Pressure jumps between two runs

Dear Sir,

I was running a test case and found that the pressure value in log file will jump between two identical runs, even though no conditions are changed for the system. For example, as shown below, “run 600” then “run 500” without changing any other parameters gave the log information like this:

600 80.011145 -73218.359 37.480322 30.646216 6.8524014 -26463.228 -26425.747 -27231.198 1.9494951 -0.0002600758 0.00010282679 0.843356
Loop time of 29.4407 on 1 procs for 500 steps with 3625 atoms

Pair time () = 29.3194 (99.5879) Bond time () = 0.000192404 (0.00065353)

Neighbor list builds = 1
Dangerous builds = 0

run 500 Memory usage per processor = 7.82065 Mbytes
Step Temp Press KinEng kexy kez PotEng TotEng Enthalpy 3 1[1] 1[2] 1[3]
600 80.011145 -57500.261 37.480322 30.646216 6.8524014 -26463.228 -26425.747 -27231.198 1.9494951 -0.0002600758 0.00010282679 0.843356

The pressure value is labeled in bold black and it jumped, but all other parameters are constant. I don’t know why this happened, is it possible that simulation box dimensions changed somehow?

Best Regards,
Z. Q

Without knowing what fixes you have applied and exactly what happened
between these two runs, no one can tell. If you suspect the volume
has changed, why not write it out to make sure?

If your setup, structure and dynamics were all good and the system
conserved energy, since H=U+PV, and H and U stayed exactly the same,
then the pressure change must be accompanied by a volume change,


If you are using shrink-wrap BC, then the volume
may change between runs, due to re-shrink-wrapping.
That only happens when neighbor lists are rebuilt.
As Ray said, it's easy to monitor the volume yourself
in the thermo output.


Dear Steve and Ray,

Thanks for your explanation and sorry for my late reply. I monitored the volume and you are right, the pressure jump is due to the change of volume which originated from the shrink-wrap BC.

Best Regards,
Z. Q.