From: Michael Robinson (michael.robinson1_at_monash.edu)
Date: Fri Oct 16 2020 - 01:23:57 CDT
Hi namd-l,
To the best of my knowledge, the latest non-alpha release of NAMD (2.14)
supports GPU acceleration for Drude simulations that do not use the
non-bonded Thole method, and the NAMD3 alpha development does not include
the use of polarisable Drude force fields. My understanding of the current
state of the CHARMM-Drude force field is that the use of the Thole
shielding scheme is typically considered the better way to go about these
simulations - and it enables a larger timestep. Given this, is there a
generally accepted decision in the choice of non-GPU accelerated Drude
Thole based simulations, or GPU-accelerated Drude simulations without the
use of Thole shielding, or is this typically handled on a case-by-case
basis based on timing?
Thank you for any help you can provide,
Regards,
Michael Robinson
This archive was generated by hypermail 2.1.6 : Fri Dec 31 2021 - 23:17:10 CST