Re: Drude minimization does not converge

From: Christian Jorgensen (chrisjorg_at_gmail.com)
Date: Fri Jan 13 2017 - 13:21:39 CST

A very long equilibration appears to allow me to run the system,

also, I am using 0.1 fs timestep as suggested for initial dynamics.

On Fri, Jan 13, 2017 at 5:26 PM, Christian Jorgensen <chrisjorg_at_gmail.com>
wrote:

> Hello, I already set it up twice and got the same error.
> I am aware membrane proteins can have randomised setup errors
> due to the initial distribution of lipids, which sometimes are not resolved
> by the minimizer algorithm.
>
> However, the non-Drude psf/pdb can run fine, with acceptable initial
> energies:
>
> ETITLE: TS BOND ANGLE DIHED
> IMPRP ELECT VDW BOUNDARY MISC
> KINETIC TOTAL TEMP POTENTIAL TOTAL3
> TEMPAVG PRESSURE GPRESSURE VOLUME
> PRESSAVG GPRESSAVG
>
>
> ENERGY: 0 3583.6527 16962.1087 11760.6333
> 325.3170 -162471.7959 5441.4535 0.0000 0.0000
> 41673.5090 -82725.1218 309.7646 -124398.6309
> -82428.7248 309.7646 267.2281 267.6599
> 829440.0000 267.2281 267.6599
>
>
>
> I will run a longer equilibration and then re-setup the Drude files, but I
> am convinced the problem lies with the positioning of the lone pairs.
>
>
> I will try to reproduce the error again, but as said I redid the setup two
> times already.
>
> On Fri, Jan 13, 2017 at 4:36 PM, <salehesam_at_gmail.com> wrote:
>
>> Christian,
>>
>> I experienced this once, however the issue for me was that two atoms
>> happen to have overlapped by random chance, causing the energy to reach
>> this somewhat infinite value.
>>
>> Upon reconstructing the system the error I received had disappeared.
>>
>> Have you tried reconstructing your system?
>>
>> Can you replicate the error?
>>
>> Best,
>> Saleh Al-Khalifa
>>
>> On Jan 13, 2017, at 12:19, Christian Jorgensen <chrisjorg_at_gmail.com>
>> wrote:
>>
>> Hello,
>>
>> I am running an energy minimization on a membrane protein
>> prepared using Drude Prepper.
>>
>>
>> I can run the normal psf/pdb prior to turning it into Drude psf/pdb.
>>
>> I am unable to run the first minimization with Drude on, as the energy at
>> step 0 is inf
>>
>> ETITLE: TS BOND ANGLE DIHED
>> IMPRP ELECT VDW BOUNDARY MISC
>> KINETIC TOTAL TEMP POTENTIAL TOTAL3
>> TEMPAVG PRESSURE GPRESSURE VOLUME
>> PRESSAVG GPRESSAVG DRUDECOM DRUDEBOND DRCOMAVG
>> DRBONDAVG
>>
>>
>> ENERGY: 0 6915.9918 23501.9638 34656.5055
>> 251.6123 9999999999.9999 31402.7078 0.0000 0.0000
>> 9999999999.9999 9999999999.9999 9999999999.9999 9999999999.9999
>> 9999999999.9999 9999999999.9999 9999999999.9999 9999999999.9999
>> 1000000.0000 9999999999.9999 9999999999.9999 9999999999.9999
>> 9999999999.9999 9999999999.9999 9999999999.9999
>>
>>
>>
>> MINIMIZER RESTARTING CONJUGATE GRADIENT ALGORITHM
>>
>> LINE MINIMIZER REDUCING GRADIENT FROM nan TO nan
>>
>> LINE MINIMIZER BRACKET: DX 0 0 DU -nan -nan DUDX -nan -nan -nan
>>
>>
>>
>> As said, this Drude psf/pdb structures come from an equilibrated
>> structure. I think the problem is with the placing of the Drude lone pairs.
>>
>> Has anyone experienced this?
>>
>>
>>
>>
>
>
> --
> ---------------------------
> Christian Jorgensen
> PhD student, King's College London
> ---------------------------
>

-- 
---------------------------
Christian Jorgensen
PhD student, King's College London
---------------------------

This archive was generated by hypermail 2.1.6 : Sun Dec 31 2017 - 23:20:59 CST