Make HPOlib intensify-aware

Right now, most HPOlib functions work on the basis of configurations, e.g. the number of jobs is measured in configurations and not in the number of target algorithm runs. Also, most plotting functions behave this way. It would be more logical to change this behaviour in a way that the standard is the number of target algorithm run instead of number of configurations.

The issue names no files, tests, or entry points. Start by locating the functions that count jobs and the plotting functions, then trace how configurations and target algorithm runs are represented. Done would require a consistent run-based default across the affected functions, with tests or examples confirming the new behavior.

Written by the indexing model from the issue text.