[gmx-users] Issue with nrexcl (g_rdf related)

[João Henriques]

Dear all,

A little context first: I have written my own CG bead necklace model for
use with gromacs. Production simulations use nrexcl = 1.

One of the validation analysis involves computing the rdf, and for that I
need a .tpr with higher nrexcl, as suggested in the tool's help.

Interestingly, two different .tpr files compiled using two similar
topologies, differing only on the nrexcl, produce similar outputs when
checked with gmxdump (the init. velocities differ but that's normal,
because the seed is random).

Needless to say that rdf outputs are similar regardless of which .tpr I use
(they are the same!).

I double checked that the .top files differ uniquely on the nrexcl value.

Tested with gmx-4.6.7 and gmx-5.0.4.

P.S.: I use an harmonic potential for the bonds (func. type 6). Don't know
if this matters, but I have a feeling it does.

Thank you for your attention,
Best regards,
João Henriques