Hi all,
I am attempting to run in vacuo simulations and I would like to know which
would be
the best setting with sander module.
As stated at p.60 of Amber14 user guide, igb=6 would correspond to an in
vacuo
simulation and I quote here what it is written "This option is logically
equivalent to
setting igb=0 and eedmeth=4, although the implementation (and computational
efficiency)
is not the same".
It was not clear to me in which of the 2 cases (igb=6 or ntb=0, igb=0,
eedmeth=4)
the computational efficiency should improve, therefore I made some tests
with the
3 different temperature couplings available in Amber12, here are the
results:
*igb=6*
Berendsen: 113.7 ns/day
Andersen(vrand=1000): 70.12 ns/day
Langevin(gamma_ln=2): 61.79 ns/day
*ntb=0, igb=0, eedmeth=4*
Berendsen: 66.48 ns/day
Andersen(vrand=1000): 68.94 ns/day
Langevin(gamma_ln=2): 72.70 ns/day
I ran a simulation of a 758 atoms molecule, on a blade with 8 CPUs dual core
(hence 16 cores) and hyper-threading (hence 32 threads), however I
utilised only 16 cores of such a blade and with the following
sander input file (example from the simulation with Berendsen and igb=6):
dyn in vacuo (igb=6)
&cntrl
imin = 0, irest = 1, ntx = 5,
nstlim = 100000, dt = 0.001,
ntc = 2, ntf = 2,
cut = 1000.0,
igb = 6,
ntpr = 5000, ntwr = 10000, ntwx = 2000,
ntt = 1,
temp0 = 300.0,
ioutfm = 1
/
Therefore I would obviously be tempted to utilise Berendsen thermostat with
igb=6,
as it is the fastest one (113.7 ns/day)
But, provided that Berendsen coupling is kind of deprecated for several
reasons (Amber14
user guide, p.290), especially for simulations in implicit solvent, in
which a uniform
temperature is not guaranteed because collisions with water molecules are
missing, do you think that it would be a sensible choice to use Berendsen
coupling
for in vacuo simulations?
My answer would be yes for two reasons:
1) performance efficiency
2) with in vacuo simulations one does not actually want
any "random kick" (like it occurs with Andersen or Langevin
coupling) that would emulate the collisions with the solvent.
Am I right/wrong?
Any comments/hints on this issue would be very appreciated.
Many thanks!
PS Just to be meticulous (running 1 step of minimisation) I have also
checked if the potential functions utilised were
actually the same between igb=6 and igb=0, ntb=0, eedmeth=4 settings and
they
perfectly match :) as you can see below (therefore igb=6 with Berendsen is
actually
the most efficient implementation):
*igb=6*
NSTEP ENERGY RMS GMAX NAME NUMBER
1 1.5157E+03 2.1721E+01 9.0558E+01 C2 367
BOND = 196.4394 ANGLE = 427.5654 DIHED =
570.5536
VDWAALS = -413.8957 EEL = 3712.9904 EGB =
0.0000
1-4 VDW = 205.9521 1-4 EEL = -3183.9443 RESTRAINT =
0.0000
*ntb=0, igb=0, eedmeth=4*
NSTEP ENERGY RMS GMAX NAME NUMBER
1 1.5157E+03 2.1721E+01 9.0558E+01 C2 367
BOND = 196.4394 ANGLE = 427.5654 DIHED =
570.5536
VDWAALS = -413.8957 EEL = 3712.9904 HBOND =
0.0000
1-4 VDW = 205.9521 1-4 EEL = -3183.9443 RESTRAINT =
0.0000
--
Dr Massimiliano Porrini
Valérie Gabelica Team
U869 ARNA - Inserm / Bordeaux University
Institut Européen de Chimie et Biologie (IECB)
2, rue Robert Escarpit
33607 Pessac Cedex
FRANCE
Tel : 33 (0)5 40 00 63 31
http://www.iecb.u-bordeaux.fr/teams/GABELICA
Emails: m.porrini.iecb.u-bordeaux.fr
mozz76.gmail.com
_______________________________________________
AMBER mailing list
AMBER.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber
Received on Thu May 22 2014 - 13:00:03 PDT
