[gmx-users] replica exchange 1%

[Daniel Burns]

Hi,

I'm using the Plumed plugin to perform a "Hamiltonian" replica exchange.

I have histograms of the total energies of each system that share a huge
overlap - on the order of 80%.

Only a small subset of the residues have perturbed energy functions.

Replica exchange attempts only happen every 2000 steps.

Pressure coupling is every 5 steps.

Despite the above, I am only experiencing successful exchanges 1% of the
time.  Any suggestions on what might be the reason for such low exchange
probabilities?

Thank you!

Dan