Hi,
I don't think it is necessarily related. You get this kind of error when
you lose track of atoms for a certain reason. In this case it looks like
you might be pulling too fast? (13A in 100 ps, which would be ~46 km/hr?).
Have you tried with lower velocities? Try doing the simulation at about 1
A/ns and tell us if then it works. Also please tell us what version of
AMBER you are using, with what patches and make sure you start from an
"equilibrated" structure. If it is in anyway related to the COM restraint
issues, I know that a patch from Charles Lin is on the way, so it should be
working in the next days.
Cheers,
El mié., 13 mar. 2019 a las 21:48, Pratul Agarwal (<pratul.agarwal-lab.org>)
escribió:
> This sounds similar to the issues being discussed here:
> http://archive.ambermd.org/201903/0095.html
> or
> http://archive.ambermd.org/201903/0056.html?
>
> Pratul K. Agarwal, Ph.D.
> (Editorial Board Member: PLoS ONE, Microbial Cell Factories)
> Web: http://www.agarwal-lab.org/
>
>
> On 3/13/2019 1:21 PM, Gerardo Zerbetto De Palma wrote:
>
> Hello! First time in this mailing list.
> I am having an issue making a pull between two centers of mass of two atom
> groups, using pmemd.cuda. The stdout reads:
>
> cudaMemcpy GpuBuffer::Download failed an illegal memory access was
> encountered
>
> When I run this on sander, it works just fine, so I think the system is not
> blowing up.
> I have also tried:
> - running a plain simple simulation (not pulling) with pmemd.cuda and in
> also worked just fine.
> - running a simulation pulling from 2 individual atoms (not center of mass)
> and it also worked fine.
>
> I am attaching the mdin, restraint and output files for the center of mass
> pulling.
> Thanks a lot.
> Best regards!
>
> Gerardo Zerbetto De Palma
> Universidad de Buenos Aires
>
>
>
>
> _______________________________________________
> AMBER mailing list
> AMBER.ambermd.org<mailto:AMBER.ambermd.org>
> http://lists.ambermd.org/mailman/listinfo/amber
>
>
> _______________________________________________
> AMBER mailing list
> AMBER.ambermd.org
> http://lists.ambermd.org/mailman/listinfo/amber
>
--
Stephan Schott Verdugo
Biochemist
Heinrich-Heine-Universitaet Duesseldorf
Institut fuer Pharm. und Med. Chemie
Universitaetsstr. 1
40225 Duesseldorf
Germany
_______________________________________________
AMBER mailing list
AMBER.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber
Received on Fri Mar 15 2019 - 07:30:03 PDT
