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
Orchestrator DPD is broken. Small sizes, 3x3x3 to 5x5x5 seem to work and manage to get to 10,000 timesteps, but anything larger fails after a few hundred timesteps.
This could be something to do with the changes to the asynchronous synchronisations, but I would assume this wouldn't work for any size.
Small volumes seem to be deterministic which is hopeful.
I'm going to start by counting messages in and out, as I am worried that some are getting lost, but if not this is going to turn into a real head-scratcher.
The text was updated successfully, but these errors were encountered:
Orchestrator DPD XML seemed to be fixed by pulling from the Orchestrator FEATURE-0167-etc branch.
It included a bug fix, and now everything seems to work quite nicely with bucket fill placement. I'm currently getting results for 3x3x3 to 40x40x40 on one box just to see how that goes.
Next is to try it with random placement. If that works then this can be closed. DPD XML is still way too large (1000+ bytes over), but short of taking away features it needs, I can't think of a way to slim it down much more.
Orchestrator DPD is broken. Small sizes, 3x3x3 to 5x5x5 seem to work and manage to get to 10,000 timesteps, but anything larger fails after a few hundred timesteps.
This could be something to do with the changes to the asynchronous synchronisations, but I would assume this wouldn't work for any size.
Small volumes seem to be deterministic which is hopeful.
I'm going to start by counting messages in and out, as I am worried that some are getting lost, but if not this is going to turn into a real head-scratcher.
The text was updated successfully, but these errors were encountered: