[gmx-users] Problem configuring openMPI

Manik Mayur manik.mayur at gmail.com
Tue Mar 3 20:35:30 CET 2009


this time upon executing :
./configure\
       --prefix /home/local/gromacs\
       --enable-mpi\
       --program-suffix="_mpi"\
       --disable-nice\
       --enable-shared\
       MPICC=/home/openmpi/mpicc\

it exited with the following error:

../mdlib/.libs/libmd_mpi.so: undefined reference to `al_append'
../mdlib/.libs/libmd_mpi.so: undefined reference to `rw'
../mdlib/.libs/libmd_mpi.so: undefined reference to `lam_deregister_object'
../mdlib/.libs/libmd_mpi.so: undefined reference to `nsend'
../mdlib/.libs/libmd_mpi.so: undefined reference to `dsend'
../mdlib/.libs/libmd_mpi.so: undefined reference to `lam_trpoint'
../mdlib/.libs/libmd_mpi.so: undefined reference to `lam_rtrnamepub'
../mdlib/.libs/libmd_mpi.so: undefined reference to `mrev4'
../mdlib/.libs/libmd_mpi.so: undefined reference to `blktype'
../mdlib/.libs/libmd_mpi.so: undefined reference to `lam_kexit'
../mdlib/.libs/libmd_mpi.so: undefined reference to `ah_free'
../mdlib/.libs/libmd_mpi.so: undefined reference to `lam_ksigretry'
../mdlib/.libs/libmd_mpi.so: undefined reference to `lam_rtrnamedel'
../mdlib/.libs/libmd_mpi.so: undefined reference to `getnodeid'
../mdlib/.libs/libmd_mpi.so: undefined reference to `rpdoom'
../mdlib/.libs/libmd_mpi.so: undefined reference to `al_next'
../mdlib/.libs/libmd_mpi.so: undefined reference to `lam_rtrnameget'
../mdlib/.libs/libmd_mpi.so: undefined reference to `getrent'
../mdlib/.libs/libmd_mpi.so: undefined reference to `getntype'
../mdlib/.libs/libmd_mpi.so: undefined reference to `lam_ksigsetretry'
../mdlib/.libs/libmd_mpi.so: undefined reference to `argvfree'
../mdlib/.libs/libmd_mpi.so: undefined reference to `lam_errorstr'
../mdlib/.libs/libmd_mpi.so: undefined reference to `getroute'
../mdlib/.libs/libmd_mpi.so: undefined reference to `getnodes'
../mdlib/.libs/libmd_mpi.so: undefined reference to `kenter'
../mdlib/.libs/libmd_mpi.so: undefined reference to `lam_ksignal'
../mdlib/.libs/libmd_mpi.so: undefined reference to `lam_register_objects'
../mdlib/.libs/libmd_mpi.so: undefined reference to `bfselect'
../mdlib/.libs/libmd_mpi.so: undefined reference to `lpattach'
../mdlib/.libs/libmd_mpi.so: undefined reference to `ldogetlinks'
../mdlib/.libs/libmd_mpi.so: undefined reference to `al_free'
../mdlib/.libs/libmd_mpi.so: undefined reference to `al_delete'
../mdlib/.libs/libmd_mpi.so: undefined reference to `kexit'
../mdlib/.libs/libmd_mpi.so: undefined reference to `nrecv'
../mdlib/.libs/libmd_mpi.so: undefined reference to `microsleep'
../mdlib/.libs/libmd_mpi.so: undefined reference to `ah_init'
../mdlib/.libs/libmd_mpi.so: undefined reference to `bfrecv'
../mdlib/.libs/libmd_mpi.so: undefined reference to `show_help'
../mdlib/.libs/libmd_mpi.so: undefined reference to `next_prime'
../mdlib/.libs/libmd_mpi.so: undefined reference to `stoi'
../mdlib/.libs/libmd_mpi.so: undefined reference to `kinit'
../mdlib/.libs/libmd_mpi.so: undefined reference to `sfh_sock_accept_tmout'
../mdlib/.libs/libmd_mpi.so: undefined reference to
`sfh_sock_open_srv_inet_stm'
../mdlib/.libs/libmd_mpi.so: undefined reference to `_kio'
../mdlib/.libs/libmd_mpi.so: undefined reference to `ttime'
../mdlib/.libs/libmd_mpi.so: undefined reference to `ah_find'
../mdlib/.libs/libmd_mpi.so: undefined reference to `al_insert'
../mdlib/.libs/libmd_mpi.so: undefined reference to `al_init'
../mdlib/.libs/libmd_mpi.so: undefined reference to `ah_insert'
../mdlib/.libs/libmd_mpi.so: undefined reference to `bfiprobe'
../mdlib/.libs/libmd_mpi.so: undefined reference to `ah_delete'
../mdlib/.libs/libmd_mpi.so: undefined reference to `sfh_sock_set_buf_size'
../mdlib/.libs/libmd_mpi.so: undefined reference to `ah_expand'
../mdlib/.libs/libmd_mpi.so: undefined reference to `al_find'
../mdlib/.libs/libmd_mpi.so: undefined reference to `getorigin'
../mdlib/.libs/libmd_mpi.so: undefined reference to `lam_rtrnamesweep'
../mdlib/.libs/libmd_mpi.so: undefined reference to `_lam_signal'
../mdlib/.libs/libmd_mpi.so: undefined reference to
`sfh_sock_open_clt_inet_stm'
../mdlib/.libs/libmd_mpi.so: undefined reference to `rpstate'
../mdlib/.libs/libmd_mpi.so: undefined reference to `tprintf'
../mdlib/.libs/libmd_mpi.so: undefined reference to `kpause'
../mdlib/.libs/libmd_mpi.so: undefined reference to `ah_next'
../mdlib/.libs/libmd_mpi.so: undefined reference to `rev8'
../mdlib/.libs/libmd_mpi.so: undefined reference to `argvadd'
../mdlib/.libs/libmd_mpi.so: undefined reference to `bfiselect'
../mdlib/.libs/libmd_mpi.so: undefined reference to `lam_rtrstore'
collect2: ld returned 1 exit status
make[1]: *** [mdrun] Error 1

what is the problem now?

On Wed, Mar 4, 2009 at 12:43 AM, Manik Mayur <manik.mayur at gmail.com> wrote:

>
>
> On Wed, Mar 4, 2009 at 12:28 AM, Nicolas <nsapay at ucalgary.ca> wrote:
>
>> Manik Mayur a écrit :
>>
>>>
>>>
>>> On Tue, Mar 3, 2009 at 11:47 PM, Nicolas <nsapay at ucalgary.ca <mailto:
>>> nsapay at ucalgary.ca>> wrote:
>>>
>>>    Manik Mayur a écrit :
>>>
>>>        Hi,
>>>
>>>        If I have 2 mpi environments, namely LAM and openMPI, and I
>>>        want gromacs to use openMPI of the two, what extra option
>>>        should I pass during the ./configure step? In my case openMPI
>>>        is installed locally(in my /home folder).
>>>
>>>    Sorry, I've misunderstood your message. Anyway, I would first
>>>    check twice the paths to the binaries in my script. Then, I would
>>>    check my environment variables, especially PATH, LD_LIBRARY_PATH
>>>    and LD_RUN_PATH. You might need to modify them to make sure any
>>>    Lam-related stuff is loaded by default in your environment.
>>>
>>>
>>> ok, is there any specific option or environment variable through which I
>>> can make gromacs aware of openMPI and not LAM while installation? If I carry
>>> on the default --enable-mpi thing, upon executing the following:
>>>
>> Just a silly question: Did you recompile Gromacs with openMPI? If not, you
>> should with the correct CPPFLAGS, LDFLAGS and MPICC. Typically, you should
>> have something like:
>>
>>   ./configure\
>>        --prefix /home/local/gromacs\
>>        --enable-mpi\
>>        --program-suffix="_mpi"\
>>        --without-x\
>>        --disable-nice\
>>        MPICC=/path/to/mpicc\
>
>
> thanks, I think that is what I needed. Let me try and then I will let you
> know.
>
>
>>        CC=/path/to/gcc
>>        CPPFLAGS=-I/path/to/fftw/include/\
>>        LDFLAGS=-L/path/to/fftw/lib/
>>
>
> Why is FFTW necessary? In all my gromacs installations till now, I have
> avoided this package (out of ignorance).
>
>
>
>> mpicc is located in the bin/ directory of openMPI. gcc is the regular gcc
>> installation (gcc 3). As I said, your issue is probably due to a path
>> uncorrectly pointing to Lam. So check/modify your environment variables,
>> especially the paths to loaded libraries.
>>
>>
>>>     $/home/openmpi/bin/mpirun -np 4 mdrun_mpi -s eql_mpi.tpr -o
>>> eql_mpi.trr
>>>
>>> I get the following error:
>>>
>>> ------------------------------------------------
>>>       It seems that there is no lamd running on this host, which
>>>       indicates
>>>       that the LAM/MPI runtime environment is not operating.  The
>>>       LAM/MPI
>>>       runtime environment is necessary for MPI programs to run (the MPI
>>>       program tired to invoke the "MPI_Init" function).
>>>
>>>       Please run the "lamboot" command the start the LAM/MPI runtime
>>>       environment.  See the LAM/MPI documentation for how to invoke
>>>       "lamboot" across multiple machines.
>>>
>>>   -----------------------------------------------------------------------------
>>> so it is like mdrun_mpi was configured with LAM and hence it is looking
>>> for it.
>>>
>>>    I've tried to use a local installation of openMPI myself, but
>>>    finally asked my sysadmin to install it for all users. I had many
>>>    problems with libraries that couldn't be loaded properly.
>>>
>>>    Cheers,
>>>    Nicolas
>>>
>>>
>>>        Thanks,
>>>        Manik
>>>
>>>
>>>        On Tue, Mar 3, 2009 at 10:58 PM, Nicolas <nsapay at ucalgary.ca
>>>        <mailto:nsapay at ucalgary.ca> <mailto:nsapay at ucalgary.ca
>>>
>>>        <mailto:nsapay at ucalgary.ca>>> wrote:
>>>
>>>           Hi,
>>>
>>>           As mentioned by the error message:
>>>
>>>
>>>             Please run the "lamboot" command the start the LAM/MPI
>>>        runtime
>>>
>>>           Before using Lam, you have to boot it. In my scripts, I've got
>>>           something like:
>>>
>>>             lamboot -v $PBS_NODEFILE
>>>             lamrun -v -np $NPROC mdrun mdargs
>>>             lamhalt
>>>
>>>           I don't use Lam often, though, you might need to use
>>>        different Lam
>>>           options.
>>>
>>>           Nicolas
>>>
>>>           Manik Mayur a écrit :
>>>
>>>               Hi,
>>>
>>>               I have till now successfully used openMPI and gromacs
>>>               together. Recently I got a cluster where I tried to install
>>>               openMPI and gromacs locally as I donot have root
>>>        priviledges
>>>               on it. But it turns out that it already has MPI
>>>        environment on
>>>               it (LAM) which I donot want to use (as the version is quite
>>>               old.. the cluster still has gcc-3.2.3!).
>>>               So after installing openMPI 1.3 and gromacs 4.0.4 on it
>>>        (in my
>>>               /home folder), I tried to execute:
>>>
>>>               $/home/openmpi/bin/mpirun -np 4 mdrun_4_mpi -s
>>>        eql_mpi.tpr -o
>>>               eql_mpi.trr
>>>
>>>               It gives the following error:
>>>
>>>
>>> -----------------------------------------------------------------------------
>>>               It seems that there is no lamd running on this host, which
>>>               indicates
>>>               that the LAM/MPI runtime environment is not operating.  The
>>>               LAM/MPI
>>>               runtime environment is necessary for MPI programs to
>>>        run (the MPI
>>>               program tired to invoke the "MPI_Init" function).
>>>
>>>               Please run the "lamboot" command the start the LAM/MPI
>>>        runtime
>>>               environment.  See the LAM/MPI documentation for how to
>>>        invoke
>>>               "lamboot" across multiple machines.
>>>
>>> -----------------------------------------------------------------------------
>>>
>>>               My question is - Why it still tries to look for LAM
>>>               environment or lamd? I will be extremely thankful if
>>>        somebody
>>>               gives a hint.
>>>
>>>               -Manik
>>>               --        Graduate student
>>>               Microfluidics Lab
>>>               Dept. of Mechanical Engg.
>>>               IIT Kharagpur
>>>
>>> ------------------------------------------------------------------------
>>>
>>>               _______________________________________________
>>>               gmx-users mailing list    gmx-users at gromacs.org
>>>        <mailto:gmx-users at gromacs.org>
>>>               <mailto:gmx-users at gromacs.org
>>>        <mailto:gmx-users at gromacs.org>>
>>>
>>>               http://www.gromacs.org/mailman/listinfo/gmx-users
>>>               Please search the archive at http://www.gromacs.org/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
>>>        <mailto:gmx-users-request at gromacs.org>
>>>               <mailto:gmx-users-request at gromacs.org
>>>        <mailto:gmx-users-request at gromacs.org>>.
>>>
>>>               Can't post? Read
>>>        http://www.gromacs.org/mailing_lists/users.php
>>>
>>>
>>>
>>>           _______________________________________________
>>>           gmx-users mailing list    gmx-users at gromacs.org
>>>        <mailto:gmx-users at gromacs.org>
>>>           <mailto:gmx-users at gromacs.org <mailto:gmx-users at gromacs.org>>
>>>
>>>           http://www.gromacs.org/mailman/listinfo/gmx-users
>>>           Please search the archive at http://www.gromacs.org/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
>>>        <mailto:gmx-users-request at gromacs.org>
>>>           <mailto:gmx-users-request at gromacs.org
>>>        <mailto:gmx-users-request at gromacs.org>>.
>>>
>>>           Can't post? Read
>>> http://www.gromacs.org/mailing_lists/users.php
>>>
>>>
>>>
>>>
>>>        --        Graduate student
>>>        Microfluidics Lab
>>>        Dept. of Mechanical Engg.
>>>        IIT Kharagpur
>>>
>>>  ------------------------------------------------------------------------
>>>
>>>        _______________________________________________
>>>        gmx-users mailing list    gmx-users at gromacs.org
>>>        <mailto:gmx-users at gromacs.org>
>>>        http://www.gromacs.org/mailman/listinfo/gmx-users
>>>        Please search the archive at http://www.gromacs.org/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
>>>        <mailto:gmx-users-request at gromacs.org>.
>>>        Can't post? Read http://www.gromacs.org/mailing_lists/users.php
>>>
>>>
>>>
>>>    _______________________________________________
>>>    gmx-users mailing list    gmx-users at gromacs.org
>>>    <mailto:gmx-users at gromacs.org>
>>>    http://www.gromacs.org/mailman/listinfo/gmx-users
>>>    Please search the archive at http://www.gromacs.org/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
>>>    <mailto:gmx-users-request at gromacs.org>.
>>>    Can't post? Read http://www.gromacs.org/mailing_lists/users.php
>>>
>>>
>>>
>>>
>>> --
>>> Graduate student
>>> Microfluidics Lab
>>> Dept. of Mechanical Engg.
>>> IIT Kharagpur
>>> ------------------------------------------------------------------------
>>>
>>> _______________________________________________
>>> gmx-users mailing list    gmx-users at gromacs.org
>>> http://www.gromacs.org/mailman/listinfo/gmx-users
>>> Please search the archive at http://www.gromacs.org/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/mailing_lists/users.php
>>>
>>
>>
>> _______________________________________________
>> gmx-users mailing list    gmx-users at gromacs.org
>> http://www.gromacs.org/mailman/listinfo/gmx-users
>> Please search the archive at http://www.gromacs.org/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/mailing_lists/users.php
>>
>
>
>
> --
> Graduate student
> Microfluidics Lab
> Dept. of Mechanical Engg.
> IIT Kharagpur
>



-- 
Graduate student
Microfluidics Lab
Dept. of Mechanical Engg.
IIT Kharagpur
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://maillist.sys.kth.se/pipermail/gromacs.org_gmx-users/attachments/20090304/6f0bdffd/attachment.html>


More information about the gromacs.org_gmx-users mailing list