[gmx-users] Eror: No parallel Ewald use PME instead!

[Justin Lemkul]

On 6/4/14, 8:29 AM, Ali Alizadeh wrote:
> Dear Mark,
>
> In my first simulation we had these lines:
>
> ; Electrostatics
> coulombtype    = PME        ; Particle Mesh Ewald for long-range
> electrostatics
> pme_order    = 4        ; cubic interpolation
> fourierspacing    = 0.12        ; grid spacing for FFT
>
> -----------------------
>
> ; Electrostatics
> coulombtype    = ewald        ; Particle Mesh Ewald for long-range
> electrostatics
> fourierspacing    = 0.6        ; grid spacing for FFT
>
> -----------------
>
> My system in 2nd case is equilibrated.
>

For anyone to make sense of what you're doing, you'll need to provide complete 
.mdp files (those settings are not the only determining factors of the validity 
of the model physics) and evidence of what you're deciding is equilibrated, 
stable, etc (i.e. real numbers and output from whatever relevant analysis).

-Justin

> On Wed, Jun 4, 2014 at 4:35 PM, Mark Abraham <mark.j.abraham at gmail.com>
> wrote:
>
>>
>>
>>
>> On Wed, Jun 4, 2014 at 12:42 PM, Ali Alizadeh <
>> ali.alizadehmojarad at gmail.com> wrote:
>>
>>> Dear Mark,
>>>
>>> Thank you for your reply.
>>>
>>> When you do not change any parameters except one thing in mdp file(PME to
>>>
>>> ewald as the paper said that)and then your system converges, for me it means the problem is related to
>>>
>>> choosing the right method for  calculation of the electrostatic potential energy. What's your point on it?
>>>
>>>
>> My point is that a flawed model physics, or a flawed system preparation,
>> can get lucky and run stably, or not. On the information given, there is no
>> reason to suppose that changing the kind of reciprocal-space approximation
>> is relevant. Last time I heard a report like this, the person had changed
>> various nonbonded settings "to make things run faster," and in so doing
>> produced a junk model physics.
>>
>> Mark
>>
>>
>>> -----
>>>
>>> Mark wrote:
>>>
>>> Hi,
>>>
>>> What makes you think the PME algorithm, rather than your choices of
>>> settings for it (and twenty other things), is the problem?
>>>
>>> Mark
>>>
>>>
>>> On Wed, Jun 4, 2014 at 7:57 AM, Ali Alizadeh <ali.alizadehmojarad at gmail.com <https://maillist.sys.kth.se/mailman/listinfo/gromacs.org_gmx-users>>
>>> wrote:
>>>
>>>> * Dear All users,
>>> *
>>>>> * I have encounter this error:
>>> *>>* -------------
>>> *>* Fatal error:
>>> *>* No parallel Ewald. Use PME instead.
>>> *>* --------------
>>> *>>* I used 8 cores and gromacs 4.5.5, for doing my calculations. I should use
>>> *>* ewald method because
>>> *>>* my system can not converge using PME it heats up my system. Are there any
>>> *>* suggestions?
>>> *>>>>* --
>>> *>* Sincerely
>>> *>>* Ali Alizadeh
>>> *>* --*
>>>
>>>
>>> --
>>> Sincerely
>>>
>>> Ali Alizadeh
>>>
>>
>>
>
>

-- 
==================================================

Justin A. Lemkul, Ph.D.
Ruth L. Kirschstein NRSA Postdoctoral Fellow

Department of Pharmaceutical Sciences
School of Pharmacy
Health Sciences Facility II, Room 601
University of Maryland, Baltimore
20 Penn St.
Baltimore, MD 21201

jalemkul at outerbanks.umaryland.edu | (410) 706-7441
http://mackerell.umaryland.edu/~jalemkul

==================================================