delete_atoms during the running

If I use “delete_atoms” in the “run 10000 every 1000” My calclation does not proceed any more after at a cretain specific step (e.g., 113322) while it works well without it. I’ve tested it several times but it showed the problem consistently at the same timestep.
The calulation was running in parallel computing regime and the periodic step of dump (save of the geometric configuration) was also set to be 1000.
I’ve also tested the use of “compress no” in delete_atoms.

In such cases, what would be the points suspected?
Does “delete_atoms” have some known bugs?

Thank you for your reply in advance.

Best
Young

What is the error you get ?

Arun