September 17, 2020 at 7:54 am #565
I get the following error in one of my hydro setups on Arepo. Looks like there is some error in setting up proper geometry. Can anyone please suggest a way to sort this error and get the code running?
task=72 flags=1 0 0 0 (axb)*c = 1.17124e-05 task=72 pp0=172 pp1=-1 pp2=-3 pp3=152 p=173 IDs=(850214 0 0 850045 850044) pos_0=(836.5|256.75|288.25) pos_1=(2800|2800|28000) pos_2=(2800|26558.8|-5600) pos_3=(835.5|257|289.5) pos=(836.5|256.75|288.25) task=72 imageflags=(1 0 0 1 1) TERMINATE: ******!!!!!****** Code termination on task=72, function InTetra(), file /home/alankar/arepo-clcrush/arepo/src/mesh/voronoi/voronoi_3d.c, line 3136: strange zero count
September 27, 2020 at 8:31 pm #571
- This topic was modified 4 months, 1 week ago by Alankar Dutta.
Is this happening directly after startup?
If yes, check your cell positions, especially that no cell has a position < 0 or > BoxSize.
And (via scatter plot or so) check if the cell positions fill your simulation volume and there
is not some region entirely devoid of cells. Let me know if this solves the problem.September 28, 2020 at 7:48 am #583
This is hapenning right on start up.
I increased the box size to a value bit larger than the xmax, ymax and zmax values in the IC file. Now the code runs but there is no progress beyond
DOMAIN: Peano-Hilbert order done, took 0.052613 sec.
Arepo just keeps on running without any progress until the entire job execution time.
I wish to share the links to the IC, output dump, and param files but the forum doesn’t allow that. Please let me know if there is an alternative.September 28, 2020 at 8:50 am #586
But in all cases, I have ensured that boxsize is not sharter than the boxsize in the IC file. But I still get the same error.October 3, 2020 at 7:11 pm #590
could you send me ICs, Config.sh and param.txt directly via email (see contact) or upload it to a could storage that I can access? I can have a look at it.October 4, 2020 at 3:00 am #591
Thanks a lot Rainer. For some unknown reason when I pasted this link, the post doesn’t show up in the forum. I have shared you the link via e-mail.October 5, 2020 at 7:37 pm #592
It seems like there are two cells (i.e. two mesh generating points, pos0 and pos in error message) with identical coordinates. This is not allowed in the Voronoi-construction. Check your ICs for this.October 10, 2020 at 6:58 am #594
Yes, you were right. Turns out these identical points were due to some data truncation error. Datatype specification was making the points identical in the python code that I used to generate the IC. This has been corrected and the issue is fixed. However, now the code doesn’t proceed beyond the following message (you can get it in output-log.txt). The code just keeps on running. It has been more than 10 hours but the code doesn’t progress beyond this message.
DOMAIN: Peano-Hilbert order done, took 0.594324 sec.
Relevant files had been shared by email as the form doesn’t support posting the link.
- You must be logged in to reply to this topic.