Software efficiency or performance optimization at the software and hardware architectural level

Abstract

Some computer scientists point out that the efficiency or performance of software is decreasing when the hardware is becoming more powerful. There are many reasons contributing to this situation. The most important reason is that the cost for developing software is dramatically increasing while, at the same time, computer hardware is becoming cheaper and cheaper. Hence, an intuitive idea for achieving efficient software is to think of a neat way for optimization. Software can be optimized at algorithmic level or at architectural level. The architectural efficiency or performance optimization is mostly about the resource allocation. Therefore, finding a method for resource allocation optimization is crucial to achieve software efficiency or performance optimization at the software architectural level.

This thesis proposes a method which uses the response model R to analyze the relations between software performance/efficiency and resource from the software and hardware architectural perspective, hence provide a neat way for the performance or efficiency optimization.