It might also be worth noting that, especially for small QM regions,
parallelization can actually diminish performance. I've seen this happen
with as few as 8 processors. As far as I know, this is a limitation of the
algorithms, not the sander implementation. There is probably a similar
problem for GPU accelerated QM/MM.
On a separate note, has anyone been following the GPU developments in
MOPAC? I'm sure there are some difficulties just with using the special GPU
linear algebra libraries, etc. Are there any plans to bring this to SQM
(obviously *very* long term)?
Regards,
Brian
On Mon, Mar 24, 2014 at 12:35 PM, Jason Swails <jason.swails.gmail.com>wrote:
> On Mon, 2014-03-24 at 16:17 +0000, John Smith wrote:
> > Hi AMBER users
> >
> > I am trying to run a QM/MM MD simulation on a protein ligand complex (431
> > amino acid protein). I used Antechamber initially for the prepin and
> frcmod
> > files. I then used xLeap to generate the prmtop and inpcrd files. I am
> > trying to run a QM/MM simulation where the quantum region would include
> the
> > ligand and a key residue in the protein (CYS123). I am using AMBER12 and
> > have a GPU machine. I noticed that I cannot use the PMEMD to run QM/MM
> and
> > I have to use Sander, even with Sander I am getting the error message
> > "Internal namelist read error" and the calculation stops. The input files
> > is as follows:
> >
> > Initial minimisation
> > &cntrl
> > imin=1, maxcyc=2000, ncyc=1000,
> > cut=12, ntb=1, igb=0, ifqnt=1
> > /
> > &qmmm
> > qmmask = '123, 432'
> > qmcharge = 0,
> > qm_theory = 'PM3'
> > qmcut = 12.0
> > /
>
> qmmask is not a valid mask. See Appendix C of the Amber manual for
> details about the mask syntax. (I think you meant ':123,432')
>
> Also, putting commas after every variable in the input file might
> prevent these kinds of errors with some compilers.
>
> > END
> > END
> >
> > This input file is for the initial minimisation and I was planning to use
> > the QM part on the MD stages as well but unfortunately I am being unable
> to
> > progress because of the aforementioned error message.
> >
> > Could you please help me with that?
> >
> > Also is there a command that I can use to run the calculations on the
> GPU?
>
> No. The time-consuming part is the QM/MM part, anyway, so GPU
> acceleration would not buy much speed enhancement. Semi-empirical QM/MM
> is not well-suited for GPU acceleration (and only mopac can do it right
> now, AFAIK).
>
> HTH,
> Jason
>
> --
> Jason M. Swails
> BioMaPS,
> Rutgers University
> Postdoctoral Researcher
>
>
> _______________________________________________
> AMBER mailing list
> AMBER.ambermd.org
> http://lists.ambermd.org/mailman/listinfo/amber
>
--
================================ Current Address =======================
Brian Radak : BioMaPS
Institute for Quantitative Biology
PhD candidate - York Research Group : Rutgers, The State
University of New Jersey
University of Minnesota - Twin Cities : Center for Integrative
Proteomics Room 308
Graduate Program in Chemical Physics : 174 Frelinghuysen Road,
Department of Chemistry : Piscataway, NJ
08854-8066
radak004.umn.edu :
radakb.biomaps.rutgers.edu
====================================================================
Sorry for the multiple e-mail addresses, just use the institute appropriate
address.
_______________________________________________
AMBER mailing list
AMBER.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber
Received on Mon Mar 24 2014 - 10:00:04 PDT
