Parallel Computing Research Articles

Constraint Realization‐Based Hamel Field Integrator for Geometrically Exact Planar Euler–Bernoulli Beam Dynamics

ABSTRACTIn this article, we first introduce a Hamel field integrator designed for a geometrically exact Euler–Bernoulli beam with infinite‐dimensional holonomic constraints, constructed using a Lagrange multiplier. This method addresses the complexities introduced by constraints, but the additional multiplier introduces a new degree of freedom and hence results in a system with mixed‐type partial differential equations. To address this issue, we further propose a constraint realization method based on perturbation theory for infinite‐dimensional mechanical systems within the framework of Hamel's formalism. This method circumvents the use of additional Lagrange multiplier, significantly reducing the computational complexity of modeling problems. Building on this, we construct a perturbed Hamel field integrator optimized for parallel computing and incorporate artificial viscosity to accelerate constraint convergence. While applicable to three dimensions, our method is demonstrated in a simplified context using planar Euler–Bernoulli beam examples to illustrate the effectiveness of the unified mathematical framework.

Dynamic hybrid parallel computing of the Ray Model for solving underwater acoustic fields in vast sea

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.

Construction of parallel algorithms for the study of complex systems objects with a hierarchical-network structure

The procedure for complexite evaluation of complex systems objects that have a hierarchical-network structure has been formalized.In order to realization this procedure in real time mode, parallel computation algorithms are proposed.Speed up estimates were obtained for the mentioned algorithms, which confirm their high efficiency. The proposed parallel algorithms are oriented for execution on modern computing means with shared and distributed memory: computers with multi-core processors, clusters, hybrid architectures and in high performance distributed environments.

NUMA-aware parallel sparse LU factorization for SPICE-based circuit simulators on ARM multi-core processors

In circuit simulators that resemble the Simulation Program with Integrated Circuit Emphasis (SPICE), one of the most crucial steps is the solution of numerous sparse linear equations generated by frequency domain analysis or time domain analysis. The sparse direct solvers based on lower-upper (LU) factorization are extremely time-consuming, so their performance has become a significant bottleneck. Despite the existence of some parallel sparse direct solvers for circuit simulation problems, they remain challenging to adapt in terms of performance and scalability in the face of rapidly evolving parallel computers with multiple NUMA hardware based on ARM architecture. In this paper, we introduce a parallel sparse direct solver named HLU, which re-examines the performance of the parallel algorithm from the viewpoint of parallelism in pipeline mode and the computing efficiency of each task. To maximize task-level parallelism and further minimize the thread waiting time, HLU devises a fine-grained scheduling method based on an elimination tree in pipeline mode, which employs depth-first search (DFS-like) to iteratively search for parent tasks and then place dependent tasks in the same task queue. HLU also suggests two NUMA node affinity strategies: thread affinity optimization based on NUMA nodes topology to guarantee computational load balancing and data affinity optimization to enable effective memory placement when threads access data. The rationality and effectiveness of the sparse solver HLU are validated by the SuiteSparse Matrix Collection. In comparison with KLU and NICSLU, the experimental results and analysis show that HLU attains a speedup of up to 9.14× and 1.26x (geometric mean) on a Huawei Kunpeng 920 Server, respectively.

Multi-task photonic reservoir computing: wavelength division multiplexing for parallel computing with a silicon microring resonator

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.

Automation of conceptual design and modification of aircraft type unmanned aerial vehicles using multidisciplinary optimization and evolutionary algorithms. Part 1: Methods and models

This paper proposes a method for selecting rational parameters for large-size aircraft-type unmanned aerial vehicles at the initial design stages using an optimization algorithm of differential evolution and numerical mathematical modeling of aerodynamic problems. The method assumes implementation of weight and aerodynamic balance in the main flight modes, it can consider aircraft-type unmanned aerial vehicles with one or two lifting surfaces, applies parallel calculations, and automatically generates a three-dimensional geometric model of the aircraft appearance based on the optimization results. A method for accelerating by more than three times the process of solving the problem of optimizing aircraft takeoff weight parameters by introducing the target function into the set of design variables is proposed and demonstrated. The results of assessing the reliability of the mathematical models used for aerodynamics and the correct calculation of the target function are presented, taking into account various constraints. A comprehensive check of the operability and effectiveness of the method were considered by solving demonstration problems by optimizing more than ten main design parameters of the appearance of two existing heavy-class unmanned aerial vehicles with known characteristics from open sources. Examples of using the optimization results to modify prototypes are provided.

A novel daily runoff forecasting model based on global features and enhanced local feature interpretation

The development of artificial intelligence has introduced new perspectives to the field of hydrological forecasting. However, there is still a lack of research on efficiently identifying the physical characteristics of runoff sequences and developing prediction models that consider global and local sequence features. This study proposes a parallel computing prediction model called IMCAEN (Integrated Multi-Feature Causal Dilated Convolutional Attention Encoder Network) to address these issues. Unlike existing models, this model can monitor fluctuations and anomalies in time series. Incorporating the CDC-AA (Causal Dilated Convolutional Network with Aggregation Attention) and encoder structure captures both local sequence variations and global abrupt anomalies, allowing for comprehensive attention to sequence features. When predicting runoff data from three different hydrological conditions, the IMCAEN model achieved NSEC (Nash-Sutcliffe Efficiency Coefficient) values of 0.98, 0.97, and 0.88, respectively, and outperformed benchmark models in other evaluation indicators as well. Given the opacity of the feature distribution process in AI models, SHAP (SHapleyAdditive exPlanations) analysis and spatial expression of feature distribution are used to assess the contribution of each feature variable to the long-term trend of runoff and to verify the distribution of features trained in each module. The proposed IMCAEN model efficiently captures local and global information in the runoff evolution process through parallel computing and shared features, enabling accurate runoff forecasting and providing critical references for timely warnings and predictions.

A multiscale model for efficiently simulating laser powder bed fusion process with detailed microstructure and mechanical performance results

Numerical simulation is regarded as an important method for studying the mechanisms of additive manufacturing (AM), especially in building a multiscale simulation model. However, the conversion of physical phenomena and parameters across scales is complex, and the number of simulation elements often changes extremely, leading to an inherent conflict between simulation accuracy and efficiency. Therefore, this work introduces a model that balances the need for detailed grain morphology and overall simulation efficiency, to investigate the relationship between process parameters, microstructure, and mechanical performance during Inconel 718 laser powder bed fusion (L-PBF) process. The cellular automata - finite element method (CA-FEM) was used to simulate the melt pool forming and predict the microstructure evolution. Based on the CA result, the tensile process was simulated through crystal plasticity-finite element method (CP-FEM). Specially, the CA-FEM model was simplified and sliced layer-by-layer to reduce simulation time and data storage, supporting parallel computation. The prediction time for the microstructure growth of millions of elements was less than 1.5 h. The CP-FEM model was simplified by reasonably reducing elements in the representative volume element (RVE) model, improving the simulation efficiency by 73 %. The simulation results were validated through EBSD observation and tensile tests, showing good agreement with experimental results, with the final simulation error for mechanical performance not exceeding 10 %. The effects of process parameters on microstructure and mechanical performance were comprehensively discussed. This model supports the optimization of L-PBF process parameters for Ni-based alloys and provides a reference for improving the efficiency of multiscale simulations for other AM processes and materials.

A lightweight dual-link accelerated authentication protocol based on NLFSR-XOR APUF

Physical Unclonable Function (PUF) circuits, as a lightweight hardware security primitive, can provide reliable authentication for resource-constrained Internet of Things (IoT) devices. However, in real-time environments and systems, authentication of embedded devices has strict requirements on resources and low latency. Therefore, this paper proposes a lightweight dual-link accelerated authentication protocol design based on NLFSR-XOR APUF, through the study of Non-Linear Feedback Shift Registers (NLFSR) and XOR APUF circuits. First, the scheme utilizes the symmetric path delay deviation characteristics of APUF and the complexity of NLFSR state transitions to form a nonlinear output function that changes with the challenge signal. Then, a lightweight and attack-resistant authentication protocol is established by combining the random probability of shuffling array bits with the XOR confusion mechanism. Finally, the advantages of GPU parallel computing and AES T-table reconfiguration scheme are used to achieve an accelerated and side-channel attack-resistant authentication protocol. Experimental results show that the PUF circuit can effectively resist various modeling attacks, including logistic regression (LR), artificial neural network (ANN), and support vector machine (SVM). The security of the protocol has been formally verified, and the prototype has been implemented on the Xilinx xc100T development board, effectively resisting deception attacks, physical attacks, and modeling attacks. The protocol's area overhead in terms of LUT and encryption time is reduced by 58.6 % and 67.8 %, respectively, compared to similar protocols.

Predictive pretrained transformer (PPT) for real-time battery health diagnostics

Modeling and forecasting the evolution of battery systems involve complex interactions across physical, chemical, and electrochemical processes, influenced by diverse usage demands and dynamic operational patterns. In this study, we developed a predictive pre-trained Transformer (PPT) model equipped with 1,871,114 parameters that enhance identification of both short-term and long-term patterns in time-series data. This is achieved through the integration of convolutional layers and probabilistic sparse self-attention mechanisms, which collectively enhance prediction accuracy and efficiency in diagnosing battery health. Moreover, the customized hybrid-model fusion supports parallel computing and employs transfer learning, reducing computational costs while enhancing scalability and adaptability. Consequently, this allows for precise real-time health estimations across various battery cycles. We validated this method using a public dataset of 203 commercial lithium iron phosphate (LFP)/graphite batteries charged at rates ranging from 1C to 8C. By using only partial charge data—from an 80 % state of charge to the maximum charging voltage (3.6 V for LFP batteries, 4.2 V for ternary batteries)—and avoiding complex feature engineering, error metrics were achieved below 0.3 % for root mean square error (RMSE), weighted mean absolute percentage error (WMAPE), and mean absolute error (MAE), with an R2 of 98.9 %. The generalization capabilities were further demonstrated across 36 different testing protocols, encompassing 23,480 cycles throughout the entire life cycle, with a total inference time of 9.88 s during the testing phases. Further experiments on 30 nickel cobalt aluminum (NCA) batteries and 36 nickel cobalt manganese (NCM) batteries, across different battery types and operational scenarios, resulted in RMSE, WMAPE, and MAE all below 0.9 %, with R2 values of 94.1 % and 94.4 %, respectively. These findings highlight the potential of our customized deep transfer neural networks to enhance diagnostic accuracy, accelerate training, and improve generalization in real-time applications.

Bayesian blockwise inference for joint models of longitudinal and multistate data with application to longitudinal multimorbidity analysis.

Multistate models provide a useful framework for modelling complex event history data in clinical settings and have recently been extended to the joint modelling framework to appropriately handle endogenous longitudinal covariates, such as repeatedly measured biomarkers, which are informative about health status and disease progression. However, the practical application of such joint models faces considerable computational challenges. Motivated by a longitudinal multimorbidity analysis of large-scale UK health records, we introduce novel Bayesian inference approaches for these models that are capable of handling complex multistate processes and large datasets with straightforward implementation. These approaches decompose the original estimation task into smaller inference blocks, leveraging parallel computing and facilitating flexible model specification and comparison. Using extensive simulation studies, we show that the proposed approaches achieve satisfactory estimation accuracy, with notable gains in computational efficiency compared to the standard Bayesian estimation strategy. We illustrate our approaches by analysing the coevolution of routinely measured systolic blood pressure and the progression of three important chronic conditions, using a large dataset from the Clinical Practice Research Datalink Aurum database. Our analysis reveals distinct and previously lesser-known association structures between systolic blood pressure and different disease transitions.

Object Detection in High-Resolution UAV Aerial Remote Sensing Images of Blueberry Canopy Fruits

Blueberries, as one of the more economically rewarding fruits in the fruit industry, play a significant role in fruit detection during their growing season, which is crucial for orchard farmers’ later harvesting and yield prediction. Due to the small size and dense growth of blueberry fruits, manual detection is both time-consuming and labor-intensive. We found that there are few studies utilizing drones for blueberry fruit detection. By employing UAV remote sensing technology and deep learning techniques for detection, substantial human, material, and financial resources can be saved. Therefore, this study collected and constructed a UAV remote sensing target detection dataset for blueberry canopy fruits in a real blueberry orchard environment, which can be used for research on remote sensing target detection of blueberries. To improve the detection accuracy of blueberry fruits, we proposed the PAC3 module, which incorporates location information encoding during the feature extraction process, allowing it to focus on the location information of the targets and thereby reducing the chances of missing blueberry fruits. We adopted a fast convolutional structure instead of the traditional convolutional structure, reducing the model’s parameter count and computational complexity. We proposed the PF-YOLO model and conducted experimental comparisons with several excellent models, achieving improvements in mAP of 5.5%, 6.8%, 2.5%, 2.1%, 5.7%, 2.9%, 1.5%, and 3.4% compared to Yolov5s, Yolov5l, Yolov5s-p6, Yolov5l-p6, Tph-Yolov5, Yolov8n, Yolov8s, and Yolov9c, respectively. We also introduced a non-maximal suppression algorithm, Cluster-NMF, which accelerates inference speed through matrix parallel computation and merges multiple high-quality target detection frames to generate an optimal detection frame, enhancing the efficiency of blueberry canopy fruit detection without compromising inference speed.

Quantum Nuclear Dynamics on a Distributed Set of Ion-Trap Quantum Computing Systems.

Quantum nuclear dynamics with wavepacket time evolution is classically intractable and viewed as a promising avenue for quantum information processing. Here, we use IonQ, Inc.'s 11-qubit trapped-ion quantum computer, Harmony, to study the quantum wavepacket dynamics of a shared-proton within a short-strong hydrogen-bonded system. We also provide the first application of distributed quantum computing for chemical dynamics problems, where a distributed set of quantum processes is constructed using a tensor network formalism. For a range of initial states, we experimentally drive the ion-trap system to emulate the quantum nuclear wavepacket as it evolves along the potential surface generated from the electronic structure. Following the experimental creation of the nuclear wavepacket, we extract measurement observables such as its time-dependent spatial projection and its characteristic vibrational frequencies to good agreement with classical results. Vibrational eigenenergies obtained from quantum computation are in agreement with those obtained from classical simulations to within a fraction of a kilocalorie per mole, thus suggesting chemical accuracy. Our approach opens a new paradigm for studying the quantum chemical dynamics and vibrational spectra of molecules and also provides the first demonstration of parallel quantum computation on a distributed set of ion-trap quantum computers.

Efficient memristor accelerator for transformer self-attention functionality

The adoption of transformer networks has experienced a notable surge in various AI applications. However, the increased computational complexity, stemming primarily from the self-attention mechanism, parallels the manner in which convolution operations constrain the capabilities and speed of convolutional neural networks (CNNs). The self-attention algorithm, specifically the matrix-matrix multiplication (MatMul) operations, demands a substantial amount of memory and computational complexity, thereby restricting the overall performance of the transformer. This paper introduces an efficient hardware accelerator for the transformer network, leveraging memristor-based in-memory computing. The design targets the memory bottleneck associated with MatMul operations in the self-attention process, utilizing approximate analog computation and the highly parallel computations facilitated by the memristor crossbar architecture. Remarkably, this approach resulted in a reduction of approximately 10 times in the number of multiply-accumulate (MAC) operations in transformer networks, while maintaining 95.47% accuracy for the MNIST dataset, as validated by a comprehensive circuit simulator employing NeuroSim 3.0. Simulation outcomes indicate an area utilization of 6895.7 μm2\\documentclass[12pt]{minimal} \\usepackage{amsmath} \\usepackage{wasysym} \\usepackage{amsfonts} \\usepackage{amssymb} \\usepackage{amsbsy} \\usepackage{mathrsfs} \\usepackage{upgreek} \\setlength{\\oddsidemargin}{-69pt} \\begin{document}$$\\mu m^2$$\\end{document}, a latency of 15.52 seconds, an energy consumption of 3 mJ, and a leakage power of 59.55 μW\\documentclass[12pt]{minimal} \\usepackage{amsmath} \\usepackage{wasysym} \\usepackage{amsfonts} \\usepackage{amssymb} \\usepackage{amsbsy} \\usepackage{mathrsfs} \\usepackage{upgreek} \\setlength{\\oddsidemargin}{-69pt} \\begin{document}$$\\mu W$$\\end{document}. The methodology outlined in this paper represents a substantial stride towards a hardware-friendly transformer architecture for edge devices, poised to achieve real-time performance.

The jump filter in the discontinuous Galerkin method for hyperbolic conservation laws

When simulating hyperbolic conservation laws with discontinuous solutions, high-order linear numerical schemes often produce undesirable spurious oscillations. In this paper, we propose a jump filter within the discontinuous Galerkin (DG) method to mitigate these oscillations. This filter operates locally based on jump information at cell interfaces, targeting high-order polynomial modes within each cell. Besides its localized nature, our proposed filter preserves key attributes of the DG method, including conservation, L2 stability, and high-order accuracy. We also explore its compatibility with other damping techniques, and demonstrate its seamless integration into a hybrid limiter. In scenarios featuring strong shock waves, this hybrid approach, incorporating this jump filter as the low-order limiter, effectively suppresses numerical oscillations while exhibiting low numerical dissipation. Additionally, the proposed jump filter maintains the compactness of the DG scheme, which greatly aids in efficient parallel computing. Moreover, it boasts an impressively low computational cost, given that no characteristic decomposition is required and all computations are confined to physical space. Numerical experiments validate the effectiveness and performance of our proposed scheme, confirming its accuracy and shock-capturing capabilities.

Large scale multi-GPU based parallel traffic simulation for accelerated traffic assignment and propagation

Traffic simulation is a critical tool for congestion analysis, travel time estimation, and route optimization in urban planning, benefiting navigation apps, transportation network companies, and state agencies. Traditionally, traffic micro-simulation frameworks are based on road segments and can only support a limited number of main roads. Efficient traffic simulation on a regional scale remains a significant challenge due to the complexity of urban mobility and the large scale of spatiotemporal data. This paper introduces a Large Scale Multi-GPU Parallel Computing based Regional Scale Traffic Simulation Framework (LPSim), which leverages graphical processing unit (GPU) parallel computing to address these challenges. LPSim utilizes a multi-GPU architecture to simulate extensive and dynamic traffic networks with high fidelity and reduced computation time. Using the parallel processing capabilities of GPUs, LPSim can perform tens of millions of individual vehicle dynamics simulations simultaneously, significantly outperforming traditional CPU-based approaches. The framework is designed to be scalable and can easily accommodate the increasing complexity of traffic simulations. We present the theory behind GPU-based traffic simulation, the architecture of single- and multi-GPU based simulations, and the graph partition strategies that enhance computation resource load balance. Our experimental results demonstrate the effectiveness of LPSim in simulating large-scale traffic scenarios. LPSim is capable of completing simulations of 2.82 million trips in just 6.28 min on a single GPU machine equipped with 5120 CUDA cores (Tesla V100-SXM2). Furthermore, utilizing a Google Cloud instance with two NVIDIA V100 GPUs, which collectively offer 10240 CUDA cores, LPSim successfully simulates 9.01 million trips within 21.16 min. We further tested our simulator with the same demand on dual NVIDIA A100-PCIE-40GB GPUs, which finished the simulation in 0.0398 h, approximately 113 times faster than the same simulation scenario running on an Intel(R) Xeon(R) Gold 6326 CPU @ 2.90 GHz, which takes 4.49 h to complete. This performance not only demonstrates its speed and scalability advantages over traditional simulation techniques but also highlights LPSim’s unique position as the first traffic simulation framework that is scalable for both single- and multiple-GPU configurations. Consequently, LPSim provides an invaluable tool for individuals and extensive research teams alike, enabling the acquisition of large-scale traffic simulation results in a time-efficient manner. LPSim code is available at: https://github.com/Xuan-1998/LPSim

RSD-based high-performance radix-4 Montgomery Modular Multiplication for Elliptic Curve Cryptography

This paper proposes a high-performance radix-4 Montgomery Modular Multiplication (MMM) algorithm and its corresponding hardware architecture for Elliptic Curve Cryptography (ECC), in which the quotient and the partial product accumulation are computed in parallel in each iteration. Additionally, in this MMM, the Redundant Signed Digit (RSD) representation and the Signed Digit Adder (SDA) are used to eliminate the long carry chain and achieve parallel computation, as well as remove pre-computation and integrate modular reduction operations. Our MMM algorithm is implemented in 256-bit and 1024-bit versions on Xilinx Virtex-6 and Virtex-7 FPGAs, respectively. It consumes only 1.55k/10.18k Look-Up Tables (LUTs), takes 133/517 clock cycles, and runs at maximum frequencies of 558.8/641.7 MHz. According to the comparison in terms of Area Time Product (ATP), our design can achieve the ATP of 0.369 over the 256-bit NIST prime domain, which is approximately half of that of the state-of-the-art works. The Scalar Point Multiplication (SPM) scheme using this MMM algorithm consumes 14.19k LUTs and completes a single Scalar Point Multiplication (SPM) operation in 0.217 ms, and it also has a lower ATP than most other SPM algorithms currently in existence.

Trends in brain MRI and CP association using deep learning.

Cerebral palsy (CP) is a neurological disorder that dissipates body posture and impairs motor functions. It may lead to an intellectual disability and affect the quality of life. Early intervention is critical and challenging due to the uncooperative body movements of children, potential infant recovery, a lack of a single vision modality, and no specific contrast or slice-range selection and association. Early and timely CP identification and vulnerable brain MRI scan associations facilitate medications, supportive care, physical therapy, rehabilitation, and surgical interventions to alleviate symptoms and improve motor functions. The literature studies are limited in selecting appropriate contrast and utilizing contrastive coupling in CP investigation. After numerous experiments, we introduce deep learning models, namely SSeq-DL and SMS-DL, correspondingly trained on single-sequence and multiple brain MRIs. The introduced models are tailored with specialized attention mechanisms to learn susceptible brain trends associated with CP along the MRI slices, specialized parallel computing, and fusions at distinct network layer positions to significantly identify CP. The study successfully experimented with the appropriateness of single and coupled MRI scans, highlighting sensitive slices along the depth, model robustness, fusion of contrastive details at distinct levels, and capturing vulnerabilities. The findings of the SSeq-DL and SMSeq-DL models report lesion-vulnerable regions and covered slices trending in age range to assist radiologists in early rehabilitation.

RPEPL: Tibetan Sentiment Analysis Based on Relative Position Encoding and Prompt Learning

Sentiment analysis is a critical task for natural language processing. Much research has been done for high-resource languages such as English and Chinese. However, Tibetan is an extremely low-resource language with less reference information. According to the practical demands, this paper proposes a Tibetan sentiment analysis method based on relative position encoding and prompt learning, the method is abbreviated as RPEPL. First, Word information is introduced to syllable sequences by converting the directed acyclic lattice into a squashed structure. Second, a relative position encoding is used to encode the position information of syllables and words. Then the association relations and semantic information of tokens are identified by leveraging the multi-attention. Finally, the sentiment category of Tibetan sentence is obtained through a prompt learning framework. Experimental results demonstrate that RPEPL significantly outperforms the baseline methods on the TUSA dataset and TNEC dataset. Additionally, Traditional recurrent neural networks cannot perform large-scale parallel computation and convolutional neural networks have difficulty modeling long-distance dependencies in Tibetan text both of which are resolved using RPEPL. Furthermore, the use of a multi-attention not only enriches the association relations between syllables and words, but also enhances the understanding of sentence semantic and syntactic structure information, and improves the performance of Tibetan sentiment analysis.

PARALLEL COMPUTING

Parallel computing is a computational technique in which two or more tasks or instructions can be processed concurrently by means of several processors resulting into enhanced computational efficiency as well as decreased time for solving complex problems. As an approach to solve a particular problem, parallel computing is an attempt that decomposes a problem into constituent subproblems to be solved concurrently; it uses multiple cores, hardware systems to gain advantage in speed The use of parallel computing is prominent in sciences that call for massive computations such as simulations, data mining, artificial intelligence and HPC. Key Words- Parallel computing, Concurrent processing, Computational efficiency, Problem decomposition, Subproblems, Simulations, Data mining, Artificial intelligence, High-performance computing (HPC), Multiple processors, Multiple cores, Hardware systems