AlgorithmBase class
An abstract base class of load/work balancing algorithms.
Classes
- class __metaclass__
- Metaclass for defining Abstract Base Classes (ABCs).
Static methods
- def factory(algorithm_name: str, parameters: dict, work_model: Model.lbsWorkModelBase.WorkModelBase, logger: logging.Logger, rank_qoi: str, object_qoi: str)
- Instantiate the necessary concrete algorithm.
Methods
- def _AlgorithmBase__print_QOI(self, rank_or_obj)
- Print list of implemented QOI when invalid QOI is given.
- def _initialize(self, p_id, phases, distributions, statistics)
- Factor out pre-execution checks and initalizations.
- def _report_final_mapping(self, logger)
- Report final rank object mapping in debug mode.
- def _update_distributions_and_statistics(self, distributions: dict, statistics: dict)
- Compute and update run distributions and statistics.
- def execute(self, p_id, phases, distributions, statistics, a_min_max)
- Execute balancing algorithm on Phase instance.
- def get_initial_communications(self)
- Return the initial phase communications.
- def get_rebalanced_phase(self)
- Return phased assigned for processing by algoritm.
Special methods
- def __init__(self, work_model: Model.lbsWorkModelBase.WorkModelBase, parameters: dict, logger: logging.Logger, rank_qoi: str, object_qoi: str)
- Class constructor.
Method documentation
def lbaf.
Execute balancing algorithm on Phase instance.
:param: p_id: index of phase to be rebalanced (all if equal to _) :param: phases: list of Phase instances :param: distributions: dictionary of load-varying variables :param: statistics: dictionary of statistics :param: a_min_max: possibly empty list of optimal arrangements.
def lbaf.
Class constructor.
:param work_model: a WorkModelBase instance :param parameters: a dictionary of parameters :param rank_qoi: rank QOI to track :param object_qoi: object QOI to track.