[CP2K-user] SCF convergence in md slowing down

[Daniele Fontanari]

Hello, I notice an issue when I perform an MD run on a large system.

I have that the first steps are fast but they get slower and slower (from 
the timing in the .ener file), in the output I notice that the SCF run take 
more and more steps to converge.
If I stop my calculation and restart it from the restart input file the 
times go down again.

I added an extract from my ener file:

md calculation starts at step 2036 (restart from an existing WFN file), 
first step take a little more than the subsequents ones. Until 2332 we see 
the time increase, then I stop and restart CP2K and times go down again

if I grep "SCF run" from the output you have the results in grep_out, and 
the whole output in raw_out. in md.inp and step.xyz the necessary files for 
running the md simulation (minus the WFN file) (If you prefer I can attach 
the restart file, but it is less readable)

Does anyone see something wrong in my input file or know the issue?

Cheers

Daniele
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.cp2k.org/archives/cp2k-user/attachments/20181121/26a824d7/attachment.htm>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: extract.ener
Type: application/octet-stream
Size: 40086 bytes
Desc: not available
URL: <https://lists.cp2k.org/archives/cp2k-user/attachments/20181121/26a824d7/attachment.obj>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: grep_out
Type: application/octet-stream
Size: 12814 bytes
Desc: not available
URL: <https://lists.cp2k.org/archives/cp2k-user/attachments/20181121/26a824d7/attachment-0001.obj>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: md.inp
Type: chemical/x-gamess-input
Size: 1447 bytes
Desc: not available
URL: <https://lists.cp2k.org/archives/cp2k-user/attachments/20181121/26a824d7/attachment.inp>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: step.xyz
Type: chemical/x-xyz
Size: 64399 bytes
Desc: not available
URL: <https://lists.cp2k.org/archives/cp2k-user/attachments/20181121/26a824d7/attachment.xyz>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: raw_out
Type: application/octet-stream
Size: 1550108 bytes
Desc: not available
URL: <https://lists.cp2k.org/archives/cp2k-user/attachments/20181121/26a824d7/attachment-0002.obj>