Article
Computer Science, Theory & Methods
Jianhua Gao, Weixing Ji, Zhaonian Tan, Yizhuo Wang, Feng Shi
Summary: This article presents a new approach for compressed binary sparse matrix-vector multiplication, which reduces data transfer and improves computational performance through partitioning and encoding.
IEEE TRANSACTIONS ON PARALLEL AND DISTRIBUTED SYSTEMS
(2022)
Article
Computer Science, Hardware & Architecture
Thaha Mohammed, Aiiad Albeshri, Iyad Katib, Rashid Mehmood
Summary: Sparse linear algebra plays a crucial role in various fields like engineering, science, and business. Current SpMV kernels underutilize the potential of GPUs and lack in-depth research on GPU performance. DIESEL, a deep learning-based tool, is proposed to predict and execute the best-performing SpMV kernel, improving performance metrics compared to existing AI-based tools.
JOURNAL OF SUPERCOMPUTING
(2021)
Article
Chemistry, Multidisciplinary
Shizhao Chen, Jianbin Fang, Chuanfu Xu, Zheng Wang
Summary: Optimizing sparse matrix-vector multiplication is challenging due to the non-uniform distribution of non-zero elements. This paper presents a new hybrid storage format and employs machine learning to automatically select the appropriate storage format based on the target matrix and hardware. Experimental results show significant performance improvement on different multi-core CPU platforms.
APPLIED SCIENCES-BASEL
(2022)
Article
Computer Science, Interdisciplinary Applications
Nileshchandra K. Pikle, Shailesh R. Sathe, Arvind Y. Vyavahare
Summary: Assembly free FEM improves performance by optimizing Matrix-vector Products (MvP) strategy for coalesced global memory access on GPU using shared memory and texture cache. Despite the issue of low occupancy, the proposed strategy outperforms both element based and DoF based strategies on GPU.
ENGINEERING WITH COMPUTERS
(2022)
Article
Computer Science, Hardware & Architecture
Hengshan Yue, Xiaohui Wei, Jingweijia Tan, Nan Jiang, Meikang Qiu
Summary: This article proposes an energy-efficient ECC mechanism called Eff-ECC for GPGPU register files, utilizing the error sensitivity of instructions, the duplicate characteristics of the same-named registers, and the error sensitivity of data bits. Experimental results demonstrate that Eff-ECC significantly reduces energy consumption.
IEEE TRANSACTIONS ON COMPUTER-AIDED DESIGN OF INTEGRATED CIRCUITS AND SYSTEMS
(2022)
Article
Computer Science, Hardware & Architecture
Ehsan Atoofian
Summary: This study presents a method for dynamically detecting and bypassing trivial instructions in GPGPUs, reducing energy consumption by minimizing waste of hardware resources and improving energy efficiency. By bypassing trivial instructions, energy consumption of GPGPUs can be reduced by 8% with negligible impact on performance.
IEEE EMBEDDED SYSTEMS LETTERS
(2021)
Article
Mathematics, Applied
Siham Boukhris, Artem Napov, Yvan Notay
Summary: We propose a new sparse matrix format for discretized partial differential equations with piecewise-constant coefficients. This format saves memory and is suitable for parallel computing on GPUs. It is well suited for algebraic multigrid methods and has been shown to outperform other solvers in terms of both run time and memory usage.
SIAM JOURNAL ON SCIENTIFIC COMPUTING
(2023)
Article
Computer Science, Hardware & Architecture
Ernesto Dufrechou, Pablo Ezzatti, Enrique S. Quintana-Orti
Summary: This work reviews the development of efficient GPU routines for the sparse matrix-vector product and evaluates different implementations using machine learning techniques. The experiments show that selecting the optimal method for a given matrix is challenging due to the complexity of matrix structures, but machine learning can accurately predict the best method with over 80% accuracy, potentially leading to significant reductions in execution time and energy consumption.
INTERNATIONAL JOURNAL OF HIGH PERFORMANCE COMPUTING APPLICATIONS
(2021)
Article
Computer Science, Software Engineering
Jose I. Aliaga, Hartwig Anzt, Enrique S. Quintana-Orti, Andres E. Tomas
Summary: In this work, a hybrid sparse matrix layout called CSRC is proposed, which combines the flexibility of well-known sparse formats and offers appealing properties, such as low cost conversion from CSR format, similar storage requirements, and high performance for both direct and transposed products on modern graphics accelerators. This solution significantly improves performance when integrated into iterative algorithms for truncated singular value decomposition.
CONCURRENCY AND COMPUTATION-PRACTICE & EXPERIENCE
(2023)
Article
Computer Science, Artificial Intelligence
Mengjia Yin, Xianbin Xu, Tao Zhang, Conghuan Ye
Summary: This paper focuses on establishing a performance evaluation model, analyzing performance bottlenecks quantitatively, and providing guidance for optimizing them. Using matrices as an example, it examines dense and sparse matrices, developing evaluation models and relational functions based on performance analysis. Through practical tests, the model proves to be practical with stable deviation thresholds.
INTERNATIONAL JOURNAL OF PATTERN RECOGNITION AND ARTIFICIAL INTELLIGENCE
(2021)
Article
Computer Science, Theory & Methods
Aleksei Sorokin, Sergey Malkovsky, Georgiy Tsoy
Summary: This research focuses on the performance of general matrix multiplication routines on modern heterogeneous computing systems, comparing IBM and Intel CPUs and finding that IBM systems perform best on GPUs for matrix multiplication.
JOURNAL OF PARALLEL AND DISTRIBUTED COMPUTING
(2022)
Article
Computer Science, Theory & Methods
Najeeb Ahmad, Buse Yilmaz, Didem Unat
Summary: Sparse Triangular Solve (SpTRSV) is a crucial part of scientific computing, but its performance can be limited by varying parallelism characteristics. This work introduces a split-execution model for SpTRSV to automatically divide computation and improve parallelism, with experimental results showing significant speedups on CPU-GPU platforms.
IEEE TRANSACTIONS ON PARALLEL AND DISTRIBUTED SYSTEMS
(2021)
Article
Computer Science, Software Engineering
Jose I. Aliaga, Hartwig Anzt, Thomas Gruetzmacher, Enrique S. Quintana-Orti, Andres E. Tomas
Summary: Our study introduces a variant of the coordinate sparse matrix format to optimize the sparse matrix-vector product. This approach balances workload distribution and compresses indexing arrays and numerical information, leading to improved performance on various platforms including multicore processors and GPUs.
CONCURRENCY AND COMPUTATION-PRACTICE & EXPERIENCE
(2022)
Article
Computer Science, Hardware & Architecture
Jose Aliaga, Hartwig Anzt, Thomas Gruetzmacher, Enrique S. Quintana-Orti, Andres E. Tomas
Summary: Krylov methods provide a fast and parallel tool for solving large-scale sparse linear systems iteratively. This article investigates techniques to reduce communication and memory costs in order to improve the performance of practical implementations. By integrating a communication-reduction strategy into the GMRES solver using Ginkgo's memory accessor, the memory accesses can be effectively hidden, resulting in accelerated iterative steps. The customization of the storage format, based on the properties of the orthogonal basis, further enhances the performance of the solver.
INTERNATIONAL JOURNAL OF HIGH PERFORMANCE COMPUTING APPLICATIONS
(2023)
Article
Computer Science, Information Systems
Seog Chung Seo
Summary: Researchers proposed an efficient implementation of the SIKE mechanism on GPUs, optimizing underlying field arithmetic and taking full advantage of the GPU architecture. Experimental results showed that the GPU software outperformed the SIKE CPU software on Intel i9-10900K CPU by a factor of 140.64-146.81.
Article
Engineering, Manufacturing
Jufeng Wang, Chunfeng Liu, MengChu Zhou, Tingting Leng, Aiiad Albeshri
Summary: This study selects a proper number of required types of processing modules (PMs) to process wafers, ensuring the highest productivity of a wafer-residency-time-constrained dual-arm cluster tool. It proposes the necessary and sufficient conditions for tool schedulability and develops a polynomial-complexity algorithm for finding an optimal cyclic schedule. Examples are provided to demonstrate its superiority over existing approaches, advancing the field of cluster tool scheduling and promoting green manufacturing of wafers for semiconductor producers.
IEEE TRANSACTIONS ON SEMICONDUCTOR MANUFACTURING
(2023)
Article
Computer Science, Information Systems
Bander Alzahrani, Nikos Fotiou, Aiiad Albeshri, Abdullah Almuhaimeed, Khalid Alsubhi
Summary: ICN is an emerging paradigm that enables secure retrieval of content items independent of their location. This paper proposes a solution that allows third-party storage nodes to verify user authorization for accessing specific content items, leveraging Verifiable Credentials to build trust chains and express user capabilities. The solution enables users to prove authorization using a single message integrated into a content request and eliminates the need for verifying entities to store any secrets. It also supports lightweight delegation.
INTERNATIONAL JOURNAL OF INFORMATION SECURITY
(2023)
Article
Automation & Control Systems
SiYa Yao, Qi Kang, MengChu Zhou, Muhyaddin J. Rawa, Aiiad Albeshri
Summary: This article proposes an efficient discriminative manifold distribution alignment (DMDA) approach, which improves feature transferability by aligning both global and local distributions and refines a discriminative model by learning geometrical structures in manifold space. Extensive experiments show that DMDA outperforms other methods in both classification accuracy and time efficiency in domain adaptation tasks.
IEEE TRANSACTIONS ON SYSTEMS MAN CYBERNETICS-SYSTEMS
(2023)
Article
Chemistry, Multidisciplinary
Xiaoling Wang, Qi Kang, Mengchu Zhou, Zheng Fan, Aiiad Albeshri
Summary: Multi-task optimization (MTO) is a new evolutionary computation paradigm that solves multiple optimization tasks concurrently by utilizing task similarities and historical knowledge. This work proposes the individually guided multi-task optimization (IMTO) framework, which explores each individual to learn from other tasks, selects individuals with higher solving ability, and only inferior individuals learn from other tasks to improve knowledge transfer. The advantage of IMTO over multifactorial evolutionary framework and baseline solvers is verified through benchmark studies.
APPLIED SCIENCES-BASEL
(2023)
Article
Computer Science, Information Systems
Thaha Mohammed, Si-Ahmed Naas, Stephan Sigg, Mario Di Francesco
Summary: Today's DNNs are accurate but require large amount of data. This work proposes a knowledge sharing method by exchanging weights of pretrained DNNs and using transfer learning. It utilizes a market-based approach for optimal knowledge sharing and introduces a weight fusion technique. Evaluation shows that the proposed solution is efficient and significantly improves inference accuracy without the need of federated learning.
IEEE INTERNET OF THINGS JOURNAL
(2023)
Article
Computer Science, Information Systems
Yuzhe Xu, Thaha Mohammed, Mario Di Francesco, Carlo Fischione
Summary: This article addresses the problem of DNN inference allocation in edge computing, proposing a realistic DNN inference model and a distributed algorithm to solve it. Experimental results show that the proposed solution significantly outperforms existing techniques in terms of inference time, load balance, and convergence speed.
IEEE INTERNET OF THINGS JOURNAL
(2023)
Review
Computer Science, Information Systems
Sardar Usman, Rashid Mehmood, Iyad Katib, Aiiad Albeshri
Summary: Big data has transformed science and technology, bringing about societal changes. High-performance computing (HPC) supports big data analysis using AI and methods. Efforts have been made to combine HPC and big data into converged architectures for improved performance and resource efficiency.
Article
Environmental Sciences
Sarah Alswedani, Rashid Mehmood, Iyad Katib, Saleh M. Altowaijri
Summary: Mental health issues have significant impacts and addressing the root causes is crucial for prevention and sustainability. A holistic approach is needed to understand mental health in the context of social and environmental factors. More research, awareness, and interventions are necessary to address these issues, including studying the effectiveness and risks of medications.
Article
Mathematics, Interdisciplinary Applications
Fang Guo, Mengzhuo Luo, Jun Cheng, Iyad Katib, Kaibo Shi
Summary: This paper investigates the problem of nonfragile observer-based tracking control for a class of fuzzy fast sampling singularly perturbed systems with sensor saturation, event-triggered scheme, and random cyber-attacks. The proposed control protocol improves design flexibility and reduces conservativeness to avoid asynchronous phenomenon between the systems. By integrating the fuzzy nonfragile observer and reference model signal into the tracking controller design, the tracking error can be reduced.
CHAOS SOLITONS & FRACTALS
(2023)
Article
Engineering, Civil
Junqi Zhang, Huan Liu, Peng Zu, Mengshi Zhao, Cheng Wang, Aiiad Albeshri, Abdullah Abusorrah, MengChu Zhou
Summary: Recently, the use of a particle swarm optimizer (PSO) to guide robots in a source location problem has gained attention. Traditional obstacle avoidance strategies are not effective when robots lack prior information. This work proposes a novel PSO based on Tabu Search (PSO-TS) that sets trapping areas as tabu objects to enable robots to locate multiple sources without prior knowledge or expensive hardware.
IEEE TRANSACTIONS ON INTELLIGENT TRANSPORTATION SYSTEMS
(2023)
Article
Automation & Control Systems
Gang Qin, An Lin, Jun Cheng, Mengjie Hu, Iyad Katib
Summary: This study investigates the issue of event-triggered fault detection filtering for memristive neural networks with dynamic quantization in the discrete-time domain. A novel event-triggered protocol is proposed based on dynamic quantization parameter, fault occurrence probability, and network bandwidth utilization rate, and an asynchronous filter framework is developed to ensure the stability of the system.
JOURNAL OF THE FRANKLIN INSTITUTE-ENGINEERING AND APPLIED MATHEMATICS
(2023)
Article
Automation & Control Systems
Wei Kang, Gang Qin, Jun Cheng, Huaicheng Yan, Iyad Katib, Jinde Cao
Summary: This paper proposes a security control method for a discrete-time switched power system using a probabilistic event-triggered protocol, which effectively optimizes network resource utilization and improves system security and stability under multi-strategy deception attacks.
JOURNAL OF THE FRANKLIN INSTITUTE-ENGINEERING AND APPLIED MATHEMATICS
(2023)
Article
Computer Science, Artificial Intelligence
Junhui Wu, Gang Qin, Jun Cheng, Jinde Cao, Huaicheng Yan, Iyad Katib
Summary: This paper proposes an innovative approach to mitigate the effects of deception attacks in Markov jumping systems by developing an adaptive neural network control strategy. The approach effectively approximates the unbounded false signals injected by deception attacks and establishes a connection between the joint Markov chain and controller.
Article
Computer Science, Information Systems
Iyad Katib, Fatmah Y. Assiri, Turki Althaqafi, Zenah Mahmoud Alkubaisy, Diaa Hamed, Mahmoud Ragab, Heung-Il Suk
Summary: Smart Fintech, empowered by data science and artificial intelligence, drives automated, intelligent, personalized financial and economic businesses, playing a crucial role in today's technology-driven society and economies.
Editorial Material
Computer Science, Information Systems
Juan M. M. Corchado, Sara Rodriguez, Fernando de la Prieta, Pawel Sitek, Vicente Julian, Rashid Mehmood