Test rerun with FEP
When using FEP states with rerun, the FEP state is set to 0 instead of the index read from the rerun trajectory file.
#6 Updated by Berk Hess almost 2 years ago
No, I never used this code. I only noticed it can't be correct when refactoring the code.
Do you agree with my assessment that you always get lambda state 0?
I would think that running a few steps with a setup with fep_state > 0 and then doing a rerun will show the issue.
#7 Updated by Mark Abraham almost 2 years ago
Michael Shirts wrote:
Berk, could you post a sample file that cause the problem? I'd love to get a regression test in the free energy section that covers this bug.
Possibly this explains the issue I found at https://gerrit.gromacs.org/#/c/5435/ but didn't try to find out how to solve. (I have further progress on that patch offline, but won't get back to it until we get a 2017 beta out!)