Berkeley COMPSCI C267 - Replica Exchange Molecular Dynamics (REMD) for Amber’s Particle-Mesh Ewalds MD (PMEMD) code (6 pages)

Previewing pages 1, 2 of 6 page document View the full content.
View Full Document

Replica Exchange Molecular Dynamics (REMD) for Amber’s Particle-Mesh Ewalds MD (PMEMD) code



Previewing pages 1, 2 of actual document.

View the full content.
View Full Document
View Full Document

Replica Exchange Molecular Dynamics (REMD) for Amber’s Particle-Mesh Ewalds MD (PMEMD) code

184 views

Other


Pages:
6
School:
University of California, Berkeley
Course:
Compsci C267 - Applications of Parallel Computers
Applications of Parallel Computers Documents

Unformatted text preview:

Replica Exchange Molecular Dynamics REMD for Amber s Particle Mesh Ewalds MD PMEMD code Lia Ball Teresa Head Gordon Group Goal Converging A monomer Molecular Dynamics simulations Use Amber ff99SB force field and TIP4P ew explicit water model to sample an all atom representation of A conformational ensemble High temperature simulations allow sampling of minima separated by large energy barriers Replica Exchange Molecular Dynamics REMD runs several independent simulations run at different temperatures in parallel REMD from 450 K to 287 K Run PMEMD simulations 450 K independently for all replicas Every picosecond of simulation time attempt to exchange two replicas that are close in temperature Energy minima accessed at high temperature will exchange down to the low temperature replica over time sync sync 287 K final ensemble PMEMD is faster than Sander PMEMD is a version of Sander MD code that is optimized to perform parallelized particle mesh Ewalds calculations which use Fast Fourier Transforms to calculate the long range forces on atoms On 16 processors PMEMD takes 33 s to perform 1ps of simulation on one 25 000 atom system including setup Sander takes 55 s If exchanges and synchronization time are less than 22 s my code will be faster than Sander REMD Exchange criteria are monte carlo criteria that depend only on replica temperatures and energies Structure information does not need to be transferred between processors Subroutine Organization Instead of one master task I create a master for each replica that does everything that the master does in the original code Each master has its own set of global variables that are never shared with the other replicas pmemd get cmdline master setup runmd bcast dat pme force rep setup exchange check my atom movement check new list limit Md setup distribute crds Light blue boxes indicate subroutines that I modified from the original code but that are still only run once for the entire program Pink subroutines are those that I



View Full Document

Access the best Study Guides, Lecture Notes and Practice Exams

Loading Unlocking...
Login

Join to view Replica Exchange Molecular Dynamics (REMD) for Amber’s Particle-Mesh Ewalds MD (PMEMD) code and access 3M+ class-specific study document.

or
We will never post anything without your permission.
Don't have an account?
Sign Up

Join to view Replica Exchange Molecular Dynamics (REMD) for Amber’s Particle-Mesh Ewalds MD (PMEMD) code and access 3M+ class-specific study document.

or

By creating an account you agree to our Privacy Policy and Terms Of Use

Already a member?