[Nek5000-users] A Restart Issue with NaNs

[nek5000-users at lists.mcs.anl.gov]

I will try adjusting the dt to a smaller value. 



Sorry for the confusion, let me try to clarify: 



By restart without difficulty, I mean that I had been running the simulation to obtain data, stopping it, then starting again from where I left off. So it hadn't been a problem. 



Now the issue. After a successfull restart to continue the simulation, I now have 5 new fld files progressing in 4 second intervals. These files are where I am having the issue. I only 

noticed it by trying to use the most recent file again for a restart. When it didn't work I tried the most recent after that and so on until only the "2nd fld file" worked. So somewhere during that time 

this happened and I'm not sure why.  



Let me try the dt and I'll post my result again. 



- Michael 


----- Original Message ----- 
From: nek5000-users at lists.mcs.anl.gov 
To: nek5000-users at lists.mcs.anl.gov 
Sent: Tuesday, November 30, 2010 8:43:53 PM GMT -06:00 US/Canada Central 
Subject: Re: [Nek5000-users] A Restart Issue with NaNs 

Hi Michael, One of my favorite tests at this point is to reduce dt to a ridiculously low value, e.g., 1.e-6 or so... At that point, inertia should conquer all and the solution should change very little... >From your earlier email, I had the impression that you had made several runs (for this case?) with restart w/o difficulty, but now it looks like you're indicating you get the problem for any of the restart files? Has something changed in the .usr file or .rea file? Paul On Tue, 30 Nov 2010, nek5000-users at lists.mcs.anl.gov wrote: > > > I don't believe it is corrupted. I min/max values given after the restart > files load looks good, and I also > > just confirmed it using glmax/glmin. > > > > Also, I have several restart files that I have tried, all resulting in the > same problem. I have also looked > > at them in visit and everything looks consistent from what I had before, > could it still be corrupted even then? > > > > I have 5   *.f  files in increments of 4 seconds. I terminated the run > to change something with 'echo -2 >ioinfo'.  > > Fld files 3-5 give the problem, while number 2 worked. > > > > > > -Michael > > > ----- Original Message ----- From: nek5000-users at lists.mcs.anl.gov To: > "Nekton User List" Sent: Tuesday, November > 30, 2010 7:21:09 PM GMT -06:00 US/Canada Central Subject: Re: [Nek5000-users] > A Restart Issue with NaNs > > Hi Micheal, How did this computation finish on the prior run? Is the data > corrupted? (i.e., are there nans in the file?) You might try setting nsteps > to zero and then checking via glmin/glmax or something to see if the restart > file is corrupted. Paul On Tue, 30 Nov 2010, nek5000-users at lists.mcs.anl.gov > wrote: > > > Hi, > > > > I am having an issue with a simulation in which I > restart using the restart options in the Rea file. At the very first time > step > > I get NaN's in the Helmholtz VelZ .  I have done several restarts > of the same simulation at earlier points to arrive at where I am at > > now, > however I get this upon trying to continue the simulation further.  > > > > > Any ideas on why this could be, or how to troubleshoot it?  I'm not sure > what I could do since this is showing up on the very > > first time step. > > > > > Thanks a lot, > > Michael > > > > _______________________________________________ Nek5000-users mailing > > > list Nek5000-users at lists.mcs.anl.gov > > > https://lists.mcs.anl.gov/mailman/listinfo/nek5000-users 
_______________________________________________ Nek5000-users mailing list Nek5000-users at lists.mcs.anl.gov https://lists.mcs.anl.gov/mailman/listinfo/nek5000-users
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.mcs.anl.gov/pipermail/nek5000-users/attachments/20101130/698db3b1/attachment.html>