Hi!
Problem solved.
Thank you Ross for your kind and fast reply!
This is just out of curiosity ...Despite the serious speed issue, do you
think that would come in newer cuda versions, i.e., is there somebody
working on this?
BTW: Great job, awesome code and extensive support!
All the best!
Guilherme
On Thu, May 16, 2013 at 4:58 PM, Ross Walker <ross.rosswalker.co.uk> wrote:
> Hi Guilherme,
>
> No COM is not supported on the GPU in AMBER 12. The fact the code did not
> quit is considered a bug. Please update your AMBER 12 tree and it will
> correctly quit with an error message if you try to run with COM restraints.
>
> Note calculating the center of mass of a large group of atoms is a
> difficult problem ultimately requiring a reduction in parallel. Hence
> using such restraints (even in the CPU code) will unfortunately lead to a
> reduction in efficiency.
>
> All the best
> Ross
>
>
>
> On 5/16/13 1:44 AM, "José Guilherme Vilhena Albuquerque d'Orey"
> <guilhermevilhena.gmail.com> wrote:
>
> >Dear Sir/Madam,
> >
> >I was wondering if the COM steered MD is already available for the
> >pmemd.cuda in amber12. I have tried it but it doesn't give me any error
> >nor
> >it makes what it should.
> >Furthermore how can I increase the number of residues that belong to the
> >COM (sander or pmemd). I need to pull a whole protein up (~ 1200
> >residues),
> >and apparently there is some kind of maximal limit to the string I can
> >use
> >to define the COM. I have tried to change that variable inside AMBER
> >(pmemd)
> >nmr_calls.F90: integer maxigr
> >The default is 200. I have tried to increase to 500 and it is still not
> >working. The error that is outputted is what follows.
> >Note that decreasing the number of residues in the COM it works.
> >
> >
> >pmemd.MPI:12508 terminated with signal 11 at PC=579d27 SP=7fff988ba9d0.
> > Backtrace:
> >/home/gvilhena/CODES/amber12_MOD/bin/pmemd.MPI(pme_slab_recip_mod_mp_fill_
> >charge_grid_+0x937)[0x579d27]
> >/home/gvilhena/CODES/amber12_MOD/bin/pmemd.MPI(pme_slab_recip_mod_mp_do_sl
> >ab_pmesh_kspace_+0x14d4)[0x5790e4]
> >/home/gvilhena/CODES/amber12_MOD/bin/pmemd.MPI(pme_force_mod_mp_pme_force_
> >+0x2d56)[0x562796]
> >/home/gvilhena/CODES/amber12_MOD/bin/pmemd.MPI(runmd_mod_mp_runmd_+0x1382)
> >[0x552932]
> >/home/gvilhena/CODES/amber12_MOD/bin/pmemd.MPI(MAIN__+0x808)[0x4eed78]
> >/home/gvilhena/CODES/amber12_MOD/bin/pmemd.MPI(main+0x3c)[0x41dd6c]
> >/lib64/libc.so.6(__libc_start_main+0xf4)[0x3bf881d994]
> >/home/gvilhena/CODES/amber12_MOD/bin/pmemd.MPI[0x41dc79]
> >
> >
> >Thanks in advance!
> >All the best,
> >Guilherme Vilhena
> >
> >--
> >====================================================
> >*Guilherme Vilhena, Ph.D*
> >
> >Universidad Autonoma de Madrid,
> >Departamento de Fisica Teorica de la Materia Condensada
> >Facultad de Ciencias, Modulo C-5
> >E-28049 Madrid, Spain
> >
> >tel: +34 91 497 2789
> >fax: +34-91-4974950
> >====================================================
> >_______________________________________________
> >AMBER mailing list
> >AMBER.ambermd.org
> >http://lists.ambermd.org/mailman/listinfo/amber
>
>
>
> _______________________________________________
> AMBER mailing list
> AMBER.ambermd.org
> http://lists.ambermd.org/mailman/listinfo/amber
>
--
====================================================
*Guilherme Vilhena, Ph.D*
Universidad Autonoma de Madrid,
Departamento de Fisica Teorica de la Materia Condensada
Facultad de Ciencias, Modulo C-5
E-28049 Madrid, Spain
tel: +34 91 497 2789
fax: +34-91-4974950
====================================================
_______________________________________________
AMBER mailing list
AMBER.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber
Received on Thu May 16 2013 - 09:00:03 PDT
