Code for computing the covariance matrix decomposition shown in "Classical algorithm for simulating experimental Gaussian boson sampling".

• get_decomposition.py contains the original code used by Changhun Oh et al.. This file uses the 'CVXOPT' solver by default. The output of this code is the covariance matrix of an optimized pure state.
• opt_cov_example.py contains a slight modification of the original code. In particular, it shows an error when using 'CVXOPT' for large covariance matrices. It also has the option to change the solver to 'SCS'. The output of this code is the covariance matrix of an optimized pure state.
• cov_0.npy contains an array of shape (200, 200). This array corresponds to the covariance matrix of the ground truth of the Jiuzhang 1.0 experiment.
• requirements.txt contains all the requirements to run these files.

Note that the original files can be found in here.