Re: [AMBER] ERROR: NaNs found in input coordinates.

From: Daniel Roe <daniel.r.roe.gmail.com>
Date: Thu, 3 Oct 2019 08:37:58 -0400

Hi,

On Wed, Oct 2, 2019 at 8:13 PM HOCHEOL LIM <ihcdatabank.gmail.com> wrote:
> 2. There is no error message in heat.out and min.out. And There are output

That's not true - look at the first step of your output:
```
| TI region 1


 NSTEP = 0 TIME(PS) = 0.000 TEMP(K) = 74.01 PRESS = 0.0
 Etot = ************** EKtot = 3651.3528 EPtot = **************
 BOND = 1861.9838 ANGLE = 1064.6312 DIHED = 1233.2188
 1-4 NB = 527.8747 1-4 EEL = 4546.6513 VDWAALS = **************
 EELEC = -48885.9027 EHBOND = 0.0000 RESTRAINT = 12.0262
 EAMBER (non-restraint) = **************
 DV/DL = -780.8373
```
Asterisks indicate numerical overflow (i.e. the number was too large
to print for the given width). The overflow occurs in the van der
Waals part of your energy which usually indicates bad atomic overlaps
leading to catastrophic forces, which in turn leads to the NaNs. You
need to check your starting coordinates for bad atomic overlaps etc.

-Dan

 min.rst7
> files from heat and min steps,
> - heat.en / heat.info / heat.nc / heat.out / heat.rst7
> - min.en / min.info / min.out / min.rst7
>
> I attached 0.0.zip file for output files.
>
> 3. This is the File Assignments in ti001.out,
> File Assignments:
> | MDIN: ti.in
>
> | MDOUT: ti001.out
>
> | INPCRD: heat.rst7
>
> | PARM: ti.parm7
>
> | RESTRT: ti001.rst7
>
> | REFC: refc
>
> | MDVEL: mdvel
>
> | MDEN: ti001.en
>
> | MDCRD: ti001.nc
>
> | MDINFO: ti001.info
>
> | MDFRC: mdfrc
>
>
> 4. I got the following error in ti001.out,
>
> ERROR: NaN(s) found in input coordinates.
> This likely means that something went wrong in the previous
> simulation.
>
> I think heat.rst7 must have had NaN(s) input coordinates.
>
> Looking forward to your feedback, thank you!
> _______________________________________________
> AMBER mailing list
> AMBER.ambermd.org
> http://lists.ambermd.org/mailman/listinfo/amber

_______________________________________________
AMBER mailing list
AMBER.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber
Received on Thu Oct 03 2019 - 06:00:04 PDT
Custom Search