[CP2K-user] [CP2K:18283] DFTD3 / DFT + D slowdown

[Dobromir A Kalchevski]

Hello,

I had some of my optimizations run unreasonably slowly compared to others 
and I found the problem: it is in the DFTD3 / DFT + T. Similar number of 
atoms, all other settings the same. It doesn't matter whether 
diagonalization or OT is being used. Every other setting changes nothing. 
Resetting DFTD3 fixes the problem. It is easy to spot when in the beginning 
of every optimization step, just before the SCF cycle the program uses only 
1 cpu and gives no output for a long time. It happens only with some 
systems.

One way to fix it to reduce R_CUTOFF. There will be a value below which the 
calculation time jumps down back to normal, instead of gradually getting 
smaller.

Another way that sometimes works is setting CALCULATE_C9_TERM F.

I am using cp2k 2022.1

Regards,
Dobromir

-- 
You received this message because you are subscribed to the Google Groups "cp2k" group.
To unsubscribe from this group and stop receiving emails from it, send an email to cp2k+unsubscribe at googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/cp2k/b19d4c47-4852-4377-a375-744892328c53n%40googlegroups.com.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.cp2k.org/archives/cp2k-user/attachments/20230103/3ba5ed10/attachment.htm>