Skip to content

L3fp_params

compute_tau(basis_matrix, precision_level='default')

Computes TAU parameter based on the average norm of the input basis matrix.

Parameters:

Name Type Description Default
basis_matrix ndarray

A 2D NumPy array of shape (n, m) representing a lattice basis, where each column is treated as a basis vector.

 required
precision_level str

Determines the scaling of TAU. high: favors higher TAU values for increased precision. TAU in range [30,80]. low: favors lower TAU values for faster computation. TAU in range [10,40]. default: balanced TAU scaling. TAU in range [20,60].

 'default'

Returns:

Type Description
int

A dynamically computed TAU value, scaled based on the average norm of the basis vectors and the selected precision level.
Source code in bkz/L3FP/L3fp_params.py 
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
def compute_tau(basis_matrix, precision_level="default"):
	"""Computes TAU parameter based on the average norm of the input basis matrix.

	Args:
	    basis_matrix (np.ndarray): A 2D NumPy array of shape (n, m) representing a lattice basis, where each column is treated as a basis vector.
	    precision_level (str): Determines the scaling of TAU. `high`: favors higher TAU values for increased precision. TAU in range [30,80]. `low`: favors lower TAU values for faster computation. TAU in range [10,40]. `default`: balanced TAU scaling. TAU in range [20,60].

	Returns:
	    (int): A dynamically computed TAU value, scaled based on the average norm of the basis vectors and the selected precision level.
	"""
	norms = np.linalg.norm(basis_matrix, axis=0)
	avg_norm = np.mean(norms)

	if precision_level == "high":
		return max(30, min(80, int(np.log2(avg_norm) * 2)))
	elif precision_level == "low":
		return max(10, min(40, int(np.log2(avg_norm))))
	else:  # default
		return max(20, min(60, int(np.log2(avg_norm) * 1.5)))

update_tau(basis_matrix, precision_level='default')

Calls compute_tau and updates the module-global TAU.

Parameters:

Name Type Description Default
basis_matrix ndarray

A 2D NumPy array of shape (n, m) representing a lattice basis, where each column is treated as a basis vector. precision_level (str): Determines the scaling of TAU.

 required
Source code in bkz/L3FP/L3fp_params.py 
12
13
14
15
16
17
18
19
20
def update_tau(basis_matrix, precision_level="default"):
	"""Calls compute_tau and updates the module-global TAU.

	Args:
	    basis_matrix (np.ndarray): A 2D NumPy array of shape (n, m) representing a lattice basis, where each column is treated as a basis vector.
		precision_level (str): Determines the scaling of TAU.
	"""
	global TAU
	TAU = compute_tau(basis_matrix, precision_level)