Slight Discrepancies when Reproducing Ti HCP/BCC Transition

[if3598]

Hello! I am trying to reproduce Figure 2 from Automated free energy calculation from atomistic simulations using the same EAM model that was used in the paper.

Overall, I get pretty good qualitative agreement with the published results:

but was wondering if any minor quantitative disagreements (e.g. in my run, the predicted transition temperature at 1 GPa is higher than the predicted transition temperature at 0 GPa) could be explained by settings in my input.yaml file.

I've attached a zip folder with my input file and starting structures for your reference. Thanks so much in advance for any advice or feedback!

Ti_HCP_BCC.zip

[srmnitc]

@if3598 thanks for raising this! Its a bit hard for me to point out exactly. But, this is what I have left of the original input. Calphy's internal pressure converge threshold has since changed, that could be a reason. Another thing I can notice is that I used 50 ps. Overall, I think maybe the best thing to do is to repeat the calculations 3-5 times, and add some error bars.