Articles

High-Performance Computing (HPC) is indispensable for large-scale scientific simulations, but achieving optimal performance on modern supercomputers is increasingly challenging. As HPC systems scale toward exascale, they face escalating complexity in hardware, software, and workloads. Traditional optimization methods (manual tuning and heuristic algorithms) struggle to cope with dynamic workloads and intricate system behaviors. Artificial Intelligence (AI) techniques offer a promising approach to address these challenges. This article provides an overview of AI-powered optimization in HPC, focusing on how machine learning and related AI methods enhance performance, efficiency, and scalability of scientific simulations. We survey key AI techniques applied to HPC optimization including machine learning for performance modeling, deep reinforcement learning for resource management, and AI-driven surrogate models for accelerating simulations and illustrate their impact through case studies in domains such as job scheduling and fluid dynamics. We discuss the practical applications of these techniques, highlighting reported performance gains (e.g., substantial reductions in simulation run time and improved resource utilization). We also examine the challenges in integrating AI with HPC (such as training overhead, data movement, and reliability concerns) and outline future directions for research. The convergence of AI and HPC is poised to produce “smart” simulation workflows that intelligently adapt and optimize in real time, pushing the frontiers of scientific computing.

The recent development of Mobile System-on-Chip (SoC) has completely changed the functionality of the contemporary smartphones, wearables, and IoT devices through enhanced computational strength but with stringent power and heat constraints. Mobile applications are becoming more complex (e.g., real-time computer vision, high-resolution imaging, on-device artificial intelligence, etc.) and SoCs have to provide high throughput, good work load distribution, and responsive multitasking without sacrificing battery life. To overcome these issues, modern SoCs are based on heterogeneous multicore designs, with big. LITTLE designs, special purpose accelerators, like NPUs and ISPs, and memory arranged hierarchies and backed with high bandwidth buses. Also, more sophisticated power-management approaches like Dynamic Voltage and Frequency Scaling (DVFS), clock gating, and power gating provide fine-tuning energy management in the active and idle conditions. Most recently, scheduling and thermal sensitivity resource allocation using AI has become a promising approach in increasing real-time refinement and flexibility. This paper gives a detailed discussion of performance-driven and power-efficient design methods in mobile SoCs, including architectural innovations, workload optimization, and built-in AI-assisted energy management. The discussion identifies major research trends, system-level developments, and new challenges defining next-generation mobile computing systems.

Data warehousing is a technique for collecting, managing, and presenting data to help people analyze and use that data effectively. It involves a large database designed to support management-level staff by providing all the relevant historical data for analysis. This chapter begins with a definition of data warehousing, followed by an overview of large-scale policy management to highlight the need for data warehousing. Next, an overview of an ETL framework is presented, along with a discussion of advanced ETL techniques. The chapter concludes with an outline of performance optimization techniques for data warehousing. Data warehousing is considered a key enabler for efficient reporting and analysis, with implementation choices ranging from cost-effective desktop systems to large-scale, mission-critical data marts and warehouses containing petabytes of data. Extract, transform, and load (ETL) systems remain one of the largest cost and effort areas within data warehouse development projects, requiring significant planning and resources to build, manage, and monitor the flow of data from source systems into the data warehouse. The technology and techniques used for ETL can greatly influence the success or failure of a data warehouse. Complex business requirements for data cleansing, loading, transformation, and integration have intensified, while operational plans for real-time and near-real-time reporting add additional challenges. Parallel loading mechanisms, incremental data loading, and runtime update and insert strategies not only improve ETL performance but also optimize data warehousing performance, particularly for large-scale policy management.