학술논문

Energy efficient middleware: Design and development for mobile applications
Document Type
Conference
Source
2017 19th International Conference on Advanced Communication Technology (ICACT) Advanced Communication Technology (ICACT), 2017 19th International Conference on. :541-549 2017
Subject
Communication, Networking and Broadcast Technologies
Computing and Processing
Engineering Profession
Batteries
Smart phones
Power demand
Layout
Arrays
Memory management
Semiconductor optical amplifiers
Data Layout Transformation DLT
Mobile Computing
Energy Efficiency
Array of structure AOS
Structure of arrays SOA
Language
Abstract
Over the recent years, the popularity of smartphones has increased dramatically. The advanced integrated technology in smartphones like GPS, high-speed CPU, a real world coloured display, Wi-Fi and Bluetooth etc. All these within small size light weight device attracts people a lot to obtain them. The stated capacities of those components motivate developers to create millions of useful applications. However, smartphone devices are energy constraint as they rely on limited battery power supply that has not been increased at the same pace to support the power demands. As both the hardware and the software tend to drain the battery power, the demand for energy efficient applications has increased to keep the mobile devices useful. Optimization related to memory data access create significant difference to performance and power consumption of broad range of dataintensive application. Memory Data layout transformation represents a very interesting class of optimizations. Transform Array of Structure (AOS) to Structure of Array (SOA) is one of the commonly applied and recognized transformation. The transformation reduces the memory access count and subsequently reduces the memory access energy. Thereby, we introduce data layout transformation service as solution to minimize the power consumed by application. The Service will convert the data layout in memory from AOS to SOA. The conversion will reduce the power consumed by memory and processor. Eventually, result in efficient and extended battery life.