Effective Mass Analyzer Hangs

[ams_nanolab]

When I'm using the effective mass analyzer, by dragging and dropping a NetCDF file in it, it seems to hang frequently. Say I'm performing calculations for a number of .nc files, after every 3rd or 4th file the analyzer hangs and vnl needs to be restarted again. I have tried with Auto Update disabled option but it still hangs after every 3rd-4th time.

I'm running ATK 12.8.0 on a RHEL server of good configuration, DFT calculations etc. finish quite well so I don't think teher's anything wrong with my hardware. The trouble is with the analyzer only. Is there a bug?

[Anders Blom]

So it's for several NC files, not several different k-points or other parameters in the Analyzer itself?

Any chance there is something particular about the NC files, like some calculation which is very large?

If you have the chance, please monitor the memory usage of VNL while you try this, maybe there is an accumulation and you run out of RAM?

[ams_nanolab]

Yes that happens sometimes to if I look to find out eff mass in different directions for the same K point, the analyzer hangs sometimes.
Thanks for pointing out, I'll check the memory usage.
Is there any way to clean the accumulated data before starting fresh calculation so I don't run out of RAM?

[Anders Blom]

In principle it should take care of it automatically, so I doubt it's necessarily a memory problem, but it's best to rule out all cases. It should not accumulate data, but if it does one can look at the design of the Analyzer perhaps.

I have myself used it extensively with many re-calculations with many k-points and never had any issues. It would be nice if you could identify a 123 workflow which always results in a hang, with full details (drop this file here, change parameter X to 4, etc, etc), otherwise it will be hard to troubleshoot.