Please login to be able to save your searches and receive alerts for new content matching your search criteria.
A combination of replica exchange Monte Carlo sampling techniques and energy landscape paving approach is presented. This hybrid algorithm combines the features of the energy landscape paving (ELP) and replica exchange methods (REM). I have tested its performance in studying the low-energy conformations of the benchmark peptide Met-enkephalin.
Replica exchange molecular dynamics (REMD) code has been developed for predicting protein tertiary structure and optimized assuming that simulations are performed on the Earth Simulator. A new parallel programming approach using an MPI/OpenMP model with a three-level hierarchy was developed to circumvent the problems of overhead inherent in the MPI model and shown to achieve an 88% parallel efficiency using a 256-node REMD simulation on the Earth Simulator with this model.