bug? of calculation for energies (lammps-14Oct12 versus lammps-11Nov12)

Dear all,

I compared energies calculation on lammps-14Oct12 and lammps-11Nov12.
There were large differences in pairwise energy and total energy.

I made two graphs:
figure (1)= pairwise energy "14Oct12" vs "11Nov12" .
http://twitpic.com/bc0gx3
figure (2)= total energy: "14Oct12" vs "11Nov12".
http://twitpic.com/bc0hb4

Do these show a bug in computations of energies?

I attach a lammps script and a data file for this calculations and results.

regards
chibaf (Fumihiro CHIBA)

in.fcc111 (1.04 KB)

data.fcc111 (79.3 KB)

log.14oct12 (6.89 KB)

log.11nov12 (6.89 KB)

figure 2 2012-11-10 at 22.58.45 PM.png

figure 1 2012-11-10 at 22.58.40 PM.png

we'll take a look - thanks

Steve

There will be a patch for this later today. The bug
fix was for an uninitialized variable. But the variable
was not directly related to your input script, so please
try the new version and verify that your problem goes
away (we can't reproduce it).

Steve