[gmx-users] 4. QM/MM Calculation with Orca (Minos Matsoukas)

[Minos Matsoukas]

Thank you very much for your answer,
After setting up some qm/mm runs,  I clearly understand what you mean.
Wow, the qm calculations are some cpu eating monsters...
M.

On Fri, Jun 8, 2012 at 3:46 PM, Gerrit Groenhof <ggroenh at gwdg.de> wrote:
> YOucan only use one thread in mdrun, but more than one in ORCA. Try to estimate the ratio of computation time spent between the QM and MM calculation to get an idea of why we never bothered to parallellize the MM part.
>
> Hope this helps,
>
> Gerrit
>
>
>>
>>   4. QM/MM Calculation with Orca (Minos Matsoukas)
>>
>>
>> Dear GROMACS Users,
>>
>> I am using Gromacs with ORCA for a qm calculation, but I can only use
>> 1 processor for mdrun.
>>
>> If I place the line !PAL4 in the topol.ORCAINFO and run mdrun -nt 1,
>> it works, and ORCA parallelizes correctly for 4 processors, although
>> mdrun only runs one thread.
>>
>> If I place the line !PAL4 in the topol.ORCAINFO and run mdrun -nt 4,
>> it doesn‚t work. mdrun crashes with a segmentation fault error :
>>
>> Back Off! I just backed up md.log to ./#md.log.35#
>> Reading file topol.tpr, VERSION 4.5.5 (single precision)
>> Starting 4 threads
>> QM/MM calculation requested.
>> QM/MM calculation requested.
>> QM/MM calculation requested.
>> QM/MM calculation requested.
>> there we go!
>> there we go!
>> there we go!
>> there we go!
>> Layer 0
>> nr of QM atoms 73
>> QMlevel: B3LYP/STO-3G
>>
>> /opt/soft/orca_2_9_1_linux_x86-64/opt/soft/orca_2_9_1_linux_x86-64...
>> orca initialised...
>> Segmentation fault (core dumped)
>>
>>
>> Is this because mdrun can only run with 1 thread when used with ORCA ?
>>
>> Here‚s some information :
>> OS: CentOS 6.2 x86_64 kernel 2.6.32-220
>> GROMACS: 4.5.5 compiled with options : --with-qmmm-orca --without-qmmm-gaussian
>> ORCA: 2.9.1 for x86_64
>> OPENMPI: 1.4.3
>>
>> My minimization mdp file is:
>> ------------
>> ;
>> ;     Input file
>> ;
>> title               =  Yo                     ; a string
>> cpp                 =  /lib/cpp                       ; c-preprocessor
>> integrator          =  cg                     ;
>> ;integrator               = l-bfgs
>>
>> rlist               =  1.0                    ; cut-off for ns
>> rvdw                =  1.0                    ; cut-off for vdw
>> rcoulomb            =  1.5                    ; cut-off for coulomb
>> ;       Energy minimizing stuff
>> ;
>> nsteps              =  200
>> emtol               =  10
>> emstep              =  0.1
>> ;define              = -DPOSRES
>> constraints         = none
>> QMMM                     = yes
>> QMMM-grps                = MOL
>> QMMMscheme               = normal
>> QMbasis                  = STO-3g
>> QMmethod                 = b3lyp
>> QMcharge                 = 1
>> QMmult                   = 1
>> -------------
>>
>> Thanks in advance
>
> --
> gmx-users mailing list    gmx-users at gromacs.org
> http://lists.gromacs.org/mailman/listinfo/gmx-users
> Please search the archive at http://www.gromacs.org/Support/Mailing_Lists/Search before posting!
> Please don't post (un)subscribe requests to the list. Use the
> www interface or send it to gmx-users-request at gromacs.org.
> Can't post? Read http://www.gromacs.org/Support/Mailing_Lists