This study extends the concept of time complexity to energy, i.e., energy complexity, by showing a strong correlation between time complexity and energy consumption for sorting algorithms: Bubble Sort, Counting Sort, Merge Sort and Quick Sort, written in Java and run on single kernels. We investigate the correlation between wall time and time complexity, as well as the correlation between energy consumption and wall time. The primary finding is that time complexity can be used as a guideline to estimate the energy consumption of O(n*n), O(nlog(n)) and O(n + k) sorting algorithms. The secondary finding is that the inputs producing the theoretical worst cases for Merge Sort and Bubble Sort did not produce the worst case wall time nor the worst case energy consumption.
翻译:暂无翻译