`SpikeMonitor` recording from subgroups on weave can record spikes it shouldn't record #777

mstimberg opened this Issue Nov 15, 2016 · 0 comments


None yet
1 participant

mstimberg commented Nov 15, 2016

Argh, I just found another bug, very similar to #772. This time it is about SpikeMonitor and it also affects recording from subgroups. To trigger it you need a similar situation as for #772: record from a subgroup and have no neuron in that subgroup spike during that time step. However, you also need to:

  • use the weave target (in contrast to #772, Cython and C++ standalone are not affected)
  • record from a subgroup in the middle of a group

Also, the error should have been pretty obvious, except when only asking the SpikeMonitor for its total number of spikes: the erroneously recorded spikes have indices that are outside of the range of the subgroup. For example, with SpikeMonitor(group[100:200]) you should get indices in the range [0, 100[, but due to this bug you could record spikes with indices >=100.

The fix is trivial, I'll open a PR right away.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment