accelMD and GBIS?

From: Brian Radak (brian.radak.accts_at_gmail.com)
Date: Fri Feb 27 2015 - 18:15:13 CST

Google seems to bring up various threads involving potential bugs with
accelerated MD and GBIS. However I seem to be seeing similar problems in a
slightly different context.

I am running implicit solvent aMD with a dipeptide on NAMD multicore (*not*
CUDA). In normal dihedral only mode, the dihedral energy in the aMD and
ENERGY fields never seem to differ, even when dV comes back non-zero. Here
is my first frame:

ACCELERATED MD: STEP 0 dV 1.10188 dVAVG 1.10188 BOND 4.04848 ANGLE 12.6857
DIHED 16.3605 IMPRP 0.371159 ELECT -155.46 VDW -1.45463 POTENTIAL -123.449
ETITLE: TS BOND ANGLE DIHED
    IMPRP
             ELECT VDW BOUNDARY MISC
KINETIC
              TOTAL TEMP POTENTIAL TOTAL3
TEMPAVG

ENERGY: 0 4.0485 12.6857 16.3605
      0.3712
          -155.4598 -1.4546 0.0000
0.0000 20.6767
           -102.7719 320.1538 -123.4486 -102.5823
    320.1538

Here accelMDE = 18.0 and accelMDalpha = 0.8. I'm a little new to aMD, so I
can't tell if this is a particularly aggressive boost or not (based on the
NAMD aMD paper and alanine dipeptide, it should be pretty good). However,
my phi/psi coverage seems to be pretty poor after 5,000,000 steps.

Is something wrong here or am I misunderstanding the output and/or using
poor boost parameters?

Thanks,
Brian

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

This archive was generated by hypermail 2.1.6 : Thu Dec 31 2015 - 23:21:41 CST