Calabash automates the execution of comparative software level energy experiments. It uses Scaphandre under the hood, to sample Intel RAPL for power at the host and process level.

Calabash is comprised of two components: Experiment and Analysis.

If using the experimenter module, the following requirements must be met:

• Must be on Linux, preferably with Ubuntu 22.04
• Docker
• Intel processor

The easiest way to use Calabash is by using pip install calabash-experimenter

You may also choose to clone the git repository and run the main function directly. Before doing this install are requirements with pip install -r requirements.txt from the root directory.

Then use add the desired command to the end of $python src/main.py <command>

Before using Calabash, you must create a configuration file according to the specification below.

images:
 - "<dockerhub image name>"
 - ...
out: "<path to directory for output>"
procedure:
  internal_repetitions: <number of repititons within the application (passed into the container)>
  external_repetitions: <number of repititions of the entire image>
  freq: <sampling frequency in nanoseconds>
  cooldown: <seconds in between image runs>
analysis:
    mode: <regex | pid>
    regex: "<regular expression to match on if regex mode is specified>"

With the configuration file you can use the experiment <config_path> or analyze <config_path> to run either module. Note that for proper analysis, the same configuration file should be provided for both.