mendevi.models.power_cores¶

Power prediction based on CPU utilisation rate.

Classes

PowerCores(*args, **kwargs)

Affine model to predict the power from the core utilisation.

Details

class mendevi.models.power_cores.PowerCores(*args: tuple, **kwargs: dict)[source]

Affine model to predict the power from the core utilisation.

With \(P_{static}\) and \(P_{core}\) two hyperparameters specific to each machine, but independent of the programme being executed.

Examples¶

>>> from mendevi.models.power_cores import PowerCores
>>> model = PowerCores().fit("<multithread.db>")
>>> model.predict(["paradoxe-32.rennes.grid5000.fr"], [0.0])
{'power': [247.53499079469898]}
>>>

Initialise the model.