darkhistory.utilities.log_1_plus_x¶
-
darkhistory.utilities.
log_1_plus_x
(x)¶ Computes log(1+x) with greater floating point accuracy.
Unlike
scipy.special.log1p
, this can takefloat128
. However the performance is certainly slower. See [1] for details. If that trick does not work, the code reverts to a Taylor expansion.Parameters: - x : float or ndarray
The input value.
Returns: - ndarray
log(1+x).
Footnotes
[1] | David Goldberg. 1991. What every computer scientist should know about floating-point arithmetic. ACM Comput. Surv. 23, 1 (March 1991), 5-48. |