize the Fast Fourier Transform (FFT), and it is thus better suited for efficient parallelization in distributed-memory computer systems. In particular, the IFGF method relies on a “peer-to-peer” strategy wherein, at every level, field propagation is directly enacted via “exchanges” between “peer” polynomials of constant degree, without data accumulation in large-scale “telephone-central” mathematical constructs such as those used in the Fast Multipole Method (FMM) and pure FFT-based approaches. A variety of numerical results presented in this paper illustrate the character of the proposed parallel algorithm, in particular demonstrating scaling from 1 to all 1,680 cores available in the High Performance Computing cluster used, and for problems of up to 4,096 wavelengths in acoustic size.

A parallel rule-based approach to compute rough approximations of dominance based rough set theory

In many datasets, conditional attributes and decision classes are preference-ordered, however, the classical Rough Set Theory (RST) does not consider the preference order between the values of the attributes. An extension of RST known as a Dominance-based Rough Set Approach (DRSA) provides dominance relation in this regard. In DRSA, data analysis mainly depends on the calculations of lower and upper approximations and these two measures are computationally utilizing many resources i.e., time and memory, due to the consideration of preference order. In this paper, we have proposed a parallel technique for calculating DRSA approximation sets. The proposed method directly computes approximations by following heuristic rules without calculating dominance positive or negative relations. The proposed parallel approach is then compared with the conventional method of calculation of DRSA approximations and a recent another technique of parallel processing using ten UCI publicly available datasets. Results validated the efficiency and effectiveness of the proposed model. An average reduction of 83% was observed in execution time and 86% in memory consumption. The structural complexity of the algorithm also considerably reduced.

Optimizing 2-opt-based heuristics on GPU for solving the single-row facility layout problem

The optimization of most combinatorial problems is NP-hard, which can be solved by heuristic algorithms to obtain approximately optimal solutions, especially for large-scale problems. Many heuristic algorithms can apply the 2-opt local search to find better solutions. In complete 2-opt local search, every valid neighboring solutions of swapping mechanism will be compared, which is very time consuming especially when a sequential algorithm is adopted. Nowadays, graphic processing units (GPUs) have evolved into powerful and flexible parallel computation platforms that have been widely used to accelerate the solving of NP-hard problems. Therefore, in this study, we focus on how to use GPUs to solve the single-row facility layout problem (SRFLP) with the 2-opt-based simulated annealing (SA) heuristic algorithm. As far as we know, this study is the first to solve SRFLP by using a GPU. After analyzing the fitness function and the move gains calculation between parent and child solutions, we propose a prefix-sum formula table to eliminate a large amount of replicated computation. To calculate move gains for the 2-opt local search during each iteration, many GPU threads are created and they construct and lookup the table in parallel. According to experimental results, if the prefix-sum formula table is not adopted, the GPU version outperforms the sequential CPU counterpart with the best speedup of 123. However, if the table is used in the GPU version, the best speedup can reach up to 3208. Because the proposed parallelization approach with the prefix-sum formula table is based on the features of the 2-opt local search and the SRFLP fitness function, it can be applied to any 2-opt-based heuristic algorithm for solving SRFLP even though the SA algorithm is used in this study.

A parallel metaheuristic approach for ensemble feature selection based on multi-core architectures

Ensemble learning have emerged as a useful machine learning technique, which is based on the idea that combining the output of multiple models instead of using a single model. This practice, known as “diversity”, and it usually enhances the performance. On other hand, ensemble feature selection method is based on the same idea, where multiple feature subsets are combined to select an optimal subset of features. Learning methods have difficulties with the dimensionality curse that impact the performance and increase the time exponentially. To overcome this issue, we propose a parallel heterogeneous ensemble feature selection based on three well-regarded algorithms: genetic algorithm, particle swarm optimizer, and grey wolf optimizer. The proposed approach is based on four phases; namely, distribution phase, parallel ensemble feature selection phase, combining and aggregation phase, and testing phase. Three implementations of the proposed approach are presented: a sequential approach running on the central processing unit (CPU), a parallel approach running on multi-core CPU, and a parallel approach running on multi-core CPU with graphics processing units (GPU). To assess the performance of the proposed approach twenty-one large datasets were used. The results show that the proposed parallel approach improved the performance in terms of the prediction results and running time.

Fast parallel vessel segmentation

Background and Objective: Accurate and fast vessel segmentation from liver slices remain challenging and important tasks for clinicians. The algorithms from the literature are slow and less accurate. We propose fast parallel gradient based seeded region growing for vessel segmentation. Seeded region growing is tedious when the inter connectivity between the elements is unavoidable. Parallelizing region growing algorithms are essential towards achieving real time performance for the overall process of accurate vessel segmentation.Methods: The parallel implementation of seeded region growing for vessel segmentation is iterative and hence time consuming process. Seeded region growing is implemented as kernel termination and relaunch on GPU due to its iterative mechanism. The iterative or recursive process in region growing is time consuming due to intermediate memory transfers between CPU and GPU. We propose persistent and grid-stride loop based parallel approach for region growing on GPU. We analyze static region of interest of tiles on GPU for the acceleration of seeded region growing.Results: We aim fast parallel gradient based seeded region growing for vessel segmentation from CT liver slices. The proposed parallel approach is 1.9x faster compared to the state-of-the-art.Conclusion: We discuss gradient based seeded region growing and its parallel implementation on GPU. The proposed parallel seeded region growing is fast compared to kernel termination and relaunch and accurate in comparison to Chan-Vese and Snake model for vessel segmentation.

GPU acceleration of a model-based iterative method for Digital Breast Tomosynthesis

Digital Breast Tomosynthesis (DBT) is a modern 3D Computed Tomography X-ray technique for the early detection of breast tumors, which is receiving growing interest in the medical and scientific community. Since DBT performs incomplete sampling of data, the image reconstruction approaches based on iterative methods are preferable to the classical analytic techniques, such as the Filtered Back Projection algorithm, providing fewer artifacts. In this work, we consider a Model-Based Iterative Reconstruction (MBIR) method well suited to describe the DBT data acquisition process and to include prior information on the reconstructed image. We propose a gradient-based solver named Scaled Gradient Projection (SGP) for the solution of the constrained optimization problem arising in the considered MBIR method. Even if the SGP algorithm exhibits fast convergence, the time required on a serial computer for the reconstruction of a real DBT data set is too long for the clinical needs. In this paper we propose a parallel SGP version designed to perform the most expensive computations of each iteration on Graphics Processing Unit (GPU). We apply the proposed parallel approach on three different GPU boards, with computational performance comparable with that of the boards usually installed in commercial DBT systems. The numerical results show that the proposed GPU-based MBIR method provides accurate reconstructions in a time suitable for clinical trials.

GPU acceleration of liver enhancement for tumor segmentation

Background and objective: Medical image segmentation plays a vital role in medical image analysis. There are many algorithms developed for medical image segmentation which are based on edge or region characteristics. These are dependent on the quality of the image. The contrast of a CT or MRI image plays an important role in identifying region of interest i.e. lesion(s). In order to enhance the contrast of image, clinicians generally use manual histogram adjustment technique which is based on 1D histogram specification. This is time consuming and results in poor distribution of pixels over the image. Cross modality based contrast enhancement is 2D histogram specification technique. This is robust and provides a more uniform distribution of pixels over CT image by exploiting the inner structure information from MRI image. This helps in increasing the sensitivity and accuracy of lesion segmentation from enhanced CT image. The sequential implementation of cross modality based contrast enhancement is slow. Hence we propose GPU acceleration of cross modality based contrast enhancement for tumor segmentation.Methods: The aim of this study is fast parallel cross modality based contrast enhancement for CT liver images. This includes pairwise 2D histogram, histogram equalization and histogram matching. The sequential implementation of the cross modality based contrast enhancement is computationally expensive and hence time consuming. We propose persistence and grid-stride loop based fast parallel contrast enhancement for CT liver images. We use enhanced CT liver image for the lesion or tumor segmentation. We implement the fast parallel gradient based dynamic seeded region growing for lesion segmentation.Results: The proposed parallel approach is 104.416 ( ± 5.166) times faster compared to the sequential implementation and increases the sensitivity and specificity of tumor segmentation.Conclusion: The cross modality approach is inspired by 2D histogram specification which incorporates spatial information existing in both guidance and input images for remapping the input image intensity values. The cross modality based liver contrast enhancement improves the quality of tumor segmentation.

Efficient GPU-based parallelization of solvation calculation for the blind docking problem

Molecular docking techniques are widely used in computational drug discovery. Most of these techniques simulate the way that a ligand interacts with a protein target focusing on one binding site. Blind docking is a recent technique which is designed to search the entire surface of the protein to discover new interesting binding sites. Unfortunately, this new docking method is computationally more intensive since its complexity grows exponentially according to the number of binding sites, which severely limits its utilization in practice. This paper shows a road-map for an efficient parallelization of the calculation of the solvation energy which represents the most time-consuming part of the scoring function. The latter constitutes a bottleneck in both simple and blind docking. The proposed parallelization approach aims to efficiently exploit the large computing power offered by the latest NVIDIA GPU architectures. Toward this goal, we propose a new parallel approach that exploits the Hyper-Q capability to compute several GPU kernels simultaneously, thereby speedingup the computation process and maximizing the GPU utilization. The obtained results show the huge benefit of exploiting the latest GPU architectures as compared to serial and parallel CPU approaches. Indeed, for 100 binding sites, our results show an average speedup of 186$$\times$$ as compared to the serial implementation, and 10$$\times$$ as compared to the multi-core CPU version. Moreover, an experimental comparison shows the superiority of our approach over the state of the art.

An efficient parallel simulation of unsteady blood flows in patient-specific pulmonary artery.

Simulation of blood flows in the pulmonary artery provides some insight into certain diseases by examining the relationship between some continuum metrics, eg, the wall shear stress acting on the vascular endothelium, which responds to flow-induced mechanical forces by releasing vasodilators/constrictors. V. Kheyfets, in his previous work, studies numerically a patient-specific pulmonary circulation to show that decreasing wall shear stress is correlated with increasing pulmonary vascular impedance. In this paper, we develop a scalable parallel algorithm based on domain decomposition methods to investigate an unsteady model with patient-specific pulsatile waveforms as the inlet boundary condition. The unsteady model offers tremendously more information about the dynamic behavior of the flow field, but computationally speaking, the simulation is a lot more expensive since a problem which is similar to the steady-state problem has to be solved many times, and therefore, the traditional sequential approach is not suitable anymore. We show computationally that simulations using the proposed parallel approach with up to 10 000 processor cores can be obtained with much reduced compute time. This makes the technology potentially usable for the routine study of the dynamic behavior of blood flows in the pulmonary artery, in particular, the changes of the blood flows and the wall shear stress in the spatial and temporal dimensions.

A Fully Parallel Stochastic Multiarea Power System Operation Considering Large-Scale Wind Power Integration

Multiarea power system operation/coordination is used to increase the reliability of the interconnected power grids and maintain the consistency of the price across the integrated power systems. However, the implementation of this coordinated operation is facing challenges due to increase in size and complexity of the modern power systems, high penetration of the volatile renewable energy, and interdependence issues among various power systems. To address these concerns, this paper presents a fully parallel decision-making approach for the day-ahead scheduling of interconnected power systems with large-scale wind power integration while respecting the information privacy between different systems/areas. In the proposed parallel approach, each system/area solves its day-ahead scheduling problem along with its local subproblems for different wind generations’ scenarios, and sends its equivalent (or processed) boundary information to other systems/areas. The proposed inter-regional coordination among systems/areas and intra-regional coordination between scenarios in each system/area will continue until the tie-line power flows and the generation outputs of generating units get converged. The modified IEEE 118-bus testing system is used in this paper to show the effectiveness of the proposed approach.

A GPU-Accelerated Deformable Image Registration Algorithm With Applications to Right Ventricular Segmentation

Delineation of the cardiac right ventricle is essential in generating clinical measurements such as ejection fraction and stroke volume. Given manual segmentation on the first frame, one approach to segment right ventricle from all of the magnetic resonance images is to find point correspondence between the sequence of images. Finding the point correspondence with non-rigid transformation requires a deformable image registration algorithm, which often involves computationally expensive optimization. The central processing unit (CPU)-based implementation of point correspondence algorithm has been shown to be accurate in delineating organs from a sequence of images in recent studies. The purpose of this study is to develop computationally efficient approaches for deformable image registration. We propose a graphics processing unit (GPU) accelerated approach to improve the efficiency. The proposed approach consists of two parallelization components: Parallel compute unified device architecture (CUDA) version of the deformable registration algorithm; and the application of an image concatenation approach to further parallelize the algorithm. Three versions of the algorithm were implemented: 1) CPU; 2) GPU with only intra-image parallelization (sequential image registration); and 3) GPU with inter and intra-image parallelization (concatenated image registration). The proposed methods were evaluated over a data set of 16 subjects. CPU, GPU sequential image, and GPU concatenated image methods took an average of 113.13, 16.50, and 5.96 s to segment a sequence of 20 images, respectively. The proposed parallelization approach offered a computational performance improvement of around $19\times $ in comparison to the CPU implementation while retaining the same level of segmentation accuracy. This paper demonstrated that the GPU computing could be utilized for improving the computational performance of a non-rigid image registration algorithm without compromising the accuracy.

A Parallel Approach for the Generation of Unstructured Meshes with Billions of Elements on Distributed-Memory Supercomputers

This paper describes a parallel approach for the rapid generation of ultra-large-scale unstructured meshes on distributed-memory supercomputers. A medium-sized initial mesh is prepared first. Afterwards, a two-level domain decomposition (DD) strategy is used to split and distribute the initial mesh to different cores. Finally, the parallel mesh generation, comprising a recursive procedure which includes parallel surface recovery, parallel boundary updating, and parallel mesh multiplication, is performed. The two-level DD differentiates the intra-node and inter-node communication to reduce communication overheads. A global indexing and updating scheme is used to make the mesh multiplication devoid of communication. A new parallel surface recovery algorithm without communication is developed to maintain the fidelity of the resulting mesh model to the original geometric model. Tests of the parallel approach for some real-life problems on supercomputers (Dawning-5000A and Tianhe-2) are presented. Issues regarding the speedup, parallel efficiency, and mesh quality are discussed. Results show that the proposed parallel approach has a reasonably good scalability, that the quality of the resulting mesh is improved, and that ultra-large-scale meshes with billions of elements can be generated quickly.

An O(N) and parallel approach to integral problems by a kernel-independent fast multipole method: Application to polarization and magnetization of interacting particles

Large classes of materials systems in physics and engineering are governed by magnetic and electrostatic interactions. Continuum or mesoscale descriptions of such systems can be cast in terms of integral equations, whose direct computational evaluation requires O(N2) operations, where N is the number of unknowns. Such a scaling, which arises from the many-body nature of the relevant Green’s function, has precluded wide-spread adoption of integral methods for solution of large-scale scientific and engineering problems. In this work, a parallel computational approach is presented that relies on using scalable open source libraries and utilizes a kernel-independent Fast Multipole Method (FMM) to evaluate the integrals in O(N) operations, with O(N) memory cost, thereby substantially improving the scalability and efficiency of computational integral methods. We demonstrate the accuracy, efficiency, and scalability of our approach in the context of two examples. In the first, we solve a boundary value problem for a ferroelectric/ferromagnetic volume in free space. In the second, we solve an electrostatic problem involving polarizable dielectric bodies in an unbounded dielectric medium. The results from these test cases show that our proposed parallel approach, which is built on a kernel-independent FMM, can enable highly efficient and accurate simulations and allow for considerable flexibility in a broad range of applications.

MrBayes sMC3

MrBayes is a popular software package for Bayesian phylogenetic inference, which uses an iterative approach to derive an evolutionary tree for a collection of species whose DNA sequences are known. Computationally, MrBayes is characterized by a large number of iterations, each composed of a set of tasks that isolated are not very time-consuming, but are globally computationally demanding. To accelerate the latest MrBayes 3.2, this paper presents MrBayes sMC3, which relies on the computational power of an heterogeneous CPU+GPU platform. For this, MrBayes sMC3 exploits both task and data-level parallelism while minimizing the overheads associated with kernel launches and CPU-GPU data transfers. Experimental results indicate that the proposed parallel approach, together with the proposed set of optimizations, allow for an application acceleration of up to 10× regarding the original MrBayes, and up to 3× regarding the Beagle Library. Furthermore, by analyzing the convergence rate of MrBayes sMC3 with that of the state-of-the-art approaches, a significant reduction in execution time is observed.

A Novel Immune Optimization with Shuffled Frog Leaping Algorithm - A Parallel Approach for Unsupervised Data Clustering

Data clustering is one of the data mining task, it is used to group the data objects according to their similarity. It is an optimization problem to find optimal results apply the proposed parallel approach called P-AISFLA. This hybrid algorithm is developed by utilizing the benefits of both social and immune mechanisms. The social algorithm Shuffled Frog Leaping Algorithm is a new parameter free population based algorithm combined with Clonal selection algorithm CSA. This hybrid algorithm performs the parallel computation of social behavior based SFLA and Immune behavior based CSA to improve the ability to reach the global optimal solution with a faster and a rapid convergence rate. The proposed algorithm PAISFL is applied for the data clustering applications and proved that it produces optimal results than SFLA and PSO.

A parallel adaptive segmentation method based on SOM and GPU with application to MRI image processing

In this paper we develop a set of parallel algorithms for image segmentation basing on the authors׳ former work, the self-organizing map (SOM) based vector quantization (VQ) approach, by extending the method from serial computation to parallel processing, in order to accelerate the computation process. The parallel segmentation scheme is composed of a group of parallel algorithms for implementing the whole segmentation process, including parallel classification of the image into edge and non-edge pattern vectors, parallel training of an SOM network, parallel segmentation of the image by using the trained SOM model with VQ method, and adaptive parallel estimation of the segment number of the image being processed. In the paper, all the parallel algorithms have been implemented on graphic processing units (GPU) and applied to segmenting the human brain MRI images. The experimental results obtained in the work show that, compared with the original serial method implemented on CPU, the proposed parallel approach can achieve a significant improvement on the computation efficiency with overall speedup ratios increasing from 28.81 to 89.12 as image sizes increasing from 128×128 to 1024×1024, while keeping the segmentation performance unchanged.

A parallel algorithm for random searches

We discuss a parallelization procedure for a two-dimensional random search of a single individual, a typical sequential process. To assure the same features of the sequential random search in the parallel version, we analyze the former spatial patterns of the encountered targets for different search strategies and densities of homogeneously distributed targets. We identify a lognormal tendency for the distribution of distances between consecutively detected targets. Then, by assigning the distinct mean and standard deviation of this distribution for each corresponding configuration in the parallel simulations (constituted by parallel random walkers), we are able to recover important statistical properties, e.g., the target detection efficiency, of the original problem. The proposed parallel approach presents a speedup of nearly one order of magnitude compared with the sequential implementation. This algorithm can be easily adapted to different instances, as searches in three dimensions. Its possible range of applicability covers problems in areas as diverse as automated computer searchers in high-capacity databases and animal foraging.

A fast and energy-efficient Hamming decoder for software-defined radio using graphics processing units

The demand for scalable and fast error decoders has recently increased in software-defined radio-based communication systems. Hamming code, which is one of the promising error decoders, shows acceptable accuracy; however, the computational complexity of the decoder limits its use in real-time communication. To address this issue, this paper proposes a fully parallel implementation of the (7, 4) Hamming code on a graphics processing unit (GPU) by exploiting massive data-parallelism and increasing on-chip constant memory accesses. To further improve the performance of this proposed parallel approach, this paper explores the impact of different thread/block configurations and selects optimal thread/block configurations, which can occupy more hardware resources for performing parity checks, error detection and correction, and decoding of the received codeword. In addition, the proposed GPU-based Hamming decoder can provide significant scalability by supporting different message sizes, including 355,907 bytes, 2,959,475 bytes, and 12,835,890 bytes. To verify the effectiveness of the GPU-based parallel Hamming decoder, this paper compares its performance with that of the multi-threading central processing unit (CPU) approach which is executed on an Intel multi-core processor. Experimental results indicate that the proposed GPU-based decoder operates at least 15.13 times faster and reduces the energy consumption by up to 913.17 % compared to the multi-threading CPU-based approach.

A Map-Reduce based parallel approach for improving query performance in a geospatial semantic web for disaster response

Rapid retrieval of spatial information is critical to ensure that emergency supplies and resources can reach the impacted areas in the most efficient manner. However, it remains challenging to find out the needed spatial information efficiently because of the intensive geocomputation processes involved and the heterogeneity of spatial data. It is quite cost prohibitive to query the spatial information from geographical knowledge bases containing complex topological relationships. This research introduces a Map-Reduce based parallel approach for improving the query performance of a geospatial ontology for disaster response. The approach focuses on parallelizing the spatial join computations of GeoSPARQL queries. The proposed parallel approach makes full use of data/task parallelism for spatial queries. The results of some initial experiments show that the proposed approach can reduce individual spatial query execution time by taking advantage of parallel processes. The proposed approach, therefore, may afford a large number of concurrent spatial queries in disaster response applications.