The performance of lattice-Boltzmann solver implementations usually depends mainly on memory access patterns. Achieving high performance requires then complex code which handles careful data placement and ordering of memory transactions. In this work, we analyse the performance of an implementation based on a new approach called the data-oriented language, which allows the combining of complex memory access patterns with simple source code. As a use case, we present and provide the source code of a solver for D2Q9 lattice and show its performance on GTX Titan Xp GPU for dense and sparse geometries up to 4096 2 nodes. The obtained results are promising, around 1000 lines of code allowed us to achieve performance in the range of 0.6 to 0.7 of maximum theoretical memory bandwidth (over 2.5 and 5.0 GLUPS for double and single precision, respectively) for meshes of size above 1024 2 nodes, which is close to the current state-of-the-art. However, we also observed relatively high and sometimes difficult to predict overheads, especially for sparse data structures. The additional issue was also a rather long compilation, which extended the time of short simulations, and a lack of access to low-level optimisation mechanisms.

Increasingly, genomics is being used for the prediction of specific traits and diseases (phenotypes) among humans. Wider availability of genomics data through multiple research projects (such as International HapMap Project1 and 1000 Genomes2) has been a catalyst in that direction. With the recent advances in machine learning and big data analysis, data computation resources and data models needed for genomics data analysis are readily available. However, the prediction of traits and diseases has its own challenges in terms of computational requirements and computational analysis, statistical analysis (example: confounding variables), and limited quality of data collection. Linear Mixed Models (LMM, a type of linear regression) is a common approach for Genome-wide Association Studies (GWAS) for the prediction of common traits among humans using genomics. This paper researches the existing LMM-based approaches for Genome-wide Association Studies (GWAS), describes the experiment performed on FaST-LMM approach from Microsoft Research, and then proposes an enhanced approach (called LMM-22) on how to address computational and statistical issues. LMM-22 focuses on the parallelization of LMM computations and execution of LMM-22 on General Purpose Graphics Processing Units (GPU) as against CPUs to accelerate the LMM approach for GWAS studies.

Tablets are the most used solid dosage form in medical treatments. The realistic modeling 1 of such specific object shapes in computer simulations has been demonstrated already by the author 2 for bi-convex round tablets but needs to be extended for a wider range of tablet shapes. Moreover, 3 the number of tablets must be large enough to simulate industrial drum coaters, which are typically 4 used for film coating processes. Whilst this long lasting process where tablets flow continuously, 5 tablets are loaded by frequent changes of individual contacts. Using a tailored simulation tool named 6 pharmAssist , written for GPUs (modern massively parallel processors in graphics cards and HPC 7 accelerators), we implemented the widely used oblong tablet shape. We are able to simulate flows 8 involving industrial scale number of such tablets. Data mining allows for statistical analyzis of 9 pair-wise contacts of different geometric features. Our simulation results serves as an observation 10 space, which cannot be addressed by experimental analyzis and will help to avoid degradation of 11 tablets like chipping or other defect phenomena. This paper reflects current development work in 12 order to extend the DEM with a new shape class, which comprises the modelling of compacted or 13 extruded objects, which have a smooth exterior, a flat band and sharp contour edges

Face recognition is a valuable forensic tool for criminal investigators since it certainly helps in identifying individuals in scenarios of criminal activity like fugitives or child sexual abuse. It is, however, a very challenging task as it must be able to handle low-quality images of real world settings and fulfill real time requirements. Deep learning approaches for face detection have proven to be very successful but they require a large computation power and processing time. In this work, we evaluate the speed-accuracy tradeoff of three popular deep-learning-based face detectors on the WIDER Face and UFDD data sets in several CPUs and GPUs. We also develop a regression model capable to estimate the performance, both in terms of processing time and accuracy. We expect this to become a very useful tool for the end user in forensic laboratories in order to estimate the performance for different face detection options. Experimental results showed that the best speed-accuracy tradeoff is achieved with images resized to 50% of the original size in GPUs and images resized to 25% of the original size in CPUs. Moreover, performance can be estimated using multiple linear regression models with a Mean Absolute Error (MAE) of 0.113 what is very promising for the forensic field.

Smoothed particle hydrodynamics (SPH) is regarded as a pure Lagrangian approach, which can solve fluid dynamics problems without the creation of mesh. In this paper, a paralleled SPH solver is developed to solve particle-based computational acoustics (PCA). The aim of this paper is to study the feasibility of using SPH to solve acoustic problems and to improve the efficiency of solving processes by paralleling some procedures on GPU during calculating. A stand SPH code running serially in a CPU is proposed to solve wave equation. This is a wave propagating in a two-dimensional domain. After finishing the computation, the results are compared with the theoretical solutions and they agree well. So its feasibility is verified. There are two main methods for searching neighbor particles: all-pair search method and linked-list search method. Both methods are used in different codes to simulate an identical problem and their runtimes are compared to investigate their searching efficiencies. The runtime results show that linked-list search method has a higher efficiency, which can save a lot of searching time when simulating problems with huge amounts of particles. Furthermore, the percentages of different procedures’ runtimes in a simulation are also discussed to find the most consuming one. Then, some codes are modified to run in different GPUs and their runtimes are compared with those of serial ones on a CPU. Runtime results show that the paralleled algorithm can be more than 80 times faster than the serial one. The result shows that GPU paralleled SPH computing can achieve desirable accuracy and speed in solving acoustic problems.

The development of the smart city concept and the inhabitants’ need to reduce travel time, as well as society’s awareness of the reduction of fuel consumption and respect for the environment, lead to a new approach to the classic problem of the Travelling Salesman Problem (TSP) applied to urban environments. This problem can be formulated as “Given a list of geographic points and the distances between each pair of points, what is the shortest possible route that visits each point and returns to the departure point?” Nowadays, with the development of IoT devices and the high sensoring capabilities, a large amount of data and measurements are available, allowing researchers to model accurately the routes to choose. In this work, the purpose is to give solution to the TSP in smart city environments using a modified version of the metaheuristic optimization algorithm TLBO (Teacher Learner Based Optimization). In addition, to improve performance, the solution is implemented using a parallel GPU architecture, specifically a CUDA implementation.

With the development of engineering technology, engineering has higher requirements for the accuracy and the scale of simulation calculation. The computational efficiency of traditional se-rial program can not meet the requirements of engineering。Therefore, reducing the calcula-tion time of temperature control simulation program has important engineering significance for real-time simulation of temperature field and stress field, and then adopting more reasona-ble temperature control and crack prevention measures. GPU parallel computing is introduced into the temperature control simulation program of massive concrete to solve this problem and the optimization is carried out. Considering factors such as GPU clock rate, number of cores, parallel overhead and Parallel Region, The improved GPU parallel algorithm analysis indicator formula is proposed. It makes up for the shortcomings of traditional formula that focus only on time. According to this formula, when there are enough threads, the parallel effect is limited by the size of the parallel domain, and when the parallel domain is large enough, the efficiency is limited by the parallel overhead and the clock rate. This paper studies the optimal Kernel execu-tion configuration. Shared Memory is utilized to improve memory access efficiency by 155%. After solving the problem of bank conflicts, an accelerate rate of 437.5x was realized in the sub-routine of the matrix transpose of the solver. The asynchronous parallel of data access and logi-cal operation is realized on GPU by using CUDA Stream , which can overlap part of the data access time. On the basis of GPU parallelism, asynchronous parallelism can double the compu-ting efficiency. Compared with the serial program, the accelerate rate of inner product matrix multiplication of the GPU asynchronous parallel program is 61.42x. This study further proposed a theoretical formula of data access overlap rate to guide the selection of the number of CUDA streams to achieve the optimal computing conditions. The GPU parallel program compiled and optimized by CUDA Fortran platform can effectively improve the computational efficiency of the simulation program for concrete temperature control, and better serve for engineering computing.

This paper presents a multi-swarm PSO algorithm for the Quadratic Assignment Problem (QAP) implemented on the OpenCL platform. Our work was motivated by results of time efficiency tests performed for single-swarm algorithm implementation that showed clearly that the benefits of a parallel execution platform can be fully exploited provided the processed population is large. The described algorithm can be executed in two modes: with independent swarms or with migration. We discuss the algorithm construction as well as we report results of tests performed on several problem instances from the QAPLIB library. During the experiments the algorithm was configured to process large populations. This allowed us to collect statistical data related to values of goal function reached by individual particles. We use them to demonstrate on two test cases that although single particles seem to behave chaotically during the optimization process, when the whole population is analyzed, the probability that a particle will select a near-optimal solution grows.

This paper aims to develop a low-cost, high-performance and high-reliability computing system to process large-scale data using common data mining algorithms in the Internet of Things computing. Considering the characteristics of IoT data processing, similar to mainstream high performance computing, we use a GPU cluster to achieve better IoT services. Firstly, we present an energy consumption calculation method (ECCM) based on WSN. Then, using the CUDA Programming model, we propose a Two-level Parallel Optimization Model (TLPOM) which exploits reasonable resource planning and common compiler optimization techniques to obtain the best blocks and threads configuration considering the resource constraints of each node. The key to this part is dynamic coupling Thread-Level Parallelism (TLP) and Instruction-Level Parallelism (ILP) to improve the performance of the algorithms without additional energy consumption. Finally, combining the ECCM and the TLPOM, we use the Reliable GPU Cluster Architecture (RGCA) to obtain a high-reliability computing system considering the nodes’ diversity, algorithm characteristics, etc. The results show that the performance of the algorithms significantly increased by 34.1%, 33.96% and 24.07% for Fermi, Kepler and Maxwell on average with TLPOM and the RGCA ensures that our IoT computing system provides low-cost and high-reliability services.

This is the second paper in a cycle investigating the exact solution of loop equations in decaying turbulence. We perform numerical simulations of the Euler ensemble, suggested in the previous work, as a solution to the loop equations. We designed novel algorithms for simulation, which take a small amount of computer RAM so that only the CPU time grows linearly with the size of the system and its statistics. This algorithm allows us to simulate systems with billions of discontinuity points on our fractal curve, dual to the decaying turbulence. For the vorticity correlation function, we obtain \textbf{quantum fractal laws} with regime changes between universal discrete values of indexes. The traditional description of turbulence with fractal or multifractal scaling laws does not apply here. The measured conditional probabilities of fluctuating variables are smooth functions of the logarithm of scale, with statistical errors being negligible at our large number of random samples $T = 167,772,160$ with $N = 200,000,000$ points on a fractal curve. The quantum jumps arise from the analytical distribution of scaling variable $X = \frac{\cot^2(\pi p/q)}{q^2}$ where $\frac{p}{q}$ is a random fraction. This distribution is related to the totient summatory function and is a discontinuous function of $X$. In particular, the energy spectrum is a devil's staircase with tilted uneven steps with slopes between $-1$ and $-2.03$. The logarithmic derivative of energy decay $n(t)$ as a function of time is jumping down the stairs of universal levels between $1.50$ and $1.53$. The quantitative verification of this quantization would require more precise experimental data.

Ship wave is of great interest for wave drag, coastal erosion and ship detection. In this paper, a highly-paralleled numerical scheme is proposed for simulating three-dimensional (3-D) nonlinear Kelvin ship waves effectively. First, a numerical model for nonlinear ship waves is established based on potential flow theory, boundary integral method and Jacobian-free Newton-Krylov (JFNK) method. In order to improve computational efficiency and reduce data storage of JFNK method, a banded preconditioner method is then developed by formulating the optimal bandwidth selection rule. After that, a Graphics Process Unit (GPU) based parallel computing framework is designed, and a GPU solver is developed by using Compute Unified Device Architecture (CUDA) language. Finally, numerical simulations of 3-D nonlinear ship waves under multiple scales are performed by using the GPU and CPU solvers. Simulation results show that the proposed GPU solver is more efficient than the CPU solver with the same accuracy. More than 66% GPU memory can be saved and the computational speed can be accelerated up to 20 times. Hence, the computation time for Kelvin ship waves simulation can be significantly reduced by applying the GPU parallel numerical scheme, which lays a solid foundation for practical ocean engineering.

In this work, we introduce a scalable and efficient GPU-accelerated methodology for volumetric particle advection and finite-time Lyapunov exponent (FTLE) calculation, focusing on the analysis of Lagrangian Coherent Structures (LCS) in large-scale Direct Numerical Simulation (DNS) datasets across incompressible, supersonic, and hypersonic flow regimes. LCS play a significant role in turbulent boundary layer analysis, and our proposed methodology offers valuable insights into their behavior in various flow conditions. Our novel owning-cell locator method enables efficient, constant-time cell search, and the algorithm draws inspiration from classical search algorithms and modern multi-level approaches in numerical linear algebra. The proposed method is implemented for both multi-core CPUs and Nvidia GPUs, demonstrating strong scaling up to 32,768 CPU cores and up to 62 Nvidia V100 GPUs. By decoupling particle advection from other problems, we achieve modularity and extensibility, resulting in consistent parallel efficiency across different architectures. Our methodology was applied to calculate and visualize the FTLE on four turbulent boundary layers at different Reynolds and Mach numbers, revealing that coherent structures grow more isotropic proportional to the Mach number, and their inclination angle varies along the streamwise direction. We also observed increased anisotropy and FTLE organization at lower Reynolds numbers, with structures retaining coherency along both spanwise and streamwise directions. Additionally, we demonstrated the impact of lower temporal frequency sampling by upscaling with an efficient linear upsampler, preserving general trends with only 10% of the required storage. In summary, we present a particle search scheme for particle advection workloads in the context of visualizing LCS via FTLE that exhibits strong scaling performance and efficiency at scale. Our proposed algorithm is applicable across various domains requiring efficient search algorithms in large structured domains. While this manuscript focuses on the methodology and its application to LCS, an in-depth study of the physics and compressibility effects in LCS candidates will be explored in a future publication.

This work considers the task of representation learning on the attributed relational graph (ARG). Both the nodes and edges in an ARG are associated with attributes/features allowing ARGs to encode rich structural information widely observed in real applications. Existing graph neural networks offer limited ability to capture complex interactions within local structural contexts, which hinders them from taking advantage of the expression power of ARGs. We propose Motif Convolution Module (MCM), a new motif-based graph representation learning technique to better utilize local structural information. The ability to handle continuous edge and node features is one of MCM’s advantages over existing motif-based models. MCM builds a motif vocabulary in an unsupervised way and deploys a novel motif convolution operation to extract the local structural context of individual nodes, which is then used to learn higher-level node representations via multilayer perceptron and/or message passing in graph neural networks. When compared with other graph learning approaches to classifying synthetic graphs, our approach is substantially better in capturing structural context. We also demonstrate the performance and explainability advantages of our approach by applying it to several molecular benchmarks.

Quantitative differential phase contrast (qDPC) imaging has become an important method of optical measurement and life science research in microscopy because of its high reconstruction resolution and non-invasive, high-contrast and quantitative imaging of biological samples. Despite the continuous development of the principle and algorithm, the frame rate of the existing qDPC algorithm is still much lower than that of camera acquisition, so it is hardly applied to real-time image the fast-moving biological samples. In this paper, based on color-coded multiplexing strategy, a compact real-time quantitative phase imaging system is designed to realize multi-mode imaging. The system employs a programmable LED array to illuminate directly, and the phase reconstruction algorithm is deployed in the graphics processing unit (GPU) of the laptop to accelerate the calculation. The system can achieve high-speed quantitative phase imaging of non-stained biological samples, and the frame rate can reach 60fps. The device has the advantages of compact structure, low cost and portability. Thus, it is suitable for mobile medical applications.

Synesthesia is a psychological phenomenon where sensory signals become mixed. Input to one sensory modality produces an experience in a second, unstimulated modality. In “grapheme-color synesthesia”, viewed letters and numbers evoke mental imagery of colors. The study of this condition has implications for increasing our understanding of brain architecture and function, language, memory and semantics, and the nature of consciousness. In this work, we propose a novel application deep learning to model perception in grapheme-color synesthesia. Achromatic letter images, taken from database of handwritten characters, are used to induce synesthesia. Results show the model learns to accurately create a colored version of the inducing stimulus, according to a statistical distribution from experiments on a sample population of grapheme-color synesthetes. The spontaneous, creative mental imagery characteristic of the synesthetic perceptual experience is reproduced by the model. A model of synesthesia that generates testable predictions on brain activity and behavior is needed to complement large scale data collection efforts in neuroscience, especially when articulating simple descriptions of cause (stimulus) and effect (behavior). The research and modeling approach reported here begins to address this need.

In recent years, demand has been increasing for target detection and tracking from aerial imagery via drones using onboard powered sensors and devices. We propose a very effective method for this application based on a deep learning framework. A state-of-art embedded hardware system empowers small flying robots to carry out the real-time onboard computation necessary for object tracking. Two types of embedded modules were developed: one is designed using a Jetson TX or AGX Xavier, and the other is based on an Intel Neural Compute Stick. These are suitable for real-time onboard computing power on small flying drones with limited space. A comparative analysis of current state-of-art deep-learning-based multi-object detection algorithms was carried out utilizing the designated GPU-based embedded computing modules to obtain detailed metric data about frame rates as well as the computation power. We also introduce an effective target tracking approach for moving objects. The algorithm for tracking moving objects is based on the extension of simple online and real-time tracking. It was developed by integrating a deep-learning-based association metric approach (Deep SORT), which uses a hypothesis tracking methodology with Kalman filtering and a deep-learning-based association metric. In addition, a guidance system that tracks the target position using a GPU-based algorithm is introduced. Finally, we demonstrate the effectiveness of the proposed algorithms by real-time experiments with a small multi-rotor drone.

Clustering is a data mining task used for the data extraction from the data bases or files. Clustering is used to find unknown groups present in the data sources like files or data bases. This paper focuses on clustering algorithms performance dependency on the parallel clustering platforms and the clustering algorithms along with their clustering criteria. The problems with the present Traditional clustering algorithms were throughput and data source size changes (scalability). So they can’t address the big data. So for handling the huge volumes of data, parallel clustering algorithms along with clustering criteria were used. For processing the big Data Parallel clustering algorithms are of two types based on computing platforms used. They were the horizontal scaling platforms and vertical scaling platforms.

This study explores the integration of Artificial Intelligence (AI) and collaborative robotics in industrial surface finishing, focusing on feature detection on object reflective surfaces using neural networks in computationally constrained environments. Central to this investigation is the utilization of cameras considered as sensors that can capture detailed images of reflective surfaces. These images are then analyzed by a custom neural network, demonstrating the feasibility of feature detection with non-powerful computers. The public dataset for recognizing light reflection defects, available in the SPADD repository, supports this approach. Employing open-source tools like Darknet YOLOv3 and Yolo\_mark, this method proves particularly advantageous for small and medium-sized enterprises (SMEs) due to its low cost and maintenance requirements. This research highlights the challenges of AI and collaborative robotics in industrial applications, especially in adapting to varying conditions and the need for continuous training with new data coming from several sensors. Looking ahead, plans are in place to integrate this system with cobot-assisted polishing tasks, using cameras to enhance manufacturing efficiency and quality. This work contributes to the broader understanding of practical applications of AI in industrial automation, surface finishing processes, and the effective use of sensor technology.