Molcas Forum

adutton

Member

Registered: 2017-03-24

Posts: 3

Error message while running QMMM with Molcas/Tinker

Hello,

I thought that we could use 1000 "active" atom limit when trying to run a Molcas/Tinker QMMM.  I can run file.input, file.prm, file.xyz (this actually has all 303 water molecules in it) with the file.key shown below and the job works great!  (It just runs with all the first 303 atoms in file.xyz)
file.key:

parameters file.prm
DIGITS 8
VERBOSE
extraterm none
QMMM  303
QM -1 4
MM -5 303
QMMM-MICROITERATION ON
ACTIVE -1 303

If I change the file.key to include all 303 water molecules.

parameters file.prm
DIGITS 8
VERBOSE
extraterm none
QMMM  913
QM -1 4
MM -5 913
QMMM-MICROITERATION ON
ACTIVE -1 913

I get this error message:

RdCtl: Increase Mxdc
Mxdc=                  500

1. I thought that we could use 1000 active QM/MM atoms?
2. I tried to search for how to increase the Mxdc but good not figure it out.  I found a source code page on d_bond.f but this did not make much sense to me as to how that would give me the keywords I need in the file.input to allow more atoms in the QMMM.  Is there a way to increase the Mxdc related to QMMM?
3. For my project I can probably get away with fewer water molecules solvation shell but it would be nice to have this many active waters.  Is there a way to allow more than just 500 atoms to move (if that is what the mxdc = 500) means?

Thanks for your help,
adutton

Ignacio

Administrator

From: Uppsala

Registered: 2015-11-03

Posts: 1,031

Re: Error message while running QMMM with Molcas/Tinker

Molcas 8.0 has Mxdc = MxAtom = 500, Molcas 8.2 has 5000... If you have the source code you can modify it in src/Include/Molcas.fh (but it's not guaranteed it won't break something).

niko

Member

From: Marseille

Registered: 2015-11-08

Posts: 59

Website

Re: Error message while running QMMM with Molcas/Tinker

In addition to Ignacio's comment, I can't imagine a situation where you would need to make Molcas know about such a large number of MM atoms. Do you have any particular reason for doing that?

Actually, you make "ACTIVE" all the QM and MM atoms which are transferred to Molcas (no need to do that, all "QMMM" atoms are automatically set to "ACTIVE"), which implies that their geometry optimization would be carried out by Slapaf. Hence you shall not see any MM minimization in Tinker's output.

Usually, I only add the relevant MM atoms, essentially the ones needed to use the LAMorokuma keyword in &ESPF and let Tinker manage the MM minimization at each QM geometry optimization step (of course, using the Cartesian and rHidden keywords in &Slapaf).