From: hansang bae (baeh_at_ecn.purdue.edu)
Date: Mon Dec 29 2003 - 09:19:53 CST
I observed that simulations taking many steps (> 1000 maybe) produced
different results for a serial run and a parallel run (even without using
Langevin method). How can one avoid these errors? If it's inevitable, how
can one validate their parallel simulations?
I'm asking this basic question since I'm not a MD expert.
This archive was generated by hypermail 2.1.6 : Wed Feb 29 2012 - 15:37:16 CST