RANDOM_EXCHANGES

[GiovanniBussi]

@carlocamilloni I was checking this line: https://github.com/plumed/plumed2/blob/9b82f8fb52f7e1bd89b1938ce70f9254381af1b1/patches/gromacs-2016.4.diff/src/programs/mdrun/repl_ex.cpp#L1182

The code is like this:

   /* PLUMED */
    if(plumed_test_exchange_pattern>0) {
      for (i = 0; i < re->nrepl; i++)
      {
          re->ind[i] = i;
      }
    }
    /* END PLUMED */

    /* record which moves were made and accepted */
    for (i = 0; i < re->nrepl; i++)
    {
        re->nmoves[re->ind[i]][pind[i]] += 1;
        re->nmoves[pind[i]][re->ind[i]] += 1;
    }
    fflush(fplog); /* make sure we can see what the last exchange was */
}

Is it correct to reset the ind before accumulating statistics? Shouldn't it be the opposite?

Giovanni

[carlocamilloni]

I cannot remember, the code is quite old by now. I would say that yes, there was a good reason for this…

On 19 Feb 2018, at 20:21, Giovanni notifications@github.com wrote:

@carlocamilloni https://github.com/carlocamilloni I was checking this line: https://github.com/plumed/plumed2/blob/9b82f8fb52f7e1bd89b1938ce70f9254381af1b1/patches/gromacs-2016.4.diff/src/programs/mdrun/repl_ex.cpp#L1182 https://github.com/plumed/plumed2/blob/9b82f8fb52f7e1bd89b1938ce70f9254381af1b1/patches/gromacs-2016.4.diff/src/programs/mdrun/repl_ex.cpp#L1182 The code is like this:

/ PLUMED / if(plumed_test_exchange_pattern>0) { for (i = 0; i < re->nrepl; i++) { re->ind[i] = i; } } / END PLUMED /

/* record which moves were made and accepted */
for (i = 0; i < re->nrepl; i++)
{
    re->nmoves[re->ind[i]][pind[i]] += 1;
    re->nmoves[pind[i]][re->ind[i]] += 1;
}
fflush(fplog); /* make sure we can see what the last exchange was */

} Is it correct to reset the ind before accumulating statistics? Shouldn't it be the opposite?

Giovanni

— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub https://github.com/plumed/plumed2/issues/342, or mute the thread https://github.com/notifications/unsubscribe-auth/AEXl9lN63GzwvXvRY5K5-P2WUS0Q-nqIks5tWcm0gaJpZM4SLAQ9.

[carlocamilloni]

I just checked this while checking the new GMX2019 patch and I confirm that it works correctly as it is