Fwd: (1) Failure running QM-MM with ORCA and (2) problems preparing input with transition metals

From: Francesco Pietra (chiendarret_at_gmail.com)
Date: Mon Jan 02 2017 - 04:54:19 CST

Part (1) (running with orca) SOLVED once openmpi 1.6.5-9.1 was system-wide
installed. At least, it is running very slowly in /dev/shm on this Intel(R)
Core(TM) i3-4160 CPU @ 3.60GHz with little memory (total 7942 free 4567).

Part (2) waiting

thanks

francesco pietra
---------- Forwarded message ----------
From: Francesco Pietra <chiendarret_at_gmail.com>
Date: Mon, Jan 2, 2017 at 8:38 AM
Subject: namd-l: (1) Failure running QM-MM with ORCA and (2) problems
preparing input with transition metals
To: NAMD <namd-l_at_ks.uiuc.edu>, melomcr_at_gmail.com, rcbernardi_at_ks.uiu.edu,
trudack_at_ks.uiuc.edu

Hello:

(1)
---- Tried running Example1 as indicated

/path/to/namd2 config.namd 2>&1 | tee out.log

execution OK with MOPAC, although problems in examining the results with
VMD: unclear to me the order in which files should be added and whether tcl
scripts should be sourced or "vmd -e prepare.qm.region.tcl" should be run
instead

Using ORCA 3.0.3 installed system-wide on a linux amd64 box with a 4-core
Intel(R) Core(TM) i3-4160 CPU @ 3.60GHz

/path/to/namd2 config.ORCA.namd 2>&1 | tee out.log

---- Changes made to config.ORCA.namd:

qmBaseDir "/dev/shm/NAMD_Example1"

qmExecPath "/usr/local/bin/orca_3_0_3_linux_x86-64/orca"

---- Relevant from the .out file:

Charm++> Running on 1 unique compute nodes (4-way SMP).

TCL: Minimizing for 100 steps
Info: List of ranks running QM simulations: 0.
sh: 1: mpirun: not found
ORCA finished by error termination in ORCA_GTOInt
ERROR: Could not find QM output file!

I was unable to grasp how NAMD and ORCA try to communicate.
________________________________________

(2) The script "prepare.qm.region.tcl" requires filename.psf / filename.pdb
in a water box for the system of interest. How getting those files if the
QM region includes one (or more) transition metal(s)? "vmd -dispdev text -e
filename.gen", which I commonly use to prepare psf/pdb requires a topology
that includes the metal(s). Can that be in the frame of a non-bonded model,
or what else? Or - for the time being - are transition metals intractable?
I hope not so as QM-MM is chiefly useful when transition metals are
involved, or, at least, my current interest is in this direction.

Thanks for advice

francesco pietra

This archive was generated by hypermail 2.1.6 : Sun Dec 31 2017 - 23:20:56 CST