Re: energy minimization in namd

From: Brian Radak (brian.radak.accts_at_gmail.com)
Date: Thu Apr 23 2015 - 11:51:26 CDT

I'm not an expert with the NAMD minimizer, but it looks like the DX, DU,
and DUDX are just limits on how big the minimization steps will try to
be. If I had to guess I would say they are the displacement (DX), change
in energy (DU), and change in gradient (DUDX).

10000 steps is /a lot/ of minimization, especially if you just want to
remove bad contacts before equilibration. In my experience, 500 to 1000
steps should be plenty, even for large systems (>100,000 atoms). Also
(and again, at least in my experience), excessive minimization will not
decrease the length of a requisite equilibration period and can in fact
make things worse depending on how the minimization is done.

As long as your equilibration period starts with rapid stabilization of
energy and volume (usually on the order of hundreds of ps), I'd say it
is likely to be successful (in the sense of not crashing your system).

Cheers,
Brian

On 04/23/2015 11:29 AM, Hasan haska wrote:
> Dear All,
>
> I run 10000 step energy minimization for my simulation system. The
> last part of my namd out file can be seen in below. What are DX , DU
> and DUDX in “LINE MINIMIZER BRACKET” line ? And How can I understand
> that 10000 step energy minimization is enough or not to start
> equilibration ?
>
> Thanks for the explanation.
>
> ************
>
> LINE MINIMIZER BRACKET: DX 0.000236595 0.000674418 DU -0.015094
> 0.0970886 DUDX -97.2742 2.11463 285.38
> PRESSURE: 9999 -573.684 -11.0714 1.23184 -11.0714 -539.603 14.7132
> 1.23184 14.7132 -162.05
> GPRESSURE: 9999 -572.433 -10.8975 1.2972 -11.0482 -538.393 14.5997
> 1.27755 14.7093 -160.769
> ENERGY: 9999 5759.6569 2577.9163 457.8570 0.0000
> -132654.3270 25903.7601 0.0000 0.0000 0.0000
> -97955.1366 0.0000 -97955.1366 -97955.1366
> 0.0000 -425.1123 -423.8652 1361250.0000 -425.1123
> -423.8652
>
> LINE MINIMIZER BRACKET: DX 2.36595e-05 0.000674418 DU -0.000307767
> 0.0970886 DUDX -7.82503 2.11463 285.38
> PRESSURE: 10000 -573.698 -11.0727 1.22639 -11.0727 -539.606 14.7258
> 1.22639 14.7258 -162.109
> GPRESSURE: 10000 -572.424 -10.8974 1.29553 -11.049 -538.377 14.6077
> 1.27498 14.7187 -160.804
> TIMING: 10000 CPU: 655.865, 0.0655963/step Wall: 655.865,
> 0.0655963/step, 0 hours remaining, 1873.847656 MB of memory in use.
> ETITLE: TS BOND ANGLE DIHED IMPRP
> ELECT VDW BOUNDARY MISC KINETIC
> TOTAL TEMP POTENTIAL TOTAL3 TEMPAVG
> PRESSURE GPRESSURE VOLUME PRESSAVG GPRESSAVG
>
> ENERGY: 10000 5759.7022 2577.9230 457.8567 0.0000
> -132654.3854 25903.7665 0.0000 0.0000 0.0000
> -97955.1369 0.0000 -97955.1369 -97955.1369
> 0.0000 -425.1379 -423.8685 1361250.0000 -425.1379
> -423.8685
>
> WRITING EXTENDED SYSTEM TO RESTART FILE AT STEP 10000
> WRITING COORDINATES TO DCD FILE AT STEP 10000
> WRITING COORDINATES TO RESTART FILE AT STEP 10000
> FINISHED WRITING RESTART COORDINATES
> The last position output (seq=10000) takes 0.004 seconds, 1873.848 MB
> of memory in use
> WRITING VELOCITIES TO RESTART FILE AT STEP 10000
> FINISHED WRITING RESTART VELOCITIES
> The last velocity output (seq=10000) takes 0.003 seconds, 1873.848 MB
> of memory in use
> REINITIALIZING VELOCITIES AT STEP 10000 TO 309 KELVIN.
>
>

-- 
Brian Radak
Postdoctoral Scholar
Gordon Center for Integrative Science, W323A
Department of Biochemistry & Molecular Biology
University of Chicago
929 E. 57th St.
Chicago, IL 60637-1454
Tel: 773/834-2812
email: radak_at_uchicago.edu

This archive was generated by hypermail 2.1.6 : Tue Dec 27 2016 - 23:21:05 CST