REMD-GPU

From: Francesco Pietra (chiendarret_at_gmail.com)
Date: Thu Jul 04 2013 - 02:36:59 CDT

Hello:
I am interested in running REMD on a simple GPU machine (intel six
core and two GTX-680, namd2.9-cuda) to have a large, complex system
(transmembrane, thus also lipidic membrane, in addition to
polypeptidic ligands) in order for the supercomputer center.

I came across a few REMD-GPU posts of two years ago. May I ask about
any subsequent experience, also as to namd2.10? I am not much
concerned about speed, nor using CPU_to_GPU switch, just the machine
as it is, or even the CPUs only if GPU is problematic.

Also, being novice to REMD, is it conceivable to restrain by harmonic
forces all non flexible part of the protein and the peptidic ligands,
leaving to REMD modeling to only the flexible part of the protein and
the surrounding water?

thanks

francesco pietra

This archive was generated by hypermail 2.1.6 : Wed Dec 31 2014 - 23:21:23 CST