The standalone calibration tool is designed to calibrate the model that is used for physical hosts.

These scripts are intended to be used on a SLURM based environment. They are to be used as follows:

submit a calibration run job using:

sbatch ./sbatchSubmit.sh

Make a note of its job id, then run the following command once it has completed:

extract_profiling.sh <job_id>

This extracts data from SLURM ready for processing. The extracted data can then be merged into a single dataset:

./merge.sh <core_count>

The count of cores the processor has is an optional extension to this command. The default is 16.
The next command is used to prepare the dataset for loading into the database of the energy modeller.

./tosql.sh <filter_time>

TThe filter time indicates how many seconds of data should be removed from the start of each load period. This is variable, but is recommended to be changed based upon the stability of the numbers gained. This may require graphing or other examination of variance to understand if the values obtained are undergoing changes, which may be due to averaging windows for example.