Search

This document specifies an interface to support the parallel features of Fortran, named the Parallel Runtime Interface for Fortran (PRIF). PRIF is a proposed solution in which the runtime library is responsible for coarray allocation, deallocation and accesses, image synchronization, atomic operations, events, and teams. In this interface, the compiler is responsible for transforming the invocation of Fortran-level parallel features into procedure calls to the necessary PRIF procedures. The interface is designed for portability across shared- and distributed-memory machines, different operating systems, and multiple architectures. Implementations of this interface are intended as an augmentation for the compiler's own runtime library. With an implementation-agnostic interface, alternative parallel runtime libraries may be developed that support the same interface. One benefit of this approach is the ability to vary the communication substrate. A central aim of this document is to define a parallel runtime interface in standard Fortran syntax, which enables us to leverage Fortran to succinctly express various properties of the procedure interfaces, including argument attributes.

High flexibility is an important feature of seru system that has received less attention. In this paper, we discuss how to do such flexible seru system formation, especially focusing on the strategic decision phase. We formulate the flexible seru system formation problem (FSFP) as a nonlinear programming model to evaluate flexibility performance in terms of flexibility–investment cost and flexibility–loss cost. To exactly obtain the optimal solution of the FSFP, we transform the nonlinear model into a linear one and solve it with Gurobi solver. For the large-scale problem, we proposed a parallel Master–Slave adaptive genetic algorithm (PMSA-GA) by transforming it into a two-stage stochastic programming model. The adaptive selection is used to improve the quality of solutions in PMSA-GA. To reduce the computational time, multiple populations of seru formation evolve in parallel with the assistance of the Master–Slave mechanism. Extensive experiments are tested to evaluate the performance of the proposed model and algorithm, and the effect of cost parameters on the system performance is discussed. The results show that the FSFP model takes the property of dynamic demand into account and is more suitable for dynamic demand environments than the task-oriented seru formation (TOSF) strategy from the previous literature. [ABSTRACT FROM AUTHOR]

The study of approximate matching in the Massively Parallel Computations (MPC) model has recently seen a burst of breakthroughs. Despite this progress, we still have a limited understanding of maximal matching which is one of the central problems of parallel and distributed computing. All known MPC algorithms for maximalmatching either take polylogarithmic time which is considered inefficient, or require a strictly superlinear space of n1+Ω(1) per machine. In this work, we close this gap by providing a novel analysis of an extremely simple algorithm, which is a variant of an algorithm conjectured to work by Czumaj, Lacki, Madry, Mitrovic, Onak, and Sankowski [15]. The algorithm edge-samples the graph, randomly partitions the vertices, and finds a random greedy maximal matching within each partition.We show that this algorithm drastically reduces the vertex degrees. This, among other results, leads to an O(log log Δ) round algorithm for maximal matching with O(n) space (or even mildly sublinear in n using standard techniques). As an immediate corollary, we get a 2 approximate minimum vertex cover in essentially the same rounds and space, which is the optimal approximation factor under standard assumptions. We also get an improved O(log log Δ) round algorithm for 1 + ε approximate matching. All these results can also be implemented in the congested clique model in the same number of rounds. [ABSTRACT FROM AUTHOR]

In this article, we address the practical needs for the rapid processing and application of aerial photogrammetry remote sensing images, proposing a method based on high-performance computing technology to achieve the real-time, dynamic orthorectified visualization of aerial remote sensing images. This method differs from traditional orthorectification processing and application workflows, utilizing high-performance parallel computing capabilities to directly achieve orthorectification based on raw aerial images. In this paper, we also compare and analyze four different real-time orthorectified visualization schemes to meet various application needs, evaluating the efficiency and accuracy of the proposed method. Finally, experimental results with unmanned aerial vehicle (UAV) flight image data indicate that this method is feasible, with real-time computational results meeting practical application requirements in terms of effect and performance. [ABSTRACT FROM AUTHOR]

Kaczmarz algorithm is an iterative projection method for solving system of linear equations that arise in science and engineering problems in various application domains. In addition to classical Kaczmarz, there are randomized and parallel variants. The main challenge of the parallel implementation is the dependency of each Kaczmarz iteration on its predecessor. Because of this dependency, frequent communication is required which results in a substantial overhead. In this study, a new distributed parallel method that reduces the communication overhead is proposed. The proposed method partitions the problem so that the Kaczmarz iterations on different blocks are less dependent. A frequency parameter is introduced to see the effect of communication frequency on the performance. The communication overhead is also decreased by allowing communication between processes only if they have shared non‐zero columns. The experiments are performed using problems from various domains to compare the effects of different partitioning methods on the communication overhead and performance. Finally, parallel speedups of the proposed method on larger problems are presented. [ABSTRACT FROM AUTHOR]

Design optimization of multilayer piezoelectric transducers is intended for efficient and practical usage of wideband transducers for fault diagnosis, biomedical, and underwater applications through adjusting layer thicknesses and volume fraction of piezoelectric material in each layer. In this context, we propose a parallel differential evolution (PDE) algorithm to mitigate the complexities of multivariate optimization as well as the computation time to achieve an optimized wideband transducer for the particular application. For lead magnesium niobate-lead titanate (PMN PT)- and PZT5h-based piezoelectric materials, the fitness function is formulated based on uniformity of mechanical pressure at the first three harmonics to achieve wide bandwidth in the required functional frequency range. It is carried out using a one-dimensional model (ODM), while input layer thicknesses and volume fractions of active material are evaluated using PDE. The simulation is performed on a parallel computing platform utilizing three different host machines to reduce computational time. Results of the proposed methodology for PDE are statistically represented in the form of minimum, maximum, mean, and standard deviation of fitness value, while graphically represented in terms of speedup and time. It can be observed that the execution time for parallel DE decreases with the increasing number of cores. [ABSTRACT FROM AUTHOR]

Large language models have sparked a lot of attention in the research community in recent days, especially with the introduction of practical tools such as ChatGPT and Github Copilot. Their ability to solve complex programming tasks was also shown in several studies and commercial solutions increasing the interest in using them for software development in different fields. High performance computing is one of such fields, where parallel programming techniques have been extensively used to utilize raw computing power available in contemporary multicore and manycore processors. In this paper, we perform an evaluation of the ChatGPT and Github Copilot tools for OpenMP-based code parallelization using a proposed methodology. We used nine different benchmark applications which represent typical parallel programming workloads and compared their OpenMP-based parallel solutions produced manually and using ChatGPT and Github Copilot in terms of obtained speedup, applied optimizations, and quality of the solution. ChatGPT 3.5 and Github Copilot installed with Visual Studio Code 1.88 were used. We concluded that both tools can produce correct parallel code in most cases. However, performance-wise, ChatGPT can match manually produced and optimized parallel code only in simpler cases, as it lacks a deeper understanding of the code and the context. The results are much better with Github Copilot, where much less effort is needed to obtain correct and performant parallel solution. [ABSTRACT FROM AUTHOR]

Nowadays, as the ever-increasing demand for more powerful computing resources continues, alternative advanced computing paradigms are under extensive investigation. Significant effort has been made to deviate from conventional Von Neumann architectures. In-memory computing has emerged in the field of electronics as a possible solution to the infamous bottleneck between memory and computing processors, which reduces the effective throughput of data. In photonics, novel schemes attempt to collocate the computing processor and memory in a single device. Photonics offers the flexibility of multiplexing streams of data not only spatially and in time, but also in frequency or, equivalently, in wavelength, which makes it highly suitable for parallel computing. Here, we numerically show the use of time and wavelength division multiplexing (WDM) to solve four independent tasks at the same time in a single photonic chip, serving as a proof of concept for our proposal. The system is a time-delay reservoir computing (TDRC) based on a microring resonator (MRR). The addressed tasks cover different applications: Time-series prediction, waveform signal classification, wireless channel equalization, and radar signal prediction. The system is also tested for simultaneous computing of up to 10 instances of the same task, exhibiting excellent performance. The footprint of the system is reduced by using time-division multiplexing of the nodes that act as the neurons of the studied neural network scheme. WDM is used for the parallelization of wavelength channels, each addressing a single task. By adjusting the input power and frequency of each optical channel, we can achieve levels of performance for each of the tasks that are comparable to those quoted in state-of-the-art reports focusing on single-task operation. We also quantify the memory capacity and nonlinearity of each parallelized RC and relate these properties to the performance of each task. Finally, we provide insight into the impact of the feedback mechanism on the performance of the system. [ABSTRACT FROM AUTHOR]

The behavior of turbulent flows remains a significant unsolved problem in physics. Recently, a large quantity of effort has been directed toward understanding the non-linear interactions of the different flow structures in order to address this challenge. In this paper, different implementations of one exact method for identifying these structures are analyzed. This includes two sequential algorithms and a parallelizable one, developed to handle large-scale data efficiently. The new parallel algorithm offers significant advantages in handling the computational demands of large simulations, providing a more scalable solution for future research. [ABSTRACT FROM AUTHOR]

This study presents an asynchronous parallel strategy coordinating central processing unit (CPU) and graphic processing unit (GPU) to accelerate neighborhood operation (NO). Specifically, we propose a data partitioning method called multi-anchor task queuing and a task scheduling method called bi-direction task scheduling, which can support CPU and GPU to find the responsible data blocks rapidly and concurrently handle their tasks via a bi-direction merge. Moreover, we optimize the organization of threads distributed among the CPU and GPU. Experimental results show that when a 1.7 GB raster dataset is processed, the speedup ratio achieved by the proposed parallel algorithm reaches 29.63, which is 19% and 18% higher than those of the GPU and standard asynchronous parallel algorithm, respectively. Additionally, the load balance index is below 0.085, which is significantly better than the value achieved by a conventional algorithm. Thus, the strategy achieves a higher speedup ratio and more adaptable load balance, thereby accelerating the NO more efficiently. Further, the impacts of the data volume, computational intensity, organization mode of the GPU threads, and granularity of the GPU stream on the parallel efficiency are evaluated and discussed. We also test the efficiency of four other common NOs with our strategy. [ABSTRACT FROM AUTHOR]

Rapid development realms of parallel architectures and its heterogeneity have inspired researchers to invent new scheduling strategies to efficiently distribute workloads among these architectures in a way that may lead to better performance. This paper presents a comprehensive study on optimizing resource utilization for large-scale problems by employing architecture-aware scheduling techniques. We conducted a series of experiments to measure the execution times of various architectures with different problem sizes. These experiments have been conducted multiple times to minimize measurement variance. The findings from these experiments are utilized to develop a scheduling strategy that enables faster completion of larger data-parallel problems while maximizing resource utilization. The proposed approach makes performance enhancement with 16.7% for large data size. It has a significant impact on enhancing computational efficiency and reducing costs in high-performance computing environments. [ABSTRACT FROM AUTHOR]

In this paper, a deep learning based time-domain inversion method is proposed to reconstruct high-contrast scatterers from the measured electromagnetic fields. The scatterers investigated in this study include four kinds of geometry shapes, which cover the arbitrary geometrical shapes, handwritings and lossy medium. After being well trained, the performance of the proposed method is evaluated from the perspective of accuracy, noise interference, and computational acceleration. It can be proven that the proposed framework can realize high-precision inversion in several milliseconds. Compared with typical reconstruction methods, it avoids the iterative calculation by utilizing the parallel computing ability of GPU and thus significantly reduce the computing time. Besides, the proposed method has shown the potential to be applied in practical scenarios with experimental results. Herein, it is confident that the proposed method has the potential to serve as a new path for real-time quantitative microwave imaging for various practical scenarios. In the end, the limitation of the method is also discussed. [ABSTRACT FROM AUTHOR]

Currently, lightweight small object detection algorithms for unmanned aerial vehicles (UAVs) often employ group convolutions, resulting in high Memory Access Cost (MAC) and rendering them unsuitable for edge devices that rely on parallel computing. To address this issue, we propose the SOD-YOLO model based on YOLOv7, which incorporates a DSDM-LFIM backbone network and includes a small object detection branch. The DSDM-LFIM backbone network, which combines Deep-Shallow Downsampling Modules (DSD Modules) and Lightweight Feature Integration Modules (LFI Modules), avoids excessive use of group convolutions and element-wise operations. The DSD Module focuses on extracting both deep and shallow features from feature maps using fewer parameters to obtain richer feature representations. The LFI Module, is a dual-branch feature integration module designed to consolidate feature information. Experimental results demonstrate that the SOD-YOLO model achieves an AP50 of 50.7% and a FPS of 72.5 on the VisDrone validation set. Compared to YOLOv7, our model reduces computational costs by 20.25% and decreases the number of parameters by 17.89%. After scaling the number of channels in the model, it achieves an AP50 of 33.4% with an inference time of 27.3ms on the Atlas 200I DK A2. These experimental results indicate that the SOD-YOLO model can effectively perform small object detection tasks in a large number of aerial images captured by UAVs. [ABSTRACT FROM AUTHOR]

Utilizing acoustic information for search route planning will greatly increase the success rate of searching for underwater targets, which requires rapid computing of numerous underwater acoustic fields. The efficiency of traditional computing methods is too low to meet the requirements of rapid applications. In this paper, a underwater multi-acoustic fields computing model is developed based on ray theory, and multi-level hybrid parallel computing strategies are designed based on the model characteristics, and a dynamic scheduling optimization algorithm at process level is introduced to solve the load imbalance problem. All parallel computing strategies are tested in Tianhe-II High Performance Computer (HPC) system, and the tests show that: 1. compared with the serial version, hybrid parallel computing strategy provides a Speedup of 75.7 under 240 cores; 2. after the introduction of the dynamic scheduling optimization algorithm, the speed of solving the underwater acoustic fields is further increased by 28.99% under the same computing resources, and the Speedup reaches 97.67; 3. the optimal combination of process/thread parameter on the Tianhe-II HPC system is given as 3/8, and the final Speedup reaches 112.13. [ABSTRACT FROM AUTHOR]

This paper introduces an open-source tool capable of performing the Composite System Reliability evaluation developed in the high-level, dynamic Julia™ programming language. Employing Monte Carlo Simulation and parallel computing, the tool evaluates probabilistic adequacy indices for combined generation and transmission systems, focusing on both individual delivery points and the broader system. Proficiency in Optimal Power Flow problem formulations is demonstrated through two distinct methods: DC and linearized AC, enabling comprehensive resource and transmission adequacy analysis with high-performance solvers. Addressing replicability and the insufficiency of available software, the tool supports diverse analyses on a unified platform. The paper discusses the tool's design and validation, particularly focusing on the two optimal power flow problem formulations. These insights significantly contribute to understanding transmission system performance and have implications for power system planning. [ABSTRACT FROM AUTHOR]

Cryogenic electron microscopy (Cryo-EM) is a widely used technique for visualizing the 3D structures of many drug design targets, including membrane proteins, at atomic resolution. However, the necessary throughput for structure-based drug design (SBDD) is not yet achieved. Currently, data analysis is a major bottleneck due to the rapid advancements in detector technology and image acquisition methods. Here we show "GoToCloud", a cloud-computing-based platform for advanced data analysis and data management in Cryo-EM. With GoToCloud, it is possible to optimize computing resources and reduce costs by selecting the most appropriate parallel processing settings for each processing step. Our benchmark tests on GoToCloud demonstrate that parallel computing settings, including the choice of computational hardware, as well as a required target resolution have significant impacts on the processing time and cost performance. Through this optimization of a cloud computing environment, GoToCloud emerges as a promising platform for the acceleration of Cryo-EM SBDD. The design philosophy, implementation, optimization, and performance evaluation of a cloud-computing platform for data analysis and management in Cryo-EM single particle analysis to achieve a significant reduction in processing time and cost. [ABSTRACT FROM AUTHOR]

The blisks are complex thin-walled parts with specific structures that have narrow channels and a large degree of bowed-twisted blades. These parts are typically machined using 5-axis machining. However, conventional feedrate scheduling of the tool tip can cause flutter and reduced machining accuracy when dealing with very small radii of curvature and dramatic changes in the tool axis vector. This is because there is a significant difference in moving speed between the tool tip and cutting contact points. To address this issue, we establish an optimization model for feedrate under constraints such as process-allowed cutting speed and machine tool drive along the toolpath. For sections of short toolpaths in the regions of leading and trailing edges with drastic changes in curvature and tool orientations, we schedule a constant feedrate for the tool contact point. For other sections of the toolpath, we construct time-optimal acceleration and deceleration velocity curves using parallel computing technology, ensuring no abrupt changes in acceleration or acceleration velocity at boundary points. Additionally, we smooth the feedrate profile curves for the entire toolpath using the parallel bidirectional scanning method. This approach improves both the efficiency of feedrate scheduling and cutting stability within the regions corresponding to the leading and trailing edges of blisks. [ABSTRACT FROM AUTHOR]

For energy and cost-efficient computing, many desktops and embedded computing devices these days use voltage frequency islands (VFIs) based processors. An island in VFIs consists of multiple homogeneous cores; however, multiple islands are generally heterogeneous in nature. In contrary to architectures (non-VFI) with per-core dynamic voltage frequency scaling (DVFS), in VFI, all the cores in an island run on the same voltage/frequency at the same time. The energy-aware scheduling of task graphs on VFI architectures is challenging and different from non-VFI architectures in which task-based DVFS is possible. However, in VFI scheduling, the time slot for which an island runs on a particular voltage will be decided. In this paper, we analyse 20 different scheduling heuristics for VFI architectures by varying the size of this time slot based on the workload properties and also by varying the voltage/frequency of the time slots. We also propose a heuristic OptSlotVFI to utilize all the slots optimally. The results show that the VFI scheduling heuristics are able to generate schedules which improve energy consumption up to 25% with an equivalent or shorter schedule length than the state-of-the-art approaches. [ABSTRACT FROM AUTHOR]

Optical network-on-chip (ONoC) is an emerging architecture to realize high-performance and low power consumption for many-core processors. Multicast in ONoC is a vital communication pattern widely used in parallel computing and computational genomics. This paper proposes an efficient scheme for multiple multicast requests in ONoC. Firstly, we design an algorithm to construct two trees for two multicasts without duplicate edges in a mesh topology, and it only uses one wavelength. We propose a graph to represent the relationship between each multicast. Secondly, a group-based wavelength assignment scheme is designed to assign wavelengths for multiple multicasts. Finally, we used data analysis to extract helpful simulation results to form conclusions. The results indicate that our algorithm is superior to other tree-based schemes in communication latency and wavelength number usage. [ABSTRACT FROM AUTHOR]

Lensless microscopy provides a wide field of view (FOV) determined by the image sensor size, allowing visualization of large sample areas. Coupled with advanced and even pixel super‐resolution phase retrieval algorithms, it can achieve resolutions up to the sub‐micron level, enabling both large‐FOV and high‐resolution imaging. However, high‐throughput lensless imaging encounters challenges in rapid data acquisition and large‐scale phase retrieval. Furthermore, when examining biological samples over a large FOV, focal plane inconsistencies often emerge among distinct regions. This study introduces a fast acquisition and efficient reconstruction method for coherent lensless imaging. Multiple measurements are manually modulated using an axial translation stage and sequentially captured by an image sensor, requiring no hardware synchronization. Optical parameter calibration, region‐wise auto‐focusing, and region‐wise phase retrieval algorithms are integrated to establish a general parallel computing framework for rapid, efficient, and high‐throughput lensless imaging. Experimental results demonstrate a 7.4 mm × 5.5 mm FOV and 1.38 µm half‐pitch resolution imaging of human skin and lung tumor sections with region‐wise focusing, requiring ≈0.5‐s acquisition time and 17‐s reconstruction time. By incorporating pixel super‐resolution, a 0.98 µm half‐pitch resolution is achieved in full‐FOV peripheral blood smears without additional data required, advantageous for discerning hollow shapes and segmenting blood cells. [ABSTRACT FROM AUTHOR]

Zero-knowledge succinct non-interactive proofs (zk-SNARK) have found extensive applications in various fields, including cryptocurrencies, due to their swift and efficient proof verification process. However, the computational intensity of the proof generation process poses a significant challenge, particularly at the number theoretic transform (NTT) stage. This paper proposed a GPU-based acceleration method for NTT computations, named PreNTT, to address this bottleneck. The method employed precomputation and optimization of twiddle factor powers to reduce the parallel computation overhead in NTT. It also introduced dynamic precomputation to enhance the efficiency of these computations. The algorithm made use of dynamic adaptive kernel scheduling, which allocated GPU resources on-chip according to the NTT input size, thereby boosting the computational efficiency for large-scale tasks. Additionally, the approach combined external global data shuffling with internal local data shuffling to avoid memory access conflicts. The use of CUDA multi-stream technology allowed for effective concealment of precomputation times during data transfer and computation processes. Experimental results indicate that the zk-SNARK system utilizing PreNTT achieves a speed-up ratio ranging from 1.7x to 9x in NTT module running times compared to Bellperson, the industryleading system. PreNTT effectively increases the parallelism of the NTT algorithm and reduces the computational overhead in zk-SNARK operations. [ABSTRACT FROM AUTHOR]

This article studies the problem of scheduling independent jobs on parallel machines with a common server, the objective of which is to minimize the makespan. In this case, the common server is responsible for performing the setup operations and, therefore, there must be no conflicts while conducting them. Four alternative time-indexed formulations for the problem are considered and evaluated computationally. Moreover, two algorithms are presented that can significantly improve the performance of the best time-indexed formulation. The results obtained on two benchmark datasets involving up to 100 jobs suggest that the proposed improved algorithms are substantially better than existing approaches. [ABSTRACT FROM AUTHOR]

The purpose of this work is to develop a mathematical apparatus and computational algorithms for implementation of parallel computing in geometric modeling and computer-aided design (CAD) systems. The analysis of existing approaches to parallel computing implementation in CAD systems is carried out. As a result, it is found that most information modeling and CAD systems do not support parallel computing at the level of the geometric kernel. A concept for the development of a CAD geometric kernel based on the invariants of parallel projection of geometric objects onto the axes of the global coordinate system is proposed. It combines the potential of constructive methods for geometric modeling, capable of parallelizing geometric constructions by tasks (message passing), and the mathematical apparatus of point calculus, capable of parallelization by data through coordinate-by-coordinate calculation (data parallel). The use of the coordinate-by-coordinate calculation for point equations not only makes it possible to parallelize computations along coordinate axes, but also ensures the consistency of computational operations with respect to threads, which significantly reduces the idle time and optimizes the CPU operation to achieve the maximum effect from the use of parallel computing. [ABSTRACT FROM AUTHOR]

The reliability of multiprocessor systems is now a crucial concern in parallel computing, which can be characterized as connectivity and diagnosability. The s-extra connectivity (s-EC) κ s (G) of a network G is the minimum number of nodes whose deletion disconnects the network G, and every remaining component has no less than s + 1 nodes. The s-extra diagnosability (s-ED) t s (G) of a network G is the maximum cardinality of faulty nodes that can be identified, given that each remaining component has at least s + 1 nodes. This paper investigates the s-EC and s-ED of the complete cubic network CCN(n). Specifically, we initially demonstrate that the s-EC of CCN(n) is κ s (C C N (n)) = (s + 1) (n + 1) - s (s + 3) 2 for n ≥ 3 and 0 ≤ s ≤ n - 2 . Subsequently, we demonstrate that the s-ED under the PMC model is t s (C C N (n)) = (s + 1) (n + 1) - s (s + 3) 2 + s for n ≥ 3 and 1 ≤ s ≤ n - 2 . Similarly, under the MM* model, the s-ED is t s (C C N (n)) = (s + 1) (n + 1) - s (s + 3) 2 + s for n ≥ 6 and 1 ≤ s ≤ n - 2 4 . Finally, we conduct simulation experiments, and the results indicate that the s-EC consistently surpasses other known connectivities, including classical connectivity and s-component connectivity. Additionally, the s-ED consistently outperforms classical diagnosability and s-component diagnosability of CCN(n). [ABSTRACT FROM AUTHOR]

Parallel computing is very important to accelerate the performance of computing applications. Moreover, parallel applications are expected to continue executing in more dynamic environments and react to changing conditions. In this context, applying self-adaptation is a potential solution to achieve a higher level of autonomic abstractions and runtime responsiveness. In our research, we aim to explore and assess the possible abstractions attainable through the transparent management of parallel executions by self-adaptation. Our primary objectives are to expand the adaptation space to better reflect real-world applications and assess the potential for self-adaptation to enhance efficiency. We provide the following scientific contributions: (I) A conceptual framework to improve the designing of self-adaptation; (II) A new decision-making strategy for applications with multiple parallel stages; (III) A comprehensive evaluation of the proposed decision-making strategy compared to the state-of-the-art. The results demonstrate that the proposed conceptual framework can help design and implement self-adaptive strategies that are more modular and reusable. The proposed decision-making strategy provides significant gains in accuracy compared to the state-of-the-art, increasing the parallel applications' performance and efficiency. [ABSTRACT FROM AUTHOR]

Object detection algorithms are widely used but involve significant amounts of computations, posing challenges for adaptation to resource-constrained application scenarios, such as Unmanned Aerial Vehicles (UAVs) and Unmanned Surface Vehicles (USVs). To address these challenges, this paper proposes an object detection accelerator architecture with multi-channel parallel computation. In this architecture, the computationally intensive modules, including the convolution layer, pooling layer, and upsampling layer, are accelerated using hardware, and other modules are dealt with CPU embedded in the FPGA. The methods of pipeline design, loop expansion, data reordering, and other technologies are fully utilized to design hardware acceleration modules. A data transmission architecture is designed, incorporating multi-channel transmission along with ping-pong buffering and employing a blocking strategy for off-chip data access. Furthermore, the architecture incorporates multiple acceleration IP cores to minimize data transmission delays. Based on this architecture, the tiny-YOLOv4 model is optimized and implemented on FPGA as a hardware accelerator for object detection. The network model is enhanced by integrating the convolutional layer with normalization, and different attention mechanisms are applied to improve feature extraction, thereby reducing computational load and enhancing accuracy. The performance of the tiny-YOLOv4 FPGA-based accelerator is validated using the SIMD dataset. Experimental results demonstrate that the hardware accelerator performs exceptionally, consuming only 2.4W and surpassing existing alternatives. Such adaptability facilitates its integration into complex environments such as intelligent transportation systems. [ABSTRACT FROM AUTHOR]

Significant progress has been made in industrial defect detection due to the powerful feature extraction capabilities of deep neural networks (DNNs). However, the high computational cost and memory requirement of DNNs pose a great challenge to the deployment of industrial edge-side devices. Although traditional binary neural networks (BNNs) have the advantages of small storage space requirements, high parallel computing capability, and low power consumption, the problem of significant accuracy degradation cannot be ignored. To tackle these challenges, this paper constructs a BNN with layered data fusion mechanism (LDF-BNN) based on BNext. By introducing the above mechanism, it strives to minimize the bandwidth pressure while reducing the loss of accuracy. Furthermore, we have designed an efficient hardware accelerator architecture based on this mechanism, enhancing the performance of high-accuracy BNN models with complex network structures. Additionally, the introduction of multi-storage parallelism alleviates the limitations imposed by the internal transfer rate, thus improving the overall computational efficiency. The experimental results show that our proposed LDF-BNN outperforms other methods in the comprehensive comparison, achieving a high accuracy of 72.23%, an image processing rate of 72.6 frames per second (FPS), and 1826 giga operations per second (GOPs) on the ImageNet dataset. Meanwhile, LDF-BNN can also be well applied to defect detection dataset Mixed WM-38, achieving a high accuracy of 98.70%. [ABSTRACT FROM AUTHOR]

In this study, we investigate the underlying causes of drift in the time history curves of measured parameters obtained through strain electrical measurements and assess their impacts on load measurements. To address the challenge of efficiently processing large volumes of aircraft load data, we propose and analyze a multi-level parallel algorithm specifically designed for the data processing of aircraft load measurements. To achieve this objective, we discuss parallel processing at both medium- and fine-grained levels and develop two distinct parallel processing algorithms: one for coarse- and medium-grained aircraft-type data streams, and another for medium- and fine-grained takeoff and landing data streams. The efficacy of these algorithms is validated through the processing of load data measured on a specific aircraft wing. The results demonstrate that the proposed approach offers a novel technical pathway for large-scale scientific computations and enhances data processing efficiency in the domain of aircraft load spectrum analysis. [ABSTRACT FROM AUTHOR]

Many computation tasks in science and engineering with large amount of data involve complex optimization problems for researchers. Nature Inspired based Algorithms (NIAs) are mostly know for solving complex problems in real world but also these algorithms have one of the major drawbacks of computation cost due to complex problems or increase in the dimension of the problems. Hence, several efforts are made to achieve better computation cost by using parallelization in Nature Inspired Algorithms (NIAs). Parallel Nature Inspired Algorithms (P-NIAs) come forward to overcome the cost of these type of optimization problems. Naturally some of the NIAs are inherently parallel and known for fast processing task. This survey presents a systematic review of existing literature related to parallelization in NIAs. We offered a systematic search to identify the literature through well-reputed publications as IEEE, Elsevier, Springer, ACM, Hindawi, Inderscience, and Wiley. Out of 100 studies 77 reputable studies were selected to identify the current research trends, challenges and opportunities for parallelization in NIAs. Moreover, some problems are also discussed related to algorithm's computation and performance. [ABSTRACT FROM AUTHOR]

Summary: We present a fully‐explicit, iteration‐free, weakly‐compressible method to simulate immiscible incompressible two‐phase flows. To update pressure, we circumvent the computationally expensive Poisson equation and use the general pressure equation which is solved explicitly. In addition, a less diffusive algebraic volume‐of‐fluid approach is used as the interface capturing technique and in order to facilitate improved parallel computing scalability, the technique is discretised temporally using the operator‐split methodology. Our method is fully‐explicit and stable with simple local spatial discretization, and hence, it is easy to implement. Several two‐ and three‐dimensional canonical two‐phase flows are simulated. The qualitative and quantitative results prove that our method is capable of accurately handling problems involving a range of density and viscosity ratios and surface tension effects. [ABSTRACT FROM AUTHOR]

Sustainable construction practices rely on carefully selecting building materials and balancing environmental and economic considerations. This study examines the complex link between local climate, market dynamics, and building material selection. Market data analysis, parametric modeling, and brute‐force optimization are used to provide insights into construction decision‐making. Across 5540 simulations, a thorough assessment of the financial and energy performance of various materials for walls, roofs, windows, and floors is conducted. Incorporating Pareto ranking, parallel simulation, and sensitivity analysis, the comprehensive evaluation reveals the intricate tradeoffs between cost, thermal properties, and energy savings. The findings highlight the potential for optimal external wall solutions to reduce U‐values by up to 30% and achieve source energy savings of up to 25% source energy savings across diverse climates. By emphasizing the importance of local context in material selection, this study highlights how energy consumption patterns and transmission losses influence financial and energy performance, thus advancing sustainable construction practices. [ABSTRACT FROM AUTHOR]

Quantum image histogram as a preprocessing result in quantum image processing contains the gray information of the image and plays an important role in subsequent image processing. As far as we know, there are only a few results on quantum image histogram, and studies on quantum video histogram have not been conducted. So a novel histogram statistic algorithm for quantum video in terms of the idea of parallel computing is proposed in the paper. To this end, the quantum version of carry-lookahead full-adder is first devised, and based on the novel full-adder, an entirely new hierarchical quantum adder for superposition states is also devised, which not only improves the delays generated by mutual carries of classical adder, but also reduces the complexity of classical adder from O (2 m × n) to O (m 2) . Subsequently, in order to realize the parallel statistics of quantum video, the algorithm and circuit implementation of image stitching are also given. Finally, combining the results of image stitching and Grover's search algorithm, the quantum video histogram statistics is ultimately realized in parallel. [ABSTRACT FROM AUTHOR]

To enhance solution accuracy and training efficiency in neural network approximation to partial differential equations, partitioned neural networks can be used as a solution surrogate instead of a single large and deep neural network defined on the whole problem domain. In such a partitioned neural network approach, suitable interface conditions or subdomain boundary conditions are combined to obtain a convergent approximate solution. However, there has been no rigorous study on the convergence and parallel computing enhancement on the partitioned neural network approach. In this paper, iterative algorithms are proposed to enhance parallel computation performance in the partitioned neural network approximation. Our iterative algorithms are based on classical additive Schwarz domain decomposition methods. For the proposed iterative algorithms, their convergence is analyzed under an error assumption on the local and coarse neural network solutions. Numerical results are also included to show the performance of the proposed iterative algorithms. [ABSTRACT FROM AUTHOR]

Despite the indispensable role of traditional electronic computers in modern society, their limitations in parallel processing capabilities, bit-width constraints, and processor bit-width are becoming increasingly apparent, especially when handling large-scale datasets and complex computational tasks. Although hardware technology and algorithm optimization continue to advance, the arithmetic units of traditional computers—adders—remain constrained by carry delay and bit-width limitations. This bottleneck is particularly pronounced in multiplication operations, mainly when adders are used for partial product accumulation. However, since 2018, the emergence of a new type of Reconfigurable Four-Valued Logic Electronic Processor (RFLEP) has provided a potential solution to these traditional limitations. With its large processor bit-width, flexible bit grouping capabilities, and dynamic hardware function reconfiguration features, this processor has brought revolutionary changes to the field of computing. In this context, this paper proposes and implements a Reconfigurable Four-Valued Logic Multiplication Routine (RFLMR) tailored explicitly for the RFLEP. The RFLMR utilizes the Modified Signed-Digit (MSD) representation method in multi-valued logic combined with the M transformation in four-valued logic to generate partial products. These partial products are then efficiently summed in parallel using the JW-MSD parallel adder, achieving the rapid execution of multiplication operations. Experimental results demonstrate that the multiplication routine based on the RFLEP performs multiplication operations accurately and meets theoretical expectations regarding implementation efficiency and performance. This research not only provides new ideas for developing next-generation high-performance computing systems but also paves the way for exploring more efficient and powerful computing models, heralding a profound transformation in future computing technology. [ABSTRACT FROM AUTHOR]

Well placement optimization is one of the most important means to control the decline of oilfields and improve the recovery rate in the development process of deep and heterogeneous reservoirs, such as deep buried carbonate oil reservoirs. However, the mapping relationship from deployed well positions to actual profits is non-linear and multi-modal. At the same time, the injection and production relationship of new wells also affects the contribution of well positions to final profits. Currently, common algorithms include gradient-based and heuristic non-gradient algorithms, which have advantages, but face problems of high computational complexity, slow optimization speed, and difficulty in convergence. We propose an evolutionary algorithm for well placement optimization in carbonate reservoirs. This algorithm improves well placement optimization and computational speed by constraining the sampling process to effective sampling spaces, integrating prior knowledge to enhance sampling efficiency, strengthening local optima exploration, and utilizing parallel computing. Additionally, it refines the optimized variable content based on actual control factors, enhancing the algorithm's robustness in practical applications. A case study from a carbonate reservoir in northwestern China demonstrated that this algorithm not only improved the performance by 50% compared to the classic DE algorithm but also achieved 15% higher optimization effectiveness than the current state-of-the-art algorithms. [ABSTRACT FROM AUTHOR]

Introduction: In response to the issues of complexity and low efficiency in line loss calculations for actual distribution networks, this paper proposes a fast power flow calculation method for distribution networks based on Neo4j graph models and a hierarchical forward-backward sweep parallel algorithm. Methods: Firstly, Neo4j is used to describe the distribution network structure as a simple graph model composed of nodes and edges. Secondly, a hierarchical forward-backward sweep method is adopted to perform power flow calculations on the graph model network. Finally, during the computation of distribution network subgraphs, the method is combined with the Bulk Synchronous Parallel (BSP) computing model to quickly complete the line loss analysis. Results and Discussion: Results from the IEEE 33-node test system demonstrate that the proposed method can calculate network losses quickly and accurately, with a computation time of only 0.175s, which is lower than the MySQL and Neo4j graph methods that do not consider hierarchical parallel computing. [ABSTRACT FROM AUTHOR]

The biological system is not a perfect system, but it is a relatively complete system. It is difficult to realize the lower power consumption and high parallelism that characterize biological systems if lower-level information pathways are ignored. In this paper, we focus on the K, M and P pathways of visual signal processing from the retina to the lateral geniculate nucleus (LGN). We model the visual system at a fine-grained level to ensure efficient information transmission while minimizing energy use. We also implement a circuit-level distributed parallel computing model on FPGAs. The results show that we are able to transfer information with low energy consumption and high parallelism. The Artix-7 family of xc7a200tsbv484-1 FPGAs can reach a maximum frequency of 200 MHz and a maximum parallelism of 600, and a single receptive field model consumes only 0.142 W of power. This can be useful for building assistive vision systems for small and light devices. [ABSTRACT FROM AUTHOR]

C++ template metaprogramming has emerged as a prominent approach for achieving performance portability in heterogeneous computing. Kokkos represents a notable paradigm in this domain, offering programmers a suite of high-level abstractions for generic programming while deferring much of the device-specific code generation and optimization to the compiler through template specializations. Kokkos furnishes a range of device-specific code specializations across multiple back ends, including CUDA and HIP. Diverging from conventional back ends, the OpenACC implementation presents a high-level, multicompiler, multidevice, and directive-based programming model. This paper presents recent advancements in the OpenACC back end for Kokkos (i.e., KokkACC) and focuses on its integration into the Kokkos ecosystem, exploration of automatic device selection capabilities to enhance productivity, and performance evaluation on modern hardware such as NVIDIA H100 GPUs. The study includes implementation details and a thorough performance assessment across various computational benchmarks, including minibenchmarks (AXPY and DOT product), miniapps (LULESH, MiniFE, and SNAP-LAMMPS), and a scientific kernel based on the lattice Boltzmann method. [ABSTRACT FROM AUTHOR]

This paper developed a scalable parallel computing method that can be used for platoon simulations and controller validations. A scalable adaptive platooning control law was firstly designed, which accommodates a variety of vehicle-to-vehicle communication topologies. A road vehicle dynamics model that considered the Magic Formula tyre model and suspension dynamics was then derived and validated. The parallel computing method adopted the Message Passing Interface technique to allow fast and scalable simulations. Platoon length changes do not require controller and algorithm changes. An 11-vehicle platoon on a real-world 10 km long road section was simulated. Different localisation sensor errors, communication delays, heterogenous vehicle masses and driving modes were considered. Results show that localisation errors have negligible influences on space errors. Aggressive driving and heterogeneous vehicle masses slightly increase space errors (increases less than 0.23 m). Communication delays are the greatest influencer for space errors. Increases for 15, 45 and 75 ms delays were 0.43, 1.41 and 2.41 m, respectively. It is further shown that parallel computing can improve the computing speed by three times on personal computers and seven to 12 times on workstations. [ABSTRACT FROM AUTHOR]

In this paper, we develop and study a new mathematical model to describe the good functioning of the biodenitrification process, combining both the biological and the mechanical aspects, and taking into account the chemotaxis-phenomenon. The model is governed by a non-linear reaction-diffusion-advection system for the bacterial activity and the Darcy flow for the porous medium. We prove the existence and uniqueness of the solution and we carry out the numerical approximation of the model within a variational framework. We propose a full discrete system based on a finite element method and we perform numerical simulations both in 2D and 3D. [ABSTRACT FROM AUTHOR]

Prior research has introduced the Single-Instruction-Multiple-Data paradigm for DNA computing (SIMD DNA). It offers the potential for storing information and performing in-memory computations on DNA, with massive parallelism. This paper introduces three new SIMD DNA operations: sorting, shifting, and searching. Each is a fundamental operation in computer science. Our implementations demonstrate the effectiveness of parallel pairwise operations with this new paradigm. [ABSTRACT FROM AUTHOR]

Energy-efficient computation is an increasingly important target in modern-day computing. Scientific computation is conducted using scientific workflows that are executed on highly scalable compute clusters. The execution of these workilows is generally geared towards optimizing run-time performance with the energy footprint of the execution being ignored. Evidently. minimizing both execution time as well as energy consumption does not have to be mutually exclusive. The aim of the research presented in this paper is to highlight the benefits of energy-aware scientific workflow execution. In this paper. a set of requirements for an energy-aware scheduler are outlined and a conceptual architecture for the scheduler is presented. The evaluation of the conceptual architecture was performed by developing a proof of concept scheduler which was able to achieve around 49.97% reduction in the energy consumption of the computation. [ABSTRACT FROM AUTHOR]

For the system containing multiple adjustable degree-of-freedom modular units, a mathematical description of the modular robot topology including a standardized process and a kinematics parallel modeling method are proposed. Firstly, a library of the modular unit motion transfer is established, and a mathematical description of the modular robot topology based on the quadruple matrix with the standardized process is proposed. Then, based on the kinematics model of one modular unit, considering the coupling effect between the branches of the modular robot with multiple motion branches, a position level and speed-level kinematics model is established, which is suitable for chain modular robots. Subsequently, a kinematics parallel modeling method of modular robots is proposed. Both theoretical analytics and stimulation results show that this method can be adapted to the rapid modeling of modular robots with various types of topologies. The introduction of topology mathematical description and parallel computing reduces the difficulty of modular robot modeling and improves modeling efficiency. [ABSTRACT FROM AUTHOR]

As virtual and actual universes merge inside the creating metaverse, requests have pointedly ascended for continuous, intuitive, and intense encounters. The ability of the metaverse to effectively analyze and render complicated links and information supplied by clients is critical for realizing that goal. These demanding computational demands are starting to be supported by parallel processing, and high-performance computing (HPC) is beyond uncertainty key to this domain. The integrative framework presented in this paper addresses the core challenges of inertness, flexibility, and ease of use while integrating equal registration into the metaverse. The system enables prompt handling of client actions and quick response times by distributing calculations over multiple processors, which is essential for the seamless client experience. It also manages the vast amount of metaverse material and interactions as well as the various data processing needs. The paper looks at intrinsic equal processing difficulties in this unique climate, including creating versatile and energy-effective equal calculations that consider load adjusting and asset designation. It features the need to democratize equal figuring assets to produce metaverse extension while accentuating the significance of information protection and security conventions in multiclient settings. The cooperative energy between metaverse development and equal registering progressions vows to push limits, empowering remarkable degrees of virtual submersion and collaboration. [ABSTRACT FROM AUTHOR]

In recent years, various graph computing architectures have been proposed to process graph data that represent complex dependencies between different objects in the world. The designs of the processing element (PE) in traditional graph computing accelerators are often optimized for specific graph algorithms or tasks, which limits their flexibility in processing different types of graph algorithms, or the parallel configuration that can be supported by their PE arrays is inefficient. To achieve both flexibility and efficiency, this paper proposes Grapher, a reconfigurable graph computing accelerator based on an optimized PE array, efficiently supporting multiple graph algorithms, enhancing parallel computation, and improving adaptability and system performance through dynamic hardware resource configuration. To verify the performance of Grapher, this paper selected six datasets from the Stanford Network Analysis Project (SNAP) database for testing. Compared with the existing typical graph frameworks Ligra, Gemini, and GraphBIG, the processing time for the six datasets using the BFS, CC, and PR algorithms was reduced by up to 39.31%, 35.43%, and 27.67%, respectively. The energy efficiency has also been improved by 1.8× compared to Hitgraph and 4.7× compared to ThunderGP. [ABSTRACT FROM AUTHOR]

Copyright of Space-Integrated-Ground Information Networks is the property of Beijing Xintong Media Co., Ltd. and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)

This paper considers jointly scheduling the production and resource-constrained maintenance activities in a manufacturing setting with unrelated parallel machines. In particular, a single maintenance activity needs to be scheduled on each machine in one of its available time windows, and the maintenance activities require a scarce resource, thereby limiting the number of maintenance activities that can be scheduled simultaneously on different machines. In addition, machine- and sequence-dependent setup times, machine eligibility constraints and job-specific release and due dates are considered. A mixed-integer linear program is formulated with objectives including the makespan and, motivated from practice, a weighted sum of total production completion times at machines and total job tardiness. Additionally, a hybrid genetic algorithm with a novel solution representation is proposed for solving industry-scale large instances. A case study is performed with real-world data from a semiconductor manufacturer, where production and maintenance are scheduled separately. The benefit of simultaneously scheduling production and maintenance is investigated. Tests with real-world data show that the proposed model results in schedules that substantially improve the current factory practice. [ABSTRACT FROM AUTHOR]

In recent years, deep neural networks (DNNs) have addressed new applications with intelligent autonomy, often achieving higher accuracy than human experts. This capability comes at the expense of the ever-increasing complexity of emerging DNNs, causing enormous challenges while deploying on resource-limited edge devices. Improving the efficiency of DNN hardware accelerators by compression has been explored previously. Existing state-of-the-art studies applied approximate computing to enhance energy efficiency even at the expense of a little accuracy loss. In contrast, bit-serial processing has been used for improving the computational efficiency of neural processing without accuracy loss, exploiting a simple design, dynamic precision adjustment, and computation pruning. This research presents Serial/Parallel Systolic Array (SPSA) and Octet Serial/Parallel Systolic Array (OSPSA) processing elements for edge DNN acceleration, which exploit bit-serial processing on systolic array architecture for improving computational efficiency. For evaluation, all designs were described at the RTL level and synthesized in 28 nm technology. Post-synthesis cycle-accurate simulations of image classification over DNNs illustrated that, on average, a sample 16 × 16 systolic array indicated remarkable improvements of 17.6% and 50.6% in energy efficiency compared to the baseline, with no loss of accuracy. [ABSTRACT FROM AUTHOR]

In this paper, we deal with the problem of monitoring network latency. Indeed, latency is a key network metric related to both network performance and quality of service, since it directly impacts on the overall user's experience. High latency leads to unacceptably slow response times of network services, and may increase network congestion and reduce the throughput, in turn disrupting communications and the user's experience. A common approach to monitoring network latency takes into account the frequently skewed distribution of latency values, and therefore specific quantiles are monitored, such as the 95th, 98th, and 99th percentiles. We present a comparative analysis of the speed of convergence of the sequential FRUGAL-1U, FRUGAL-2U, and EASYQUANTILE algorithms and the design and analysis of parallel, message-passing-based versions of these algorithms that can be used for monitoring network latency quickly and accurately. Distributed versions are also discussed. Extensive experimental results are provided and discussed as well. [ABSTRACT FROM AUTHOR]