Does the compiler or interpreter version influence the energy consumption of programming languages?

Abstract

Software plays a crucial role in our daily activities. Virtually all the technology we use contains software components written in a particular programming language. In this context, compilers and interpreters play an important role, as they are needed to convert the software source code into a format that can be executed by a machine. The significant influence of the programming language on the energy consumption of the resulting programs has been highlighted in some research. However, there is almost no research on the impact of the programming language compiler/interpreter version of the programming language on the energy consumption. This paper aims to fill this gap by investigating the impact of the compiler/interpreter version on the energy consumption of programs written in C, Java and Python. To do that we have performed a study that uses a hardware-based energy measurement approach to obtain the energy consumed by eight algorithms written in the three languages and run with different compiler/interpreter versions. The results do not show a trend of improvement between versions within each language, especially in terms of energy consumption. These results suggest that energy efficiency does not seem to be a major factor when developing compilers/interpreters and should therefore be prioritized.