Design of an energy aware real-time application development technique

Abstract

The utilization and management of energy currently is a focus area in embedded systems design. Most embedded systems devices operate with limited power budget, or they are battery-powered devices. Software has a notable influence on the global system energy consumption. The evaluation of software energy consumption of application software is vital for lowering the energy consumption to extend the life span of these embedded systems. Two approaches were used. This thesis examined the effect of nesting loop structures (For and While) in C language on energy consumption during the computation of these structures on the STM32F303RE microcontroller (MCU). In the first approach, the average energy consumption per each level of For loop was about 109.2 mJ from level 0 to level 2, and then the increase per each level was 5.25 mJ from level 2 to level 5. The average energy consumption per each level of While loop was about 78.9mJ from level 0 to level 2, and the average energy consumption was about 9.6 mJ from level 2 to level 5. The first approach demonstrated that the STM32F303 RE executing the While loop consumes about 62.41 mJ less per each level than executing the For loop structure. The second approach indicated an exponential increase from 7.39 nJ to 408.98 J when executing the For loop structure. The energy consumption increased proportionally while executing the While loop structure. This implies an increase of 0.362 mJ per each level for the While loop. The second approach also demonstrated that the STM32F303RE consumes about 81.80 J less executing the While loop structure than executing the For loop structure. The second approach also demonstrated an increase in the execution time in relation to the levels in the nesting fashion. The results demonstrated that the nesting loop structures increase the energy consumption in the MCU. The execution time to execute 40000 iterations in the While loop at 8 MHz increases by 25 ms from level 0 to level 2 per each nesting levels and 35 ms from level 2 to level 5 per each nesting levels of the While loop and the execution time to execute 40000 iterations in the For loop at 8 MHz increases by 10 times per each nesting levels of the For loop. Furthermore, it is observed that the nesting levels of a For loop structures present a higher power consumption than the nesting level of While loop structures. The methodology used to achieve this research was discussed and the outcomes were addressed, including the work to be conducted in the future as well.