[lammps-users] Version difference

Steve,

I noticed that the newest version yields different results comparing to previous ones (See below), although the difference is minor. Because this is just a small system, i dont know if this difference would amplify when shift to large system.

Step Temp E_pair E_mol TotEng Press Volume

15880 66.239506 -451.32423 0 -400.97515 81.827131
15890 63.135848 -448.96715 0 -400.97717 150.19172
15900 63.481911 -449.23083 0 -400.97781 137.62414
15910 66.826933 -451.78313 0 -400.98754 57.396634
15920 69.749395 -454.01083 0 -400.99386 -10.591064
15930 69.959519 -454.16781 0 -400.99112 -13.494106
15940 68.353431 -452.94465 0 -400.98875 28.693793
15950 66.528321 -451.55992 0 -400.9913 77.768713
15960 64.700493 -450.16865 0 -400.98938 124.85893
15970 63.364671 -449.14743 0 -400.98352 156.79905
15980 64.141659 -449.74307 0 -400.98858 139.59917
15990 66.238742 -451.33568 0 -400.98718 91.722843
16000 66.988816 -451.89658 0 -400.97795 69.972166
Loop time of 10.7741 on 1 procs for 8000 steps with 256 atoms

15880 66.900059 -451.83882 0 -400.98764 79.937352
15890 69.563042 -453.85793 0 -400.9826 12.075424
15900 70.805225 -454.80973 0 -400.99021 -28.396733
15910 70.275072 -454.40054 0 -400.984 -26.69403
15920 69.207093 -453.5942 0 -400.98944 -6.4789383
15930 68.577154 -453.11575 0 -400.98981 10.589783
15940 68.364931 -452.94568 0 -400.98104 22.815776
15950 67.439288 -452.25191 0 -400.99086 50.579879
15960 66.070249 -451.21231 0 -400.99188 85.705115
15970 65.710394 -450.93535 0 -400.98845 95.745507
15980 66.417233 -451.47112 0 -400.98694 79.171119
15990 66.923458 -451.85539 0 -400.98643 63.547875
16000 66.5768 -451.58891 0 -400.98345 62.708098
Loop time of 10.8358 on 1 procs for 8000 steps with 256 atoms

How old a version are you comparing to? The code changes
over time. If the first few 10,100,1000 steps are nearly identical
to before, then it is probably something small that changed, leading
to diverging trajectories, but statistically identical.

Steve