You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
As mentioned originally
"Multiscale run would really like to use dry-run. However, atm it is not possible because astrocytes lack an etype (I think)." @cattabiani
Also there seems to be issues with the original (CPU) load balance due to one of the mechanisms not featuring a default constructor. Quoting @jamesgking
"It is probably NEURON trying to instantiate a "stand alone" instance of mcd13.mod in order to evaluate the impact to computations. Typically mod files can be instantiated in isolation and run for a few time steps. In this case, mcd13 does require the extra components and so it still has a NULL pointer which it tries to dereference and crashes.
We have 2 options:
have some way to disable the computations that require dereferencing glu2; this means that mcd13 will look like it imposes no cost, so will not be factored into load balancing
have some default value which will cause mcd13 to do some typical workload
The text was updated successfully, but these errors were encountered:
As mentioned originally
"Multiscale run would really like to use dry-run. However, atm it is not possible because astrocytes lack an etype (I think)." @cattabiani
Also there seems to be issues with the original (CPU) load balance due to one of the mechanisms not featuring a default constructor. Quoting @jamesgking
"It is probably NEURON trying to instantiate a "stand alone" instance of mcd13.mod in order to evaluate the impact to computations. Typically mod files can be instantiated in isolation and run for a few time steps. In this case, mcd13 does require the extra components and so it still has a NULL pointer which it tries to dereference and crashes.
We have 2 options:
The text was updated successfully, but these errors were encountered: