[gmx-users] Combining .edr-files with eneconv

[Gmx QA]

Dear users

I have a simulation of a pure popc bilayer for which I am trying to
calculate the area-per-headgroup for the lipids, simply by using (box_x *
box_y) / num_lipids

My run is split into different parts, and I was looking at first combining
the .edr-files with eneconv and then using g_energy to extract the
box-vectors.

However, when I enter the following:
$ ~/gromacs-4.6.5/build/src/tools/eneconv_mpi -f prod.part000*edr -o
prod.edr

I first get this:

Opened prod.part0001.edr as single precision energy file
Reading energy frame      1 time   20.000
Opened prod.part0002.edr as single precision energy file
Energy files don't match, different number of energies:
 prod.part0001.edr: 52
 prod.part0002.edr: 52

Continue conversion using only the first 52 terms (n/y)?

This is puzzling, since a "manual" inspection with g_energy of each of
these two files show the exact same 52 energy terms (and indeed 52==52 :-)

Then web I press y, the following pops up:

Opened prod.part0002.edr as single precision energy file
Opened prod.part0003.edr as single precision energy file
Energy files don't match, different number of energies:
 prod.part0002.edr: 52
 prod.part0003.edr: 48

Continue conversion using only the first 48 terms (n/y)?

which seems even more strange, since again, using g_energy, there is
exactly 52 identical energy terms in these two files.

Any ideas about what I am missing, and if should be concerned? Obviously I
can use g_energy on each .edr-file separately, and the concatenate the
.xvg-files afterwards, but using eneconv first seems more appropriate
somehow :-)

Thanks
/PK