学位论文详细信息
Efficient mobile computing
big.LITTLE;Linux scheduler;Efficiency;Mobile processor
Bohloolizamani, Ali ; Torrellas ; Josep
关键词: big.LITTLE;    Linux scheduler;    Efficiency;    Mobile processor;   
Others  :  https://www.ideals.illinois.edu/bitstream/handle/2142/97250/BOHLOOLIZAMANI-THESIS-2017.pdf?sequence=1&isAllowed=y
美国|英语
来源: The Illinois Digital Environment for Access to Learning and Scholarship
PDF
【 摘 要 】
Smart handheld devices such as phones, tablets and watches are becoming more and more common rapidly. From a computer architect point of view, processor design for such computer systems is a complex problem. Since the Li-ion battery manufacturing technology is strictly limited by physical and technological limitations, the new generation of mobile processors should have a better energy efficiency to support an acceptable battery life. On the other hand, TLP of current mobile applications is measured to be mostly less than 2, which implies mobile processors should have high performance on user's demand to provide an acceptable QoE. By shrinking the chip manufacturing technology size, SoC design has been the most preferred integration approach in such applications. For example, Apple's A10, iPhone7 SoC, has more than 3 billion transistors including 4 big.LITTLE cores, 6 GPU cores and caches. Due to power-density and heat-dissipation constraints of such integration level, providing high performance on demand in an efficient way is a complex control problem. In the A10 architecture, assigning the right cores at the right time to running threads is a challenging complex control problem. The state of the art systems have control loops for controlling architectural parameters in different ways. In mobile devices controllers are heuristics-based mainly for simplicity. Considering the power-density and heat-dissipation issues in such systems, we propose an OS architecture and interface to provide an environment for improving the functionality of controllers in mobile computer systems.
【 预 览 】
附件列表
Files Size Format View
Efficient mobile computing 2038KB PDF download
  文献评价指标  
  下载次数:11次 浏览次数:36次