Rilei,
You do not need to know anything about group files to run MMPBSA.py. The
group files are set up within the script so the user does not need to do
anything. If you are curious about their setup and execution you can find
information in the Amber Manual about using multisander and group files. The
problem with your command is that you do not need to place "mpirun -np 3"
before "/home/ryu/amber10/bin/MMPBSA.py" when executing MMPBSA.py. If you
just execute MMPBSA.py using "/home/ryu/amber10/bin/MMPBSA.py -i
mmpbsa.in-sp ..." and set the mpi_cmd within the mmpbsa input file
(which you have
already done), then the script will automatically use the correct number of
processors if they are available. The script itself is not parallel, but it
calls programs (like sander.MPI) that are parallel.
Good luck!
-Bill
>
On Sun, Apr 4, 2010 at 6:13 AM, Rilei Yu <yulaomao1983.yahoo.com.cn> wrote:
> Hi Dr. Jason,
>
> My input mmpbsa as following:
> MMPBSA input file to run gb
> &general
> startframe=50, endframe=100, interval=10, verbose=0,
> ligand_mask=':421-432', receptor_mask=':1-210:211-420',
> initial_traj=0, mpi_cmd='mpirun -np 3', nproc=3
> /
> &gb
> saltcon=0.1, igb=5,
> /
> &pb
>
> /
> My input:(command, that is also used for MD, working well)
> lamboot
> mpirun -np 3 /home/ryu/amber10/bin/MMPBSA.py -i mmpbsa.in -sp
> complex_solvated.prmtop -cp complex.prmtop -rp re.prmtop -lp lig.prmtop -y
> complex_md2.mdcrd
> You said : since parallelization in MMPBSA.py operates via sander
> groupfiles"
> Can you give me an example to show the format of the "sander groupfiles"?
> This is my first time to run MMPBSA using parallel way.
>
> Thanks very much!
>
> Rilei Yu
>
> --- 10年4月4日，周日, Jason Swails <jason.swails.gmail.com> 写道：
>
> 发件人: Jason Swails <jason.swails.gmail.com>
> 主题: Re: [AMBER] MMPBSA-ON-CLUSTER
> 收件人: "AMBER Mailing List" <amber.ambermd.org>
> 日期: 2010年4月4日,周日,下午12:18
>
> Hello,
>
> What does your input file look like? The call to sander in MMPBSA.py
> does not appear to be starting the right number of threads. What does
> your mmpbsa input file look like?
>
> Using MPI, you may start as many threads as you please. However, if
> you start more threads than processors, some processors will have to
> pick up multiple threads, which forces the processor to work much
> harder than it should and can rather severely hurt performance. That
> being said, MMPBSA.py is only starting 1 thread for your calculation,
> and since parallelization in MMPBSA.py operates via sander groupfiles,
> you get the error
>
> Error: specified more groups ( 2 ) than the number of
> processors ( 1 ) !
>
> Good luck!
> Jason
>
> On Sat, Apr 3, 2010 at 11:42 PM, Rilei Yu <yulaomao1983.yahoo.com.cn>
> wrote:
> > The first problem also has been solved! But I still do not know how to
> set up the number of processors on the cluster of our institute! How can I
> check it (the number that can be available)? But I can run sander.MPI using
> 8 cpu work well for MD, now I am a bit of confused!
> >
> > Thanks,
> >
> > Rilei Yu
> >
> > --- 10年4月4日，周日, Bill Miller III <brmilleriii.gmail.com> 写道：
> >
> > 发件人: Bill Miller III <brmilleriii.gmail.com>
> > 主题: Re: [AMBER] MMPBSA-ON-CLUSTER
> > 收件人: "AMBER Mailing List" <amber.ambermd.org>
> > 日期: 2010年4月4日,周日,上午10:09
> >
> > If you are still getting the "unknown flag -y" error, then Amber wasn't
> > compiled correctly after the bug fixes were applied. Make sure that you
> > 'make clean' before you do the installation steps for serial and
> parallel.
> > It also looks like you are trying to run MMPBSA.py in parallel by asking
> for
> > multiple processors when you only have one available (Error: specified
> more
> > groups ( 2 ) than the number of processors ( 1 ) !).
> >
> > -Bill
> >
> > On Sat, Apr 3, 2010 at 9:56 PM, Rilei Yu <yulaomao1983.yahoo.com.cn>
> wrote:
> >
> >> Thanks for your answer,
> >>
> >> Based on your suggestion I re-installed the amber10 and this time I
> patched
> >> the bugfix.all.
> >> When I run it, the first problem ( mdfil: Error unknown flag: -y ) is
> >> still here. This is to say I still did not compile correctly?
> >>
> >> Thanks for your help again!
> >>
> >> r.yu.euphoria:~/gid-d4-amber10$ MMPBSA.py -i mmpbsa.in -sp
> >> complex_solvated.prmtop -cp complex.prmtop -rp re.prmtop -lp lig.prmtop
> -y
> >> complex_md2.mdcrd
> >> ptraj found! Using /home/ryu/amber10/bin/ptraj
> >> sander.MPI found! Using /home/ryu/amber10/bin/sander.MPI
> >>
> >> Preparing trajectories with ptraj...
> >> 6 frames were read in and processed by ptraj for use in calculation.
> >>
> >> Starting sander calls
> >>
> >> Starting gb calculation...
> >>
> >> Error: specified more groups ( 2 ) than the number of
> processors
> >> ( 1 ) !
> >> Error: specified more groups ( 2 ) than the number of
> processors
> >> ( 1 ) !
> >>
> >> mdfil: Error unknown flag:
> >> -y
> >>
> >> usage: sander [-O|A] -i mdin -o mdout -p prmtop -c inpcrd -r
> restrt
> >> [-ref refc -x mdcrd -v mdvel -e mden -idip inpdip
> -rdip
> >> rstdip -mdip mddip -ng numgroup -remlog remlog -rem [0|1|2] -inf mdinfo
> >> -radii radii]
> >> Consult the manual for additional options.
> >> Starting pb calculation...
> >>
> >> Error: specified more groups ( 2 ) than the number of
> processors
> >> ( 1 ) !
> >> Error: specified more groups ( 2 ) than the number of
> processors
> >> ( 1 ) !
> >>
> >> mdfil: Error unknown flag:
> >> -y
> >>
> >> usage: sander [-O|A] -i mdin -o mdout -p prmtop -c inpcrd -r
> restrt
> >> [-ref refc -x mdcrd -v mdvel -e mden -idip inpdip
> -rdip
> >> rstdip -mdip mddip -ng numgroup -remlog remlog -rem [0|1|2] -inf mdinfo
> >> -radii radii]
> >> Consult the manual for additional options.
> >> Error: Sander error! GB mdout file not created.
> >> NOTE: All files have been retained for debugging purposes. Type
> MMPBSA.py
> >> --clean to erase these files.
> >>
> >>
> >> Rilei Yu
> >>
> >> --- 10年4月3日，周六, Jason Swails <jason.swails.gmail.com> 写道：
> >>
> >> 发件人: Jason Swails <jason.swails.gmail.com>
> >> 主题: Re: [AMBER] MMPBSA-ON-CLUSTER
> >> 收件人: "AMBER Mailing List" <amber.ambermd.org>
> >> 日期: 2010年4月3日,周六,下午9:43
> >>
> >> Also, do not put mpirun -np 4 before MMPBSA.py. You do not want to
> >> start 4 threads of this process (since MMPBSA.py itself uses "mpi_cmd"
> >> inside the script, you'll wind up with 16 threads on this node, all of
> >> which will be overwriting each other and in general causing chaos!).
> >> For the rest of the issues, see Bill's email.
> >>
> >> On Sat, Apr 3, 2010 at 8:15 AM, Bill Miller III <brmilleriii.gmail.com>
> >> wrote:
> >> > There appears to be two different problems here. First, the "unknown
> >> flag:
> >> > -y" error means that sander is not recognizing the -y flag that is
> used
> >> for
> >> > post-processing of trajectory files. You should apply the bug fixes
> for
> >> > amber 10 (http://ambermd.org/bugfixes10.html) and recompile AMBER.
> There
> >> was
> >> > a bug fix (bugfix.4, I believe) that fixes this error. The second
> error,
> >> > "Error! Ptraj failed. Check coordinate and topology files for the
> >> complex."
> >> > is more ambiguous and it is difficult to see what is wrong given the
> >> > information you have provided. However, I noticed that you have
> specified
> >> > ligand_mask and receptor_mask. You should not have to do this because
> >> there
> >> > is a built-in mask guesser that should correctly choose these masks
> (and
> >> you
> >> > can check in the output to see that it was correct). Also, you
> requested
> >> > analysis on 50 frames, but ptraj only successfully read in 6 frames,
> >> which
> >> > means there is some prmtop/mdcrd inconsistency. Since you set
> >> > initial_traj=1, then your trajectory file has no waters or ions,
> correct?
> >> > Did you give MMPBSA.py a solvated prmtop by chance? This would explain
> >> the
> >> > discrepancy in the requested frames versus the output frames.
> >> >
> >> > Good luck!
> >> >
> >> > -Bill
> >> >
> >> > On Sat, Apr 3, 2010 at 6:17 AM, Rilei Yu <yulaomao1983.yahoo.com.cn>
> >> wrote:
> >> >
> >> >> Dear amber users,
> >> >>
> >> >> All the time I run MMPBSA on my
> >> >> private computer, but this time I try to use cluster. But it does not
> >> >> work. I have modified my input file based on the tutorial. But it
> still
> >> >> does not work. Can anyone here help me?
> >> >> Input file:
> >> >> MMPBSA input file to run gb
> >> >> &general
> >> >> startframe=50, endframe=100, interval=10, verbose=0,
> >> >> ligand_mask=':421-432', receptor_mask=':1-210:211-420',
> >> >> initial_traj=1, mpi_cmd='mpirun -np 4',
> >> >> nproc=4,
> >> >> /
> >> >> &gb
> >> >> saltcon=0.1, igb=5,
> >> >> /
> >> >> &pb
> >> >>
> >> >> /
> >> >>
> >> >> Errors:
> >> >> r.yu.euphoria:~/gid-d4-amber10$
> >> >> mpirun -np 4 MMPBSA.py -i mmpbsa.in -sp complex_solvated.prmtop -cp
> >> >> complex.prmtop -rp re.prmtop -lp lig.prmtop -y complex_md2.mdcrd
> >> >> ptraj found! Using /home/ryu/amber10/bin/ptraj
> >> >> sander.MPI found! Using /home/ryu/amber10/bin/sander.MPI
> >> >> ptraj found! Using /home/ryu/amber10/bin/ptraj
> >> >> sander.MPI found! Using /home/ryu/amber10/bin/sander.MPI
> >> >> ptraj found! Using /home/ryu/amber10/bin/ptraj
> >> >> sander.MPI found! Using /home/ryu/amber10/bin/sander.MPI
> >> >> ptraj found! Using /home/ryu/amber10/bin/ptraj
> >> >> sander.MPI found! Using /home/ryu/amber10/bin/sander.MPI
> >> >>
> >> >> Preparing trajectories with ptraj...
> >> >>
> >> >> Preparing trajectories with ptraj...
> >> >>
> >> >> Preparing trajectories with ptraj...
> >> >>
> >> >> Preparing trajectories with ptraj...
> >> >> 6 frames were read in and processed by ptraj for use in calculation.
> >> >>
> >> >> Starting sander calls
> >> >>
> >> >> Starting gb calculation...
> >> >>
> >> >> Error! Ptraj failed. Check coordinate and topology files for the
> >> complex.
> >> >> NOTE: All files have been retained for debugging purposes. Type
> >> MMPBSA.py
> >> >> --clean to erase these files.
> >> >> Error! Ptraj failed. Check coordinate and topology files for the
> >> complex.
> >> >> NOTE: All files have been retained for debugging purposes. Type
> >> MMPBSA.py
> >> >> --clean to erase these files.
> >> >> Error! Ptraj failed. Check coordinate and topology files for the
> >> complex.
> >> >> NOTE: All files have been retained for debugging purposes. Type
> >> MMPBSA.py
> >> >> --clean to erase these files.
> >> >>
> >> >> mdfil: Error
> >> >> unknown flag:
> >> >> -y
> >> >>
> >> >> usage: sander [-O|A] -i mdin -o mdout -p prmtop -c inpcrd -r
> >> restrt
> >> >>
> >> >> [-ref refc -x mdcrd -v mdvel -e mden -idip inpdip -rdip rstdip -mdip
> >> >> mddip -ng numgroup -remlog rem.log -remtype rem.type -rem [0|1|2]
> >> >> -rremd [0|1|2|3] -saveene saveene -clusterinfo cluster.info-reservoir
> >> >> reserv/frame -hybridtraj hybrid.strip.crd-ng numgroup -remlog remlog
> >> >> -rem [0|1|2] -inf mdinfo -radii radii]
> >> >> Consult the manual for additional options.
> >> >> [euphoria.imb.uq.edu.au:26298] MPI_ABORT invoked on rank 0 in
> >> communicator
> >> >> MPI_COMM_WORLD with errorcode 1
> >> >>
> >> >> mdfil: Error unknown flag:
> >> >>
> >> >> -y
> >> >> I am really appreciated for your help!
> >> >>
> >> >> Rilei Yu
> >> >>
> >> >>
> >> >>
> >> >> _______________________________________________
> >> >> 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
> >> >
> >>
> >>
> >>
> >> --
> >> ---------------------------------------
> >> Jason M. Swails
> >> Quantum Theory Project,
> >> University of Florida
> >> Ph.D. Graduate Student
> >> 352-392-4032
> >>
> >> _______________________________________________
> >> 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
> >>
> > _______________________________________________
> > 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
> >
>
>
>
> --
> ---------------------------------------
> Jason M. Swails
> Quantum Theory Project,
> University of Florida
> Ph.D. Graduate Student
> 352-392-4032
>
> _______________________________________________
> 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
>
_______________________________________________
AMBER mailing list
AMBER.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber
Received on Sun Apr 04 2010 - 06:00:07 PDT
