From: Jerome Henin (jhenin_at_vitae.cmm.upenn.edu)
Date: Wed Sep 27 2006 - 15:31:02 CDT
To get the optimal parallel efficiency, I would recommend to run all windows
simultaneously on a smaller number of processors. There is no standard
procedure to prepare a starting point for each window, this is highly
system-dependent. Sometimes it is easier to make a setup from scratch,
sometimes SMD can be used.
Only one thing has to be carefully considered: the value of the RC in the
starting configuration should be within the target range, or very close to
it, otherwise the boundary potentials will result in unreasonable forces at
the beginning of the run.
Jerome
On Tuesday 26 September 2006 23:41, Vani Krishna wrote:
> Hi,
>
> in my case, I ran a series of 'serial' runs, splitting the range into
> several windows. so for example, I would run (assuming window size is 5 A)
> -10 to -5 , and then use the restart runs for the next window, -5 to 0.
>
> ËïæÃæà <stting_at_mail.nankai.edu.cn> wrote: I want to compute free energy of
> the system by abf, I set the abf.conf file as follows :
>
> source /NAMD_2.6_Linux-i686/lib/init.tcl
> package require abf
>
> abf coordinate zCoord-1atom
>
> abf abf2 149
> abf abf1 {8 29 50 71 92 113 134}
>
> abf dxi 0.5
> abf xiMin -10
> abf xiMax 22
> abf outFile wb.dat
> abf fullSamples 500
> abf distFile wb.dist
> abf dSmooth 0.4
> abf applyBias yes
> abf historyFile wb.history
> run 100000
>
>
> But now I want to use eight windows to improve sampling efficiency,how I to
> set the abf.conf file? Thank you for any help!
>
>
>
>
> ---------------------------------
> Talk is cheap. Use Yahoo! Messenger to make PC-to-Phone calls. Great rates
> starting at 1¢/min.
This archive was generated by hypermail 2.1.6 : Wed Feb 29 2012 - 15:42:37 CST