Modeling energy consumption is a major concern when designing software systems
such as a cloud service or for a mobile device (e.g., smartphone, laptop).
Unfortunately, we have little knowledge on how decisions taken by a programmer
may impact the energy profile of a system. This paper investigates the energy
consumption in the Pharo programming language. We have closely monitored the
power consumption of five micro-benchmarks and four macro-benchmarks. Our
findings indicate that the way the memory is used has a significant impact on
the power consumption. We propose a code profiler to measure the power
consumption of any Pharo code expression. Our profiler is available under the
MIT license and run on Pharo 5.