Performance Parallel Computing Research Articles

Optimal design of electromagnetic repulsion mechanism based on hybrid infill criterion and kriging surrogate model

Abstract In the optimization design of the electromagnetic repulsion mechanism in finite element simulation, there are problems of high computational cost and long optimization time. To shorten the development cycle of the prototype and improve the efficiency of optimization, an optimization algorithm based on the Hybrid Infill Criterion and Kriging Surrogate Model (HIC-KSM) is proposed by organically combining the Expected Improvement (EI) infill criterion and the Pareto front infill criterion to make full use of the performance of parallel computation and to accelerate the convergence speed of the algorithm. The proposed algorithm is compared with the other optimization algorithms, the optimization results of five multi-objective test functions are compared to verify the efficiency of the proposed algorithm, and the algorithm is applied to the parameter optimization of the electromagnetic repulsion mechanism.

Performance on massive MIMO enabled mobile edge computing networks: Parallel computing modeling

With the development of communication and computation technologies, mobile edge computing (MEC) has become an emerging technology being deployed in 5G/B5G mobile networks to improve the mobile user experience. In this paper, we study the asymptotic performance of parallel computing in MEC networks with massive multi-input and multi-output (MIMO) in the millimeter wave band. First, a massive MIMO enabled MEC network with a parallel-limited computing model is constructed, and the task completed probability is defined as a new network performance metric. In the proposed model, we derive the task offloading and computing probabilities via the probability generation functional of the Poisson point process, from which, the task completed probability is obtained. Then we propose a computing resource allocation optimization algorithm to maximize the task completed probability within a given delay constraint. The numerical simulation results prove the accuracy of the obtained theoretical results and verify the performance of the proposed algorithm.

Adaptive neural acceleration unit based on heterogeneous multicore hardware architecture FPGA and software-defined hardware

ABSTRACT This study is anchored in a heterogeneous multicore hardware architecture, specifically Field Programmable Gate Arrays (FPGAs), and software-defined hardware. It employs dynamic planning for the reconfiguration of a software-defined System-on-Chip (SOC) to expedite neural network processes. The reconfigured software-defined SOC serves as an efficient data processor, facilitating the hybrid development and verification of domain-specific System-on-Chip architectures, adhering to the Advanced Microcontroller Bus Architecture (AMBA) standard. The proposed block data trimming methodology significantly enhances overall hybrid computing efficiency by mitigating throughput limitations inherent in Systolic array hardware. The designed Systolic array Matrix Multiply Unit (MMU) is evaluated for a maximum MMU size of 32 × 32 and 1,024 Multiply Accumulator (MAC) units. Hybrid dynamic circuits are devised for the synthesis of int8, int16, int32, and int64 data types and associated logic circuits to validate the performance of parallel computing. It is anticipated that this proposed system can find utility in the development and deployment of unmanned devices, as well as in the integration of deep learning and multi-sensory fusion involving acoustic, tactile, and force sensory components.

Sustainable Material Selection in New Constructions: A Brute‐Force Optimization Framework Using Parallel Computing, Cost Benefits, and Thermal Performance Analysis

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.

Low consumption automatic discovery protocol for DDS-based large-scale distributed parallel computing

DDS (Data Distribution Service) is an efficient communication specification for distributed parallel computing. However, as the scale of computation expands, high network load and memory consumption consistently limit its performance. This paper proposes a low consumption automatic discovery protocol to improve DDS in large-scale distributed parallel computing. Firstly, an improved Bloom Filter called TBF (Threshold Bloom Filter) is presented to compress the data topic. Then it is combined with the SDP(Simple Discovery Protocol) to reduce the consumption of the automatic discovery process in DDS. On this basis, data publication and subscription between the distributed computing nodes are matched using binarization threshold θ and decision threshold T , which can be obtained through iterative optimization algorithms. Experiment results show that the SDPTBF can guarantee higher transmission accuracy while reducing network load and memory consumption, and therefore improve the performance of DDS-based large-scale distributed parallel computing.

LES application to wind pressure prediction for tall building on complex terrain

This study conducted numerical simulations based on the large eddy simulation (LES) model, which employs an unstructured grid system and exhibits high performance in parallel computing, to investigate the effects of a circumferential environment, such as real terrain and surrounding buildings, on the unsteady pressure fields acting on a tall building with a rectangular section. There are three numerical examples using an irregularly shaped high-rise building (aspect ratios: height/length = 7/4 and length/width = 24/11), the same building on a complex terrain, and the same building with surrounding buildings on a complex terrain. To elucidate the statistical pressure characteristics, we focused on the mean, fluctuating, and peak values on the surfaces of the target building. The pressures for the present models were validated through a comparison with experimental data obtained under the same conditions. Subsequently, the velocity and pressure fields directly influenced by the complex local terrain and surrounding buildings were studied. A concave terrain located on the leeward side of the target building could cause a large negative pressure on the side surfaces as a result of access to the deformed separating flow. In addition, the surrounding buildings led to the movement of the stagnation point to a higher location on the windward surface and intensification of the conical vortex attached to the side surface.

Parallel swarm-based algorithms for scheduling independent tasks

Task scheduling is crucial for achieving high performance in parallel computing. Since task scheduling is NP-hard, the efficient assignment of tasks to compute resources remains an issue. Across the literature, several algorithms have been proposed to solve different scheduling problems. One group of promising approaches in this field is formed by swarm-based algorithms which have a potential to benefit from a parallel execution. Common swarm-based algorithms are Ant Colony Optimization (ACO) and Particle Swarm Optimization (PSO). In this article, we propose two new scheduling methods based on parallel ACO, PSO and, Hill Climbing, respectively. These algorithms are used to solve the problem of scheduling independent tasks onto heterogeneous multicore platforms. The results of performance measuements demonstrate the improvements on the makespan and the scheduling time achieved by the parallel variants.

High fidelity transient solver in STREAM based on multigroup coarse-mesh finite difference method

This study incorporates a high-fidelity transient analysis solver based on multigroup CMFD in the MOC code STREAM. Transport modeling with heterogeneous geometries of the reactor core increases computational cost in terms of memory and time, whereas the multigroup CMFD reduces the computational cost. The reactor condition does not change at every time step, which is a vital point for the utilization of CMFD. CMFD correction factors are updated from the transport solution whenever the reactor core condition changes, and the simulation continues until the end. The transport solution is adjusted once CMFD achieves the solution. The flux-weighted method is used for rod decusping to update the partially inserted control rod cell material, which maintains the solution's stability. A smaller time-step size is needed to obtain an accurate solution, which increases the computational cost. The adaptive step-size control algorithm is robust for controlling the time step size. This algorithm is based on local errors and has the potential capability to accept or reject the solution. Several numerical problems are selected to analyze the performance and numerical accuracy of parallel computing, rod decusping, and adaptive time step control. Lastly, a typical pressurized LWR was chosen to study the rod-ejection accident.

Adaptation of Parallel SaaS to Heterogeneous Co-Located Cloud Resources

Cloud computing has received increasing attention due to its promise of delivering on-demand, scalable, and virtually unlimited resources. However, heterogeneity or co-location of virtual cloud resources can cause severe degradation of the efficiency of parallel computations because of a priori unknown application-specific performance metrics, load imbalance, and limitations of memory bandwidth. This paper presents the runtime adaptation of parallel discrete element method (DEM) Software as a Service (SaaS) to heterogeneous or co-located resources of the OpenStack cloud. The computational workload is adapted by using weighted repartitioning and runtime measured performance of parallel computations on Docker containers. The high improvement in performance up to 48.7% of the execution time is achieved, applying the runtime adapted repartitioning when the load imbalance is high enough. The low load imbalance leads to the close values of computational load, when small variations in the system load and performance can cause oscillations in subsets of particles. Memory stress tests cause heterogeneity of non-isolated containers, which reduces the performance of memory bandwidth bound DEM SaaS on the co-located resources. The runtime adapted repartitioning handles the constant and periodically variable performance of non-isolated containers and decreases the total execution time of DEM SaaS.

Performance Analysis of Parallel Computing in Image Classification

The use of parallel computing can speed up the training of deep learning models. The traditional neural network model ResNet is chosen for testing in this paper on the effectiveness of data parallelism in image classification, and test data is provided in a 6-GPU environment. In this paper, it is suggested that various factors should be considered when building affordable hardware configurations to expedite model training in practical application scenarios. The communication costs are not insignificant because today's large computing clusters are primarily offered via cloud computing. Another crucial point to remember is that the number of CUDA cores is the primary hardware foundation for GPU acceleration technology. Therefore, acceleration may not be affected by more incredible video memory or fewer CUDA cores for some particular graphics cards. In addition, the issue of beyond-model performance in parallel computing is another issue that must be disregarded. Due to the parallel strategy's limitations, it is essential to tweak the super parameters while speeding up model training. The model's performance is more likely to be ensured by the lower learning rate and Batch size. This paper's experimental conclusion can support building an appropriate hardware configuration scheme.

Evaluating the Performance of Parallel Computing in Hybrid Models

Evaluating the Performance of Parallel Computing in Hybrid Models

On extending Amdahl’s law to learn computer performance

The problem of learning parallel computer performance is investigated in the context of multicore processors. Given a fixed workload, the effect of varying system configuration on performance is sought. Conventionally, the performance speedup due to a single resource enhancement is formulated using Amdahl’s law. However, in case of multiple configurable resources the conventional formulation results in several disconnected speedup equations that cannot be combined together to determine the overall speedup. To solve this problem, we propose to (1) extend Amdahl’s law to accommodate multiple configurable resources into the overall speedup equation, and (2) transform the speedup equation into a multivariable regression problem suitable for machine learning. Using experimental data from fifty-eight tests spanning two benchmarks (SPECCPU 2017 and PCMark 10) and four hardware platforms (Intel Xeon 8180M, AMD EPYC 7702P, Intel CoffeeLake 8700K, and AMD Ryzen 3900X), analytical models are developed and cross-validated. Findings indicate that in most cases, the models result in an average cross-validated accuracy higher than 95%, thereby validating the proposed extension of Amdahl’s law. The proposed methodology enables rapid generation of multivariable analytical models to support future industrial development, optimization, and simulation needs.

Thermal finite element analysis of complex heat sinks using open source tools and high-performance computing

The modeling of heat transfer phenomena in thermal systems has been extensively explored in industry and academia by using the finite element method (FEM) with commercial software. However, when the thermal problem introduces complexities in geometry and physics, the availability of licenses for high-performance computing could represent a limitation to achieving results in a reasonable time. Hence, finite element analysis (FEA) using open-source software (OSS) becomes a prominent candidate in this case. Therefore, multiple open-source tools are integrated into this work to solve the heat transfer equation, including conduction, convection, and radiation. Several geometrically complex heat sinks commonly used in the electronics industry are considered application examples. The performance of parallel computing is assessed in terms of processing time. The finite element solution engine is built by implementing the energy balance equations in their weak formulation in Firedrake, using its solver PETSc, the mesh generator GMSH and the post-processor Paraview, thus creating a fully OSS-based Python framework. Finally, the results are verified with commercial software for different case studies, and its potential to be extended to other fields of engineering is evident.

상호연결망 NSEP Graph와 Circle Graph에 대한 Mathematica 코딩 기반의 시각화 분석

Research on interconnection networks is continuously needed to improve the performance of parallel processing computers. Using Mathematica, ambiguous parts of the graph can be specifically implemented and tested, and efficient and accurate analysis is possible by reducing meaningless guesses or time. It is also reusable, making it convenient to analyze the additional properties of the graph later. The interconnection network NSEP has a graph with one edge added to the SEP graph with a fixed degree 3, and the CG graph has a degree 3. NSEP graphs and CG graphs require further research in addition to the properties analyzed to date. To this end, three fixed degree graphs, including SEP, the leading graph of NSEP, were designed and new properties were analyzed in Mathematica. NSEP, an interconnected network with a fixed degree, has a graph with one edge added to the SEP graph with a fixed degree 3, and the CG graph has a degree 3. A graph with a fixed degree has an advantage in that the degree does not increase as the dimension increases, but has a disadvantage in that the number of nodes increases rapidly.
 Compared to the existing graph definition, the diameter of SEP and NSEP analyzed in Mathematica was smaller. The diameter of the SEP is 1/8(9n^2-22n+24), and the diameter of the NSEP is 2/3n^2-3/2n+1. Since CG has the same result as the existing diameter A, it can be seen that the optimal routing algorithm is proven.
 This study showed how to generate interconnection networks using Mathematica coding and presented the results of efficiently analyzing the discrete structure by exploring the graph properties as well as providing visualization methods that can be used as field-based teaching contents for discrete mathematics.

Initial guess estimation and fast solving of petroleum complex molecular reconstruction model

AbstractWe proposed a fast method to solve the petroleum molecular reconstruction problem via initial guess estimation coupled with local optimization algorithm. A virtual oil database was constructed. After inputting the bulk properties of the target oil, several virtual oils with similar bulk properties were selected and their model parameters were used as the initial guesses. Subsequently, we proved that by using a local optimization method, the optimal model parameters can be rapidly obtained. The diesel and vacuum gas oil compositional models were exemplified as case studies. The results show that the proposed method presented faster computational time, which is within the range of 20–50 times that of frequently used global algorithms. Moreover, the optimal solution of the proposed method exhibited higher accuracy. In addition, the effect of the number of virtual oils and the parallel computation performance were also investigated.

Parallel Frequent Subtrees Mining Method by an Effective Edge Division Strategy

Most data with a complicated structure can be represented by a tree structure. Parallel processing is essential to mining frequent subtrees from massive data in a timely manner. However, only a few algorithms could be transplanted to a parallel framework. A new parallel algorithm is proposed to mine frequent subtrees by grouping strategy (GS) and edge division strategy (EDS). The main idea of GS is dividing edges according to different intervals and then dividing subtrees consisting of the edges in different intervals to their corresponding groups. Besides, the compression stage in mining is optimized by avoiding all candidate subtrees of a compression tree, which reduces the mining time on the nodes. Load balancing can improve the performance of parallel computing. An effective EDS is proposed to achieve load balancing. EDS divides the edges with different frequencies into different intervals reasonably, which directly affects the task amount in each computing node. Experiments demonstrate that the proposed algorithm can implement parallel mining, and it outperforms other compared methods on load balancing and speedup.

A Parallel Scanner for the Concurrent Execution of Lexical Analyzer Tasks on Multi-Core Machines using Dynamic Task Allocation Algorithm

Purpose: Massive multi-core architecture is rapidly becoming the standard in digital technology due to its high and parallel computational capability and performance benefits. To fully utilize the technological capabilities of multi-core systems, system software such as compilers should be re-engineered for parallel processing. Several important contributions have been made in the past to enhance the efficiency of the lexical analysis process by leveraging the innate parallel processing capability of multi-core computers. This trend of implementation shows that a parallel lexical analyzer tends to perform lexing tasks better than a conventional sequential lexical analyzer. This article discusses the way of making the tasks in parallel during the scanning of source program in the phase of lexical analysis. The objective of this study is to explore how to perform lexical analysis in parallel. On multi-core processors, multiple processes of the lexical analyzer program can run concurrently to scan multiple lines in the input stream in parallel for token detection. This is done by allocating tasks line-by-line to the core which is not engaged yet. Design/Methodology/Approach: Developing a theoretical and experimental approach for parallelizing the lexical scanning process on a multi-core system. Findings/Result: According to the theoretical and experimental results, the proposed methodology significantly outperforms the sequential approach in terms of tokenization. It considerably reduces the time required for lexical analysis during the compilation process. The result establishes unequivocally that the parallel lexical analyzer's performance should scale linearly with the number of cores. It is clearly observed that the speedup is expected to increase further if the number of CPU cores increases. This enhancement would speed up the compilation process even more. Originality/Value: A dynamic task allocation algorithm is developed for the concurrent execution of a lexical analyzer task on multi-core systems. Paper Type: Experimental Research.

Parallel-GPU-accelerated adaptive mesh refinement for three-dimensional phase-field simulation of dendritic growth during solidification of binary alloy

In the phase-field simulation of dendrite growth during the solidification of an alloy, the computational cost becomes extremely high when the diffusion length is significantly larger than the curvature radius of a dendrite tip. In such cases, the adaptive mesh refinement (AMR) method is effective for improving the computational performance. In this study, we perform a three-dimensional dendrite growth phase-field simulation in which AMR is implemented via parallel computing using multiple graphics processing units (GPUs), which provide high parallel computation performance. In the parallel GPU computation, we apply dynamic load balancing to parallel computing to equalize the computational cost per GPU. The accuracy of an AMR refinement condition is confirmed through the single-GPU computations of columnar dendrite growth during the directional solidification of a binary alloy. Next, we evaluate the efficiency of dynamic load balancing by performing multiple-GPU parallel computations for three different directional solidification simulations using a moving frame algorithm. Finally, weak scaling tests are performed to confirm the parallel efficiency of the developed code.

Метричні системи оцінки ефективності та масштабованості паралельних обчислень

"The current state of development of computer technology allows you to build parallel computer systems that use almost unlimited number of processors. The availability of such systems has aroused interest in studying the performance of parallel computers, which contain a large number of processors, in the implementation of real multidimensional problems. One way to increase the efficiency of parallel architectures is to reduce the time required to perform a time-consuming task, which should be commensurate with the number of processing resources used to solve this problem. The second direction is the development of highly scalable parallel or parallel algorithms. Under the scalability of the parallel algorithm on the parallel architecture we will consider a measure of its ability to efficiently use a growing number of processors. Scalability analysis can be used to select the best combination of algorithm architecture for a problem with different constraints on the size of the problem and the number of processors. It can be used to predict the performance of a parallel algorithm and a parallel architecture for a large number of processors based on known performance on a smaller number of processors. For a fixed size of the problem, it can be used to determine the optimal number of processors, which will be used and the maximum possible acceleration that can be obtained. Scalability analysis can also predict the impact of changes in hardware technology on performance and thus help develop the best parallel architectures to solve different problems The aim of the work is to critically evaluate the state of modern theory of analysis of aircraft performance and scalability and to demonstrate further research on the development of new and more complex analytical tools to analyze the effective use of the benefits of parallel equipment. The main task of the study is to develop new and modify existing theoretical models, methods and formalisms to study the problems of efficiency and scalability of parallel computing. Mathematically, to simplify the analysis, it is assumed that all temporal characteristics are non-negative. This means that acceleration is always limited by the number of processors, p, and efficiency - by one. For example, acceleration can be super-linear, overhead can be negative if memory is hierarchical and access time is increased discretely by increasing the memory used by the program. In this case, the effective computing speed of a large program will be slower on a serial processor than on a parallel computer with similar processors."

Performance of Communication- and Computation-Intensive SaaS on the OpenStack Cloud

The pervasive use of cloud computing has led to many concerns, such as performance challenges in communication- and computation-intensive services on virtual cloud resources. Most evaluations of the infrastructural overhead are based on standard benchmarks. Therefore, the impact of communication issues and infrastructure services on the performance of parallel MPI-based computations remains unclear. This paper presents the performance analysis of communication- and computation-intensive software based on the discrete element method, which is deployed as a service (SaaS) on the OpenStack cloud. The performance measured on KVM-based virtual machines and Docker containers of the OpenStack cloud is compared with that obtained by using native hardware. The improved mapping of computations to multicore resources reduced the internode MPI communication by 34.4% and increased the parallel efficiency from 0.67 to 0.78, which shows the importance of communication issues. Increasing the number of parallel processes, the overhead of the cloud infrastructure increased to 13.7% and 11.2% of the software execution time on native hardware in the case of the Docker containers and KVM-based virtual machines of the OpenStack cloud, respectively. The observed overhead was mainly caused by OpenStack service processes that increased the load imbalance of parallel MPI-based SaaS.