Limited bonded distribution for inner and outer table cutoff

[Saeed Norouzi]

Dear VOTCA Users,

I am trying to use ibi for bonded interaction (bond,angle, and dihedral). The problem is when I am using bond distribution we do not have data for some ranges (like for bond distribution in this tutorial), so during the iteration some bonds length is  bigger or smaller than the table outer or inner cutoff and I get error in LAMMPS. does anyone has idea what can I do for this error?  is there any option in VOTCA to replace the zeros in the probability for bonds with some small values so I can increase the table cutoff or I should do it manually?

 

I am using LAMMPS for my ibi and here is an error I  have currently.

Per MPI rank memory allocation (min/avg/max) = 7.672 | 7.674 | 7.676 Mbytes
Step Time TotEng PotEng Volume Density Temp Press
       0            0    963.12455   -1360.1188    151186.66     1.029793          600   -1094.7381
    2000           20    928.00314   -1457.0206    151186.66     1.029793    615.95537   -1217.8225
    4000           40    942.24742   -1472.8241    151186.66     1.029793    623.71551    917.70274
    6000           60    925.95597   -1483.2126    151186.66     1.029793    622.19103    3122.7127
    8000           80    921.17567   -1446.9757    151186.66     1.029793    611.59792     1398.679
   10000          100    972.68412   -1498.2288    151186.66     1.029793    638.13709    3228.6631
   12000          120    968.92421    -1504.366    151186.66     1.029793    638.75105    2570.6902
   14000          140    973.39445   -1546.9816    151186.66     1.029793    650.91141   -2645.2069
   16000          160    826.48486   -1573.7195    151186.66     1.029793    619.87591    164.50872
ERROR on proc 20: Bond length > table outer cutoff: type 3 length 4.4888905 (src/MOLECULE/bond_table.cpp:605)
Last command: run 50000

Best regards

Saeed

[Christoph Junghans]

On Mon, Jan 2, 2023 at 7:51 AM Saeed Norouzi <norou...@gmail.com> wrote:
>
> Dear VOTCA Users,
>
> I am trying to use ibi for bonded interaction (bond,angle, and dihedral). The problem is when I am using bond distribution we do not have data for some ranges (like for bond distribution in this tutorial), so during the iteration some bonds length is bigger or smaller than the table outer or inner cutoff and I get error in LAMMPS. does anyone has idea what can I do for this error? is there any option in VOTCA to replace the zeros in the probability for bonds with some small values so I can increase the table cutoff or I should do it manually?

There are interaction-specific properties inverse.lamps.table_begin
and inverse.lamps.table_end that you can set for each interaction that
can be smaller/ larger than VOTCA's min and max (the area where IBI is
performed) . VOTCA will then do an extrapolation for that extra area.

Christoph

>
> I am using LAMMPS for my ibi and here is an error I have currently.
>
> Per MPI rank memory allocation (min/avg/max) = 7.672 | 7.674 | 7.676 Mbytes
> Step Time TotEng PotEng Volume Density Temp Press
> 0 0 963.12455 -1360.1188 151186.66 1.029793 600 -1094.7381
> 2000 20 928.00314 -1457.0206 151186.66 1.029793 615.95537 -1217.8225
> 4000 40 942.24742 -1472.8241 151186.66 1.029793 623.71551 917.70274
> 6000 60 925.95597 -1483.2126 151186.66 1.029793 622.19103 3122.7127
> 8000 80 921.17567 -1446.9757 151186.66 1.029793 611.59792 1398.679
> 10000 100 972.68412 -1498.2288 151186.66 1.029793 638.13709 3228.6631
> 12000 120 968.92421 -1504.366 151186.66 1.029793 638.75105 2570.6902
> 14000 140 973.39445 -1546.9816 151186.66 1.029793 650.91141 -2645.2069
> 16000 160 826.48486 -1573.7195 151186.66 1.029793 619.87591 164.50872
> ERROR on proc 20: Bond length > table outer cutoff: type 3 length 4.4888905 (src/MOLECULE/bond_table.cpp:605)
> Last command: run 50000
>
>
> Best regards
> Saeed
>

> --
> Join us on Slack: https://join.slack.com/t/votca/signup
> ---
> 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 votca+un...@googlegroups.com.
> To view this discussion on the web visit https://groups.google.com/d/msgid/votca/de34c1df-ca77-466b-a29d-3af382daed5en%40googlegroups.com.



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

[Saeed Norouzi]

Dear Christoph,

Thank you so much for your help. I used the  <table_begin> and  <table_end> in my setting file, and now it works. Could I also do the extrapolation on my target distribution, so I do not have to do it for every iteration and remove this interaction-specific from the setting file? Also, How could I have other than linear extrapolation for my data?

Moreover, as you can see below, I have dihedral interaction in my settings file and I am trying to do ibi for this part also. but the problem is the VOTCA gives the potential file for dihedral in radian units not degree. I am using lammps and I need to have the table function in degree units, as VOTCA does for angle potential. Can you please help me out with this?

 <!-- Dihedral Part  -->
  <bonded>
    <!-- name of the interaction -->
    <name>dihedral.1</name>
    <min>-3.14159</min>
    <max>3.14159</max>
    <step>0.01</step>
    <inverse>
      <!-- target distribution -->
      <target>dihedral.1.dist.tgt</target>
   <lammps>
        <table>table_d1.xvg</table>
        <table_begin>-3.14159</table_begin>
        <table_end>3.14159</table_end>
   </lammps>
    </inverse>
  </bonded>

Best regards

Saeed

 

To view this discussion on the web visit https://groups.google.com/d/msgid/votca/CAHG27e5LqWxmX2BZL%2BfOYOCQ-MEQVHZon-c2jpu%3DkByywUZGAA%40mail.gmail.com.

[Christoph Junghans]

On Tue, Jan 3, 2023 at 9:47 AM Saeed Norouzi <norou...@gmail.com> wrote:
>
> Dear Christoph,
> Thank you so much for your help. I used the <table_begin> and <table_end> in my setting file, and now it works. Could I also do the extrapolation on my target distribution, so I do not have to do it for every iteration and remove this interaction-specific from the setting file?

That might work, but I am not sure! The problem with poorly sampled
regions in IBI is that they can lead to systematic errors in the
update.

> Also, How could I have other than linear extrapolation for my data?

Yes, there should be an option,
inverse.lammps.table_left_exprapolation for each interaction (and same
for right), options are constant, linear, quadratic, exponential and
sasha. See "csg_call potential extrapolate --help" and "csg_call table
extrapolate --help" for details.

> Moreover, as you can see below, I have dihedral interaction in my settings file and I am trying to do ibi for this part also. but the problem is the VOTCA gives the potential file for dihedral in radian units not degree. I am using lammps and I need to have the table function in degree units, as VOTCA does for angle potential. Can you please help me out with this?

Dihedrals are not fully implemented yet. But I think you could set
inverse.lammps.scale to 180/pi=57.295 for that interaction.

Christoph

> To view this discussion on the web visit https://groups.google.com/d/msgid/votca/CAJZjMJbERZcwt5CDFXznUw83UYwFiLGTLFjWeXm%2BpTm2mjQwzw%40mail.gmail.com.

[Saeed Norouzi]

Dear Christoph,

Thank you for your reply, I just realized that lammps could read dihedrals in both radian and degree. 

Also, I am trying to change the extrapolation type for bonded interaction. I am not sure how to select the extrapolation type for the setting file, and also change the avg points. Do I have to change it in the potential_extrapolate.sh script in the shared directory? Can I instead add the extrapolation type and number of avg points to the setting file ? If so, can you please help here?

<!-- Dihedral Part  -->
  <bonded>
    <!-- name of the interaction -->
    <name>dihedral.1</name>
    <min>-3.14159</min>
    <max>3.14159</max>
    <step>0.01</step>
    <inverse>
      <!-- target distribution -->
      <target>dihedral.1.dist.tgt</target>
   <lammps>
        <table>table_d1.xvg</table>
        <table_begin>-3.14159</table_begin>
        <table_end>3.14159</table_end>

       <table_left_extrapolation> exponential </table_left_extrapolation>

       <table_right_extrapolation> exponential </table_right_extrapolation>
   </lammps>
    </inverse>
  </bonded>

Best regards

Saeed 

To view this discussion on the web visit https://groups.google.com/d/msgid/votca/CAHG27e5-SFkgL8OBEkz6MM4uMD2_9EytwqjCHn07sLYRERefmQ%40mail.gmail.com.

[Christoph Junghans]

On Wed, Jan 4, 2023 at 5:36 AM Saeed Norouzi <norou...@gmail.com> wrote:
>
> Dear Christoph,
>

> Thank you for your reply, I just realized that lammps could read dihedrals in both radian and degree.
>
> Also, I am trying to change the extrapolation type for bonded interaction. I am not sure how to select the extrapolation type for the setting file, and also change the avg points. Do I have to change it in the potential_extrapolate.sh script in the shared directory? Can I instead add the extrapolation type and number of avg points to the setting file ? If so, can you please help here?

I don't think there is an option to change the avg points, but you
could add an option here:
https://github.com/votca/votca/blob/master/csg/share/scripts/inverse/potential_to_lammps.sh#L132
similar to what we do for the function.
However in
https://github.com/votca/votca/blob/master/csg/share/scripts/inverse/potential_extrapolate.sh#L36
we don't distinguish between avg points for left and right, so hence
there is only one option.
However it wouldn't be hard to implement that either.

> To view this discussion on the web visit https://groups.google.com/d/msgid/votca/CAJZjMJaXXNcVhb2CAzTPjFS%3DG4st_nWO%3DUQ_FnAudTj%2BrBaJ%3Dw%40mail.gmail.com.