"""
MDSuite: A Zincwarecode package.
License
-------
This program and the accompanying materials are made available under the terms
of the Eclipse Public License v2.0 which accompanies this distribution, and is
available at https://www.eclipse.org/legal/epl-v20.html
SPDX-License-Identifier: EPL-2.0
Copyright Contributors to the Zincwarecode Project.
Contact Information
-------------------
email: zincwarecode@gmail.com
github: https://github.com/zincware
web: https://zincwarecode.com/
Citation
--------
If you use this module please cite us with:
Summary
-------
"""
import numpy as np
[docs]def linear_scale_function(memory_usage: int, scale_factor: int = 1) -> int:
"""
Apply a linear scaling to memory usage.
Parameters
----------
memory_usage : int
naive memory usage, i.e., the exact memory of one configuration.
scale_factor : int
Scalar scaling factor for the memory usage in cases on non-size dependent
inflation.
Returns
-------
scaled_memory : int
Amount of memory required per configuration loaded.
"""
return memory_usage * scale_factor
[docs]def linearithmic_scale_function(memory_usage: int, scale_factor: int = 1) -> float:
"""
Apply a linearithmic scaling to memory usage.
Parameters
----------
memory_usage : int
naive memory usage, i.e., the exact memory of one configuration.
scale_factor : int
Scalar scaling factor for the memory usage in cases on non-size dependent
inflation.
Returns
-------
scaled_memory : float
Amount of memory required per configuration loaded.
"""
return scale_factor * memory_usage * np.log(memory_usage)
[docs]def quadratic_scale_function(
memory_usage: int, inner_scale_factor: int = 1, outer_scale_factor: int = 1
) -> int:
"""
Apply a quadratic scaling to memory usage.
Parameters
----------
memory_usage : int
naive memory usage, i.e., the exact memory of one configuration.
inner_scale_factor : int
Scalar scaling factor for the inner multiplication
outer_scale_factor : int
Scalar scaling factor for the outer multiplication
Returns
-------
scaled_memory : int
Amount of memory required per configuration loaded.
"""
return outer_scale_factor * (memory_usage * inner_scale_factor) ** 2
[docs]def polynomial_scale_function(
memory_usage: int,
inner_scale_factor: int = 1,
outer_scale_factor: int = 1,
order: int = 3,
) -> int:
"""
Apply a polynomial scaling to memory usage.
Parameters
----------
memory_usage : int
naive memory usage, i.e., the exact memory of one configuration.
inner_scale_factor : int
Scalar scaling factor for the inner multiplication
outer_scale_factor : int
Scalar scaling factor for the outer multiplication
order : int
Order of the polynomial.
Returns
-------
scaled_memory : int
Amount of memory required per configuration loaded.
"""
return outer_scale_factor * (memory_usage * inner_scale_factor) ** order