Dear Amber Users,
Well I had to ask this question. Did a lot of search
into the mailing list, have found some material but am still not very
clear and hence will greatly appreciate if I am given a bit of
direction here.
Ok so here it is. I learn that when one is using lagevin dynamics
setting "ig=-1" in the input parameter file is now recommended. I
therefore choose to set this parameter for my simulation. However,
having selected this option, I am now having queries as to what is
actually going on here. I am sorry if I am missing some very critical
technical details.
My issue is the following. I have a set of steps to perform during my
simulations from heating up to 300K in three steps of 1000K, to
equilibration and finally to production.
Part of the input file in the first round of heating:
ntx=1, irest=0,
ntt=3, gamma_ln=1.0, tempi=0.0, temp0=100.0, ig=-1
Part of the output file:
Langevin dynamics temperature regulation:
ig = 126603
temp0 = 100.00000, tempi = 0.00000, gamma_ln= 1.00000
So the random number here is as mentioned in this file.
Now when I am doing the next round of heating, part my input file is
as follows:
ntx=5, irest=1,
ntt=3, gamma_ln=1.0, tempi=100.0, temp0=200.0, ig=-1,
So the calculation should continue using the velocity information from
restart file saved at the end of the previous round of heating. But I
have still specified "ig=-1" and when I look at the output file
Langevin dynamics temperature regulation:
ig = 56545
temp0 = 200.00000, tempi = 100.00000, gamma_ln= 1.00000
So the random number now is different.
Subsequently, part of the input file for heating, equil and prodn:
ntx=5, irest=1,
ntt=3, gamma_ln=1.0, tempi=200.0, temp0=300.0, ig=-1,
For each restart, I have used "ig=-1" and consequently the random
number is different in each output.
Now I have my queries on this aspect.
1) For the first step I understand that the input file has only
coordinates and hence "ig=-1" took up a random number based on the
system wallclock and then velocities were assigned based on this. This
is fine.
2)But when I go for the next round with the restart file from the
previous round, it has the velocities, but since "ig=1", Does it mean
that the velocities are again rescaled based on this random number and
the initial temperature 'tempi" ?
Now if that is the case, then my velocities are rescaled at every
restart. This does not sound too comforting !!!!
I am sorry my mail has gone very long. But I will be deeply obliged if
someone can kindly throw some light into this aspect.
Thank you,
Dilraj.
_______________________________________________
AMBER mailing list
AMBER.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber
Received on Sat Jun 23 2012 - 01:30:02 PDT
