diff --git a/Appli-iso3dfd/ACO.py b/Appli-iso3dfd/ACO.py
index 5d3025259662a80830294f28f0ce1e0e53806e35..cd3f9eb72121ba51651a9d23cc486e9e91a1c7e8 100644
--- a/Appli-iso3dfd/ACO.py
+++ b/Appli-iso3dfd/ACO.py
@@ -9,24 +9,23 @@ import tools
import launcher_SUBP
def ACO(Me, NbP, comm, alpha, rho, Q, nb_ants, tau_0, n_iter, n1_max=256, n2_max=256, n3_max=256,
- fancy_strategy='AS', sub_threshold=0.1, sup_threshold=10**9,
- nb_threads=8, reps=100):
+ fancy_strategy='AS', sub_threshold=0.1, sup_threshold=10**9, nb_threads=8, reps=100):
""" Ant colony optimization of the cache blocking parameters for the execution of
the iso3dfd programm.
Args:
Me (int): index of process running the function
NbP (int): number of processes
- comm (): mpi communication object
+ comm (MPI object): mpi communication object
alpha (float): hyperparameter alpha of ACO
rho (float): evaporation rate between 0 and 1
Q (float): quantity of pheromones deposited by an ant on an edge
nb_ants (int): number of ants
tau_0 (float): initial quantity of pheromones on each edge
n_iter (int): number of cycles done for ACO
- n1_max (int, optional): Maximal first dimension of matrix. Defaults to 256.
- n2_max (int, optional): Maximal second dimension of matrix. Defaults to 256.
- n3_max (int, optional): Maximal third dimension of matrix. Defaults to 256.
+ n1_max (int, optional): Maximal first dimension of the problem. Defaults to 256.
+ n2_max (int, optional): Maximal second dimension of the problem. Defaults to 256.
+ n3_max (int, optional): Maximal third dimension of the problem. Defaults to 256.
fancy_strategy (string, optional): Strategy used to update tau. Defaults to 'AS'.
sub_threshold (float, optional): Inferior threshold for Min-Max strategy. Defaults to 0.1.
sup_threshold (float, optional): Superior threshold for Min-Max strategy. Defaults to 10**9.
@@ -34,7 +33,7 @@ def ACO(Me, NbP, comm, alpha, rho, Q, nb_ants, tau_0, n_iter, n1_max=256, n2_max
reps (int, optional): Max number of iteration before stopping the process. Defaults to 100.
Returns:
- (np.array, float): optimal path [cbx, cby, cbz] and the associated cost
+ (tuple: list, float): optimal path [n1, n2, n3, cbx, cby, cbz] and the associated cost
"""
# Initialisation of the graph and pheromon matrix
diff --git a/Appli-iso3dfd/compute_path.py b/Appli-iso3dfd/compute_path.py
index 820880d44ef5b9407c271aa06c91c33e9cc5728f..4ad791b76ce9e79ab2fde9f70e192ee51dc32de3 100644
--- a/Appli-iso3dfd/compute_path.py
+++ b/Appli-iso3dfd/compute_path.py
@@ -4,6 +4,24 @@ import numpy as np
#tau : pheromone matrix, size : 1 + n1_size + n2_size + n3_size + n1_max//16 + n2_max x 1 + n1_size + n2_size + n3_size + n1//16 + n2 + n3
def proba(i, alpha, tau, n1_size, n2_size, n3_size, n_cbx, n_cby, n_cbz):
+ """Based on i, deduce which parameter you want to choose from.
+ Returns possibles choices and associated weights.
+
+ Args:
+ i (int): Index in matrix
+ alpha (float): Parameter to compute weights
+ tau (np.array): Pheromon matrix
+ n1_size (int): First dimension of the problem
+ n2_size (int): Second dimension of the problem
+ n3_size (int): Third dimension of the problem
+ n_cbx (int): First dimension of the cache
+ n_cby (int): Second dimension of the cache
+ n_cbz (int): Third dimension of the cache
+
+ Returns:
+ (tuple: np.array, np.array): possibles choices (sequence) and their weights (weights)
+ """
+
n1_max = (2**(n1_size-1)) * 16
n2_max = (2**(n2_size-1)) * 256
n3_max = (2**(n3_size-1)) * 256
@@ -46,6 +64,18 @@ def proba(i, alpha, tau, n1_size, n2_size, n3_size, n_cbx, n_cby, n_cbz):
return (sequence, weights)
def compute_path(tau, alpha, n1_size, n2_size, n3_size):
+ """From the sizes of the problem and the pheromon matrix, give a ant path.
+
+ Args:
+ tau (np.array): Pheromon matrix
+ alpha (foat): Parameter impacting weights when choosing path
+ n1_size (int): First dimension of the problem
+ n2_size (int): Second dimension of the problem
+ n3_size (int): Third dimension of the problem
+
+ Returns:
+ (list[int]): path choosen by the ant (n1, n2, n3, cbx, cby, cbz)
+ """
#n_cbx, n_cby, n_cbz will be initialized after the choice of n1, n2, n3
n_cbx = 0
diff --git a/Appli-iso3dfd/launcher_SUBP.py b/Appli-iso3dfd/launcher_SUBP.py
index 3ea9f3c286e8840c3ca7f0bea2401b80dcef9b90..2e0ba4602033c732e8e0441b7f4a85b26bda4b50 100644
--- a/Appli-iso3dfd/launcher_SUBP.py
+++ b/Appli-iso3dfd/launcher_SUBP.py
@@ -13,17 +13,17 @@ def deploySUBP(n1, n2, n3, nb_threads, reps, cbx, cby, cbz):
"""Launch MPI execution based on exe file in bin/, and returns average fitness
Args:
- n1 (int): First dimension of matrix.
- n2 (int): Second dimension of matrix.
- n3 (int): Third dimension of matrix.
- nb_threads (int): Number of threads per MPI process.
- reps (int): Max number of iteration before stopping the process.
- cbx (int): First cache blocking matrix dimension.
- cby (int): Second cache blocking matrix dimension.
- cbz (int): Third cache blocking matrix dimension.
+ n1 (int): First dimension of the problem
+ n2 (int): Second dimension of the problem
+ n3 (int): Third dimension of the problem
+ nb_threads (int): Number of threads per MPI process
+ reps (int): Max number of iteration before stopping the process
+ cbx (int): First dimension of the cache
+ cby (int): Second dimension of the cache
+ cbz (int): Third dimension of the cache
Returns:
- float: calculated fitness (average of MPI processes executions caracteristic)
+ float: calculated fitness (average of MPI processes executions throughputs)
"""
# Get the name of the exe file
diff --git a/Appli-iso3dfd/tools.py b/Appli-iso3dfd/tools.py
index 19ebf8dba9ebfdaa878c06edf34a513ed202e2e4..587d155a2d5ce0cc36bf3f8d28ddcdf06ff31e89 100644
--- a/Appli-iso3dfd/tools.py
+++ b/Appli-iso3dfd/tools.py
@@ -100,14 +100,14 @@ def add_pheromones(tau, paths, costs, Q, n1_size, n2_size, n3_size, n_cbx, n_cby
paths (np.array): Array of the paths taken by the ants: paths[i]=(cb_x//16,cb_y,cb_z)
costs (np.array): Column array containing the costs.
Q (float): quantity of pheromones added by an ant on an edge
- n1_size (int): First dimension of matrix
- n2_size (int): Second dimension of matrix
- n3_size (int): Third dimension of matrix
+ n1_size (int): First dimension of the problem
+ n2_size (int): Second dimension of the problem
+ n3_size (int): Third dimension of problem
n_cbx (int): First dimension of the cache
n_cby (int): Second dimension of the cache
Returns:
- (np.array, list, float): the updated pheromon matrix, the best path and the associated cost
+ (np.array): the updated pheromon matrix
"""
for i in range(len(costs)):