I am computing viscosity using gmx energy. I observe that the shear viscosity in the visco.xvg file is different from that in the evisco.xvg file by a factor 4. What is the reason for this difference and which file should I refer to for shear viscosity?
In addition, for a simulation run of 200 ns, the data in the visco.xvg file is for 100 ns whereas data in the evisco.xvg file is for 50 ns. What would be the reason for this discrepancy?
How often do you write out your energies? For water I an output interval as small as 0.05 ps to get reasonable results due to the short correlation times.
The output is all integrals over correlation functions. These get more noisy as time progresses, so they are off, somewhat arbitrarily, all half and 1/4 of the analysis interval length.
Thank you for your response. I’m using MARTINI forcefield for which my time step is 20 fs and I write my energy every 10 timesteps, thus 200 ps. For a 200 ns run my visco.xvg has data for 100 ns whereas my evisco.xvg has data for 50 ns. Why is that discrepancy in the output file?
My viscosity from visco.xvg at 50 ns is 570 cP whereas at 100 ns it is 1143 cP (seems like a garbage value). Viscosity from evisco.xvg at 50 ns is 466 cP.
Last few data points from both the files are as below: