Flowitems seems to disappear at the end of simulation run

I try to give someone support and I hove now idea what is going on.

There is a large model (38 MB) with lots of objects, flowitems, tasksequences, messages, etc.
If we do a simulation run for 40 days, which takes a little bit more than an hour, all the items in the model are disappeared and all the NetworkNodes are orange (node edge closed).
I checked in the tree and the items (all items) are really gone.
There is a Source and a Queue in the model where just one item is generated and put into this queue. This is for testing if the items disappear.
This is now tested on three different computers and everywhere the same effect.
I checked at 35 days and every thing seemed to be right. But at the and (40 days) the items have gone.

So I just wonder, if anyone else had similar issues and have an idea or hints what is going on.

Sorry, but there is unfortunately no way to post this model at the community.


P.S.: This model also has the graphic issue we had with v4.30 if a properties window is open. If I close the properties window it seems to work. I thought this is fixed with v4.32 and was a graphic memory problem (leaking). Are there anyone else having still this problem?

Tom,
Yeah, I think that there's some piece of code that goes through the model at the end of an experimenter's replication, and deletes out all the flowitems, as if in preparation for the next replication. I'll put it on the development list to see if we can fix this for the next release.
Also, we have seen this issue with the Properties window. I believe we did fix a bug in the previous release, but this is another, similar bug, although it doesn't happen as frequently as the last one.

Anthony,

Thanks for your reply and that you put some stuff in the development list.
I did some more testing and spoke with the user again and you are right, he is using the Experimenter to have a warm up-period, but only makes one replication in the moment.

What I suggest is than to not use the StopTime set by the Experimenter, but make your own UserEventStoptime a little bit earlier with the following code or set the StopTime manually to zero again (none):

// UPDATE ALL STATES
updatestates();
// SCREEN REFRESH
repaintall();
// SHOW SCREEN MESSAGE
msg("UserEventStop",concat("Stop simulation run at time: ",numtostring(time(),0,4),".")); stop();


This is in my eyes working and should do the job.
I think, you could also make the warm up - period by a UserEvent where you just use the command resetstats().
But I am not sure what else will be done by the Experimenter at the end of the warm up-period.

Anyway, the problem seems to be solved and you gave some good hints.

Thanks a lot.
tom the (A)tom

P.S.: I installed the newest Nvidia-Driver (175.16) on the notebooks where I had the graphic issue and it seems to be gone at least with this model. So maybe there is still something, but using the latest Driver will help.