I tried to use the slope(x) special operation in LAMMPS to get Cp from dU/dT, but I got an odd numerical result. To understand the slope command I made the following trial: First I create a 1row vector with the simulation step as element, selected every 1 step. Then, use the slope(x) command to get the slope of the vector values. The vector is apparently properly constructed, and ave(x) or sum(x) give me the expected value. But slope(x) doesn’t give 1. Interestingly, the slope for 3 steps gives a slope=1.66 and the more steps, the closer to 1 the obtained value. Did anyone else report problems with slope(x)? Or is there no problem and I and missing something?

What are the values in the vector when you get a bad answer?

Can you post a small piece of code that reproduces the problem?

Can you contstruct the vector in a LAMMPS script without running a simulation

and see the same problem?