On Wed, May 2, 2018 at 4:37 PM, Alexander Alexander
<[email protected]> wrote:
> Thanks.
>
> On Wednesday, May 2, 2018 at 5:54:50 PM UTC-4, Christoph Junghans wrote:
>>
>> On Wed, May 2, 2018 at 3:37 PM, Alexander Alexander
>> <[email protected]> wrote:
>> > Hi Christoph,
>> > Thanks.
>> >
>> > On Wednesday, May 2, 2018 at 5:13:13 PM UTC-4, Christoph Junghans wrote:
>> >>
>> >> On Wed, May 2, 2018 at 1:38 PM, Alexander Alexander
>> >> <[email protected]> wrote:
>> >> > Dear all,
>> >> > I did a
>> >> > My IBI-bonded-nonbonded (after getting the nonbonded potentials via
>> >> > IBI-nonbonded only) crashes dues to the "...bonded interactions could
>> >> > not be
>> >> > calculated because some atoms
>> >> > involved moved further apart ...". By replacing the ill-bond by a
>> >> > harmonic
>> >> > potential in topol.top file "2 3   N  0.3453  K" where N = 1,2 and K
>> >> > =
>> >> > 5000,
>> >> > the simulation are going well. Now what? what is the relation between
>> >> > this
>> >> > harmonic  and the table_b2.xvg which was replaced by the harmonic
>> >> > one?
>> >> > What
>> >> > would happen for the table_b2.xvg?
>> >> Sorry, I don't understand your question! Can you rephrase?
>> >>
>> > Let's say the calculation crashes when table_b2.xvg is used but it works
>> > fine with "2 3   2  0.3453  1000" instead, then my question is that
>> > "should
>> > I continue the preregistration using "2 3   2  0.3453  1000" and forget
>> > about the table_b2.xvg? If so, what would be the final right
>> > table_b2.xvg at
>> > the end? If not so, how I can take benefit from the knowledge of  "2 3
>> > 2
>> > 0.3453  1000" to improve the table_b2.xvg so that it would work as well?
>> You know generate a table with k*(r-r_0) (with r_0=0.3453 andk=1000,
>> clip out the middle part and replace it with the structured part from
>> BI.
>>
> OK, I see.
> How about this instead:
> The reason that the calculation crashes with table_b2.xvg is that the
> table_b2.xvg represents a very low K (spring constant) so that the two atoms
> are getting apart. I was reading somewhere in the mailing list that the last
> number in "2 3   8  2  1  ; 1:bond-H1-L:1" is a kind of factor for the third
> column of the table_b2.xvg, so, don't you think it would be possible to
> increase this factor to 2 or 3 (which means multiplying the third column)
> until the calculation goes fine?
Yeah, that sounds reasonable! Increasing the K helps usually helps
with the atoms getting too far apart.

If you want to scale column 3 you will also have to modify column 2 accordingly.
Gromacs won't allow mismatching columns.
>>
>>
>> >
>> >>
>> >> > Figures in attachment are the plot of
>> >> > table_b2.xvg.
>> >> > I got the bonded tables including the table_b2.xvg by "csg_boltzmann"
>> >> > -->
>> >> > "csg_call --sloppy-tables" ---> "csg_call --ia-type bond --ia-name
>> >> > bond
>> >> > ...
>> >> > gromacs table_b2.xvg". And I am using the bond(angle)-*.pot.in as
>> >> > initial
>> >> > guesses for my IBI-all.
>> >> >
>> >> > The table_b1.xvg in gromacs manual has been defined such that the the
>> >> > r(nm),
>> >> > f(r), f'(r) are the first, second and third column of table_b2.xvg
>> >> > where
>> >> > f(r) is a cubic spline function in V_b(r_ij) = k*f(r_ij), If the same
>> >> > definition is applied for the table_b2.xvg obtained by the above
>> >> > method?
>> >> To be clear, the gromacs table format is x, f(x), -f'(x) (see gromacs
>> >> manual 4.2.14)
>> >>
>> >> I am not sure what your question is! Gromacs uses cubic spline
>> >> internal to interpolate  the table.
>> >> However, if you set cg.inverse.gromacs.table_bins in your xml file
>> >> small enough (you had something like 0.002), then there isn't much to
>> >> interpolate.
>> >>
>> >>
>> >> > What is the unit of f(r)?
>> >> KJ/mol (gromacs energy units)
>> >>
>> >> > What is the unit of each column in table_b2.xvg in VOTCA?
>> >> KJ/mol/nm (gromacs force units)
>> >
>> > Just to sum up the table format, can you please confirm me that the
>> > table_b.xvg in VOTCA has the following format:
>> > first column : Distance with unit nm
>> > Second column : Energy with unit KJ/mol
>> > Thirst column: Force with unit KJ/mol/nm
>> For bonds (not angles) in gromacs that is the format!
>
> For angle the first column is degree probably.
>>
>>
>> Also for a different simulation backend, e.g. lammps this will be
>> different of course, too!
>
>
> I have one more question which your comment on would be highly appreciated:
> In another IBI-bonded-nonbonded parameterization the "mpirun -np 64 gmx_mpi
> mdrun" work fine till step_021 of parameterization where it crashes because
> it can not find any domain decomposition for 64 ranks that is compatible
> with the system size. Even it works in the minimization simulation in
> step_021.
> I have already read most of the discussions about the error of domain
> decomposition, and I have tried different options like using different ranks
> (64, 32, 16, 8, 40, 36) and also -rdd ... or -dds. However, no success yet.
> I am totally confused with these parameters and I can not solve the issue.
> So, below I have shared the md.log files for the step_20 (which works fine)
> and step_21 (minimization which works fine) and step_21 (which crashes), and
> I would be so appreciated if one could help me overcome the issue.
> About the machine: Running on 4 nodes with total 64 cores, 64 logical cores,
> Cores per node: 16.
>
> md.log-20
> https://drive.google.com/open?id=1ZwYggDI1Hy_eYY8_nT8I50t91Wla0Obr
>
> md.log-21-mini
> https://drive.google.com/open?id=1oYE4A6fC1L8u_vx4sR8RIq3EWiG9-7YB
>
> md.log-21
> https://drive.google.com/open?id=1ff_QJrdGvLFgwGU9xMY-IcqfxzF2jkAO
Sorry, that really is a question for the gromacs user mailing list.
([email protected])

Christoph

>
>
> Thank you very much.
> Regards,
> Alex
>>
>>
>> Christoph
>> >>
>> >>
>> >> >
>> >> > And would you please explain table_b2-1-3.png figure?
>> >> Not sure what there is to explain..the code on how the third column is
>> >> calculated is here:
>> >>
>> >>
>> >> https://github.com/votca/csg/blob/master/share/scripts/inverse/table_to_xvg.pl#L95
>> >
>> > Best regards,
>> > Alex
>> >>
>> >>
>> >> Christoph
>> >>
>> >> > Thank you very much.
>> >> > Regards,
>> >> > Alex
>> >> >
>> >> > --
>> >> > You received this message because you are subscribed to the Google
>> >> > Groups
>> >> > "votca" group.
>> >> > To unsubscribe from this group and stop receiving emails from it,
>> >> > send
>> >> > an
>> >> > email to [email protected].
>> >> > To post to this group, send email to [email protected].
>> >> > Visit this group at https://groups.google.com/group/votca.
>> >> > For more options, visit https://groups.google.com/d/optout.
>> >>
>> >>
>> >>
>> >> --
>> >> Christoph Junghans
>> >> Web: http://www.compphys.de
>> >
>> > --
>> > You received this message because you are subscribed to the Google
>> > Groups
>> > "votca" group.
>> > To unsubscribe from this group and stop receiving emails from it, send
>> > an
>> > email to [email protected].
>> > To post to this group, send email to [email protected].
>> > Visit this group at https://groups.google.com/group/votca.
>> > For more options, visit https://groups.google.com/d/optout.
>>
>>
>>
>> --
>> Christoph Junghans
>> Web: http://www.compphys.de
>
> --
> You received this message because you are subscribed to the Google Groups
> "votca" group.
> To unsubscribe from this group and stop receiving emails from it, send an
> email to [email protected].
> To post to this group, send email to [email protected].
> Visit this group at https://groups.google.com/group/votca.
> For more options, visit https://groups.google.com/d/optout.



-- 
Christoph Junghans
Web: http://www.compphys.de

-- 
You received this message because you are subscribed to the Google Groups 
"votca" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to [email protected].
To post to this group, send email to [email protected].
Visit this group at https://groups.google.com/group/votca.
For more options, visit https://groups.google.com/d/optout.

Reply via email to