Break compute_source_psd_epochs by setting n_jobs too high

mne_sample_crash_psd.py

#!/usr/bin/env python
import numpy as np
import mne
from mne.minimum_norm import read_inverse_operator, compute_source_psd_epochs
import os
import sys, traceback
import multiprocessing

# The rank of these data with 21 epochs for this label is 12. If n_jobs=12, the PSD will compute without error.
# It looks like n_jobs can't exceed the lowest rank
n_jobs = 13
n_cpu = multiprocessing.cpu_count()
print('You have %i CPUs.  Try not to use them all in one place.\n' % n_cpu)
subject = 'sample'
# I can never remember where the sample data are supposed to go, this was my workaround.
data_path = os.getenv('MNE_SAMPLE') + '/MEG/' + subject
subjects_dir = os.getenv('MNE_SAMPLE') + '/subjects'
method = "MNE"
snr = 1.0
lambda2 = 1.0 / snr ** 2
fmin, fmax = 4., 20.
bandwidth = 4.  # bandwidth of the windows in Hz

fname_raw = data_path + '/sample_audvis_raw.fif'
fname_event = data_path + '/sample_audvis_raw-eve.fif'
fname_inv =  data_path + '/sample_audvis-meg-oct-6-meg-inv.fif'

# Get inverse operator
inverse_operator = read_inverse_operator(fname_inv)  
# Load data
raw = mne.io.Raw(fname_raw)
# Read events
events = mne.read_events(fname_event)
# Set up pick list
include = []
picks = mne.pick_types(raw.info, meg=True, eeg=False, stim=True, eog=True,
               exclude=raw.info['bads'])
src = inverse_operator['src']  
# Get labels for FreeSurfer a2009s cortical parcellation
labels = mne.read_annot(subject, parc='aparc.a2009s', hemi='both',
                                        subjects_dir=subjects_dir)
lh_labels = [label for label in labels if label.hemi.endswith('lh')]
lh_names = [label.name for label in lh_labels]
   

try:      
    event_id, tmin, tmax = 1, -0.1, 1.5
    epochs = mne.Epochs(raw, events, event_id, tmin, tmax, proj=True,
            picks=picks, baseline=(None, 0), preload=True, reject=None)
    dropped_epochs=np.arange(21,72)
    epochs.drop_epochs(dropped_epochs)
    # Pick G_front_inf-Orbital-lh instead of enumerating through all 75...
    label = lh_labels[13] 
    print 'Read label %s' % (label.name)           
    n_epochs_on, n_chan, n_times = epochs._data.shape
       # Read epochs
    print 'Computing PSD for %s onset, event %s in label %s' % ( subject, event_id, label.name)
    # This will also fail if return_generator=False
    stcs = compute_source_psd_epochs(epochs, inverse_operator, lambda2=lambda2,
         method=method, fmin=fmin, fmax=fmax, bandwidth=bandwidth, adaptive=True,
         label=label, n_jobs=n_jobs, return_generator=True)
        
    for i, stc in enumerate(stcs):
       if i >= n_epochs_on:
           break
       elif i == 0:
           psd_avg_on = np.mean(stc.data, axis=0)
           print i
       else:
           psd_avg_on += np.mean(stc.data, axis=0)
           print i
      
    psd_avg_on /= n_epochs_on

except:
    print('\nThat didn\'t go well -- problem with %s onset, check debug output below.\n' % subject)
    traceback.print_exc(file=sys.stdout)