Re: temperature stepping during heating

From: JC Gumbart (
Date: Thu Apr 17 2008 - 16:35:13 CDT

The minimization in fact is probably hurting you. Plot your
temperature as a function of time. After each minimization, I expect
your temp drops significantly and then takes longer than 10,000 steps
to get back to the target temp.

To be honest, I'm not sure such a procedure is necessary, but
certainly not in such small increments. You could try increasing the
temperature in 20 degree increments for example. In my experience
though, many of us just start simulating at 300 K. Since
minimization however cuts the temperature in half, it then takes a
couple hundred ps for the thermostat to raise the system back up to
300 K. So that is nearly the same as what you're trying to do anyway!

I don't know anymore about the additional commands than what is in
the manual.

On Apr 15, 2008, at 5:21 AM, Low Soo Mei wrote:

> Hi,
> I have a minimized membrane and am now heating it in steps up till
> 450 K, in NVT, to randomize the lipid tails. A few atoms of my
> lipid headgroup are lightly constrained throughout. I'm doing the
> heating very cautiously in 5 K steps, with 1000 steps minimization
> between each temperature step. This is taking a very long time.
> Would anyone here know if the in-between minimization steps are
> necessary, or if I can use a bigger temperature step?
> The last part of my config file is as below:
> # run one step to get into scripting mode
> minimize 0
> for {set i 5} {$i < 455} {incr i 5} {
> set temperature $i
> langevinTemp $temperature
> run 10000
> output heat_flash.$temperature
> minimize 1000
> output heat_flash.$temperature.min
> reinitvels $temperature
> }
> As you can see, after each minimization I am using reinitvels to
> give back the pre-minimization temperature to my system so that it
> will not restart at 0 K. I was also wondering what the differences
> are between this (reinitvels) and the temperature rescaling and
> temperature reassignment commands in the NAMD user guide.
> Thanks!
> Cheers,
> Soo Mei

This archive was generated by hypermail 2.1.6 : Wed Feb 29 2012 - 15:47:43 CST