gt.minimize_blockmodel_dl with parallel MCMC sweep throws a segmentation fault (v. 2.17 to current)
Minimum (non)working example:
import graph_tool.all as gt
import numpy as np, numpy.random as npr
print(gt.__version__)
# Erdős–Rényi
N = 5000
p = 1.5 * np.log(N)/N
g = gt.random_graph(N, deg_sampler=lambda: npr.poisson((N-1)*p), directed=False, model='erdos')
# Segmentation fault:
pt = gt.minimize_blockmodel_dl(g, mcmc_args={'parallel':True}, mcmc_equilibrate_args={'verbose':False, 'epsilon':1e-5}, verbose=True)
Version: 2.22 (commit 44bf2b92, Thu Mar 2 23:08:39 2017 +0000)