That has nothing to do with the normal mode calculation per se. The problem is the optimizer. For a large system, it is very hard to locally optimize the structures enough to be able to trust the low frequency vibrations.
Sent from my iPad
> On Apr 3, 2020, at 19:43, Liao <liaojunzhuo.aliyun.com> wrote:
>
> [External Email]
>
> Dear Amber users,
>
> I’ve been running into a persistent problem that I have never really resolved. That when using the MMPBSA.py.MPI script for energy calculations, specifically the nmode entropy calculations part, it very easily gets stuck.
>
> With drms=0.002, maxcyc=1000, I can see that in the nm.out file that gets stuck (usually stuck already by 200ns), the RMSG will go back and forth between a certain value, and never reach the drms value set. Whether maxcyc=1000 or 10000 doesn’t help, it usually gets stuck by iteration 200 or so.
>
> If all frames are good, they finish in an hour or so. When a single frame has this problem, after a few hours, it finishes with an error, but it still saves a result file, minus one (stuck) entropy frame. But there are times when more than one frame is stuck and nothing gets saved after a few hours.
>
> Have others been also getting this problem, frequently? If so, is it the structure problem, or a calculation set up problem?
>
> Many thanks.
>
> _______________________________________________
> AMBER mailing list
> AMBER.ambermd.org
> https://urldefense.proofpoint.com/v2/url?u=http-3A__lists.ambermd.org_mailman_listinfo_amber&d=DwIGaQ&c=sJ6xIWYx-zLMB3EPkvcnVg&r=dl7Zd5Rzbdvo14I2ndQf4w&m=jjXougxnRokDID3VyEKVh1n7VuqQy0QVAXSCvUAMakU&s=Ijt220kQgdjtFiFHs21o0WEXObQb7Sx-t6ZjyvhikaI&e=
_______________________________________________
AMBER mailing list
AMBER.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber
Received on Fri Apr 03 2020 - 17:00:02 PDT
