Re: [AMBER] Problem with parallel install

From: Daniel Roe <daniel.r.roe.gmail.com>
Date: Thu, 21 Sep 2017 11:01:37 -0400

One more thing to check - make sure that you are using the MPI that
you compiled inside the Amber directory, i.e. 'which mpirun` should
return mpirun in $AMBERHOME/bin. This should be the case if you're
sourcing $AMBERHOME/amber.sh.

-Dan

On Thu, Sep 21, 2017 at 9:54 AM, Maria Clelia Milletti
<mmilletti.emich.edu> wrote:
> Yes, I did
> $ export DO_PARALLEL='mpirun -n 2'
> before 'make test'.
> I've used the same procedure for two other MacPro's and it worked fine, so
> I'm not sure what went wrong with this machine.
> I've attached the config.log file for mpich.3-2. It exits 0, so I thought
> everything was fine.
> Is there anything else I can look at to see where the problem originates?
>
> Maria C. Milletti
> Professor of Chemistry
> 503A Science Complex
> Eastern Michigan University
> (734) 487-1183
>
> On Wed, Sep 20, 2017 at 7:49 PM, Bill Ross <ross.cgl.ucsf.edu> wrote:
>
>> Looks like your MPI environment isn't set up. I've never used it myself.
>>
>> Fatal error in MPI_Init: Other MPI error, error stack:
>> MPIR_Init_thread(474)..............:
>> MPID_Init(190).....................: channel initialization failed
>> MPIDI_CH3_Init(89).................:
>> MPID_nem_init(320).................:
>> MPID_nem_tcp_init(173).............:
>> MPID_nem_tcp_get_business_card(420):
>> MPID_nem_tcp_init(379).............: gethostbyname failed,
>> MCMErwin.local (errno 1)
>> ./Run.sff: Program error
>>
>> Maybe if you did
>>
>> $ export DO_PARALLEL='mpirun -n 2'
>>
>> before 'make test', it would work.
>>
>> On 9/20/17 1:54 PM, Maria Clelia Milletti wrote:
>> > attached.
>> >
>> > Maria C. Milletti
>> > Professor of Chemistry
>> > 503A Science Complex
>> > Eastern Michigan University
>> > (734) 487-1183
>> >
>> > On Wed, Sep 20, 2017 at 4:44 PM, Bill Ross <ross.cgl.ucsf.edu> wrote:
>> >
>> >> I would have to see the output.
>> >>
>> >> Bill
>> >>
>> >>
>> >> On 9/20/17 1:13 PM, Maria Clelia Milletti wrote:
>> >>> it seems to work fine. I've attached the test.out file.
>> >>> Any thoughts as to why when I make test after ./configure -mpi gnu
>> every
>> >>> test fails?
>> >>> thanks,
>> >>>
>> >>> Maria C. Milletti
>> >>> Professor of Chemistry
>> >>> 503A Science Complex
>> >>> Eastern Michigan University
>> >>> (734) 487-1183
>> >>>
>> >>> On Wed, Sep 20, 2017 at 3:33 PM, Bill Ross <ross.cgl.ucsf.edu> wrote:
>> >>>
>> >>>> Try with a semicolon:
>> >>>>
>> >>>> $ DO_PARALLEL='mpirun -n 2' ; ./RunTest.sh
>> >>>>
>> >>>> Bill
>> >>>>
>> >>>>
>> >>>> On 9/20/17 12:30 PM, Maria Clelia Milletti wrote:
>> >>>>> It does not. The command returns what looks like a prompt (>) and
>> stays
>> >>>>> suspended that way, but no output is written to test.out in the
>> >>>>> Test_Parallel folder.
>> >>>>>
>> >>>>> Maria C. Milletti
>> >>>>> Professor of Chemistry
>> >>>>> 503A Science Complex
>> >>>>> Eastern Michigan University
>> >>>>> (734) 487-1183
>> >>>>>
>> >>>>> On Tue, Sep 19, 2017 at 1:26 PM, Daniel Roe <daniel.r.roe.gmail.com>
>> >>>> wrote:
>> >>>>>> OK - in that case the next step would be to try testing an
>> individual
>> >>>>>> program.
>> >>>>>>
>> >>>>>> $ cd $AMBERHOME/AmberTools/src/cpptraj/test/Test_Parallel
>> >>>>>> $ DO_PARALLEL='mpirun -n 2' ./RunTest.sh
>> >>>>>>
>> >>>>>> See if that succeeds.
>> >>>>>>
>> >>>>>> -Dan
>> >>>>>>
>> >>>>>> On Tue, Sep 19, 2017 at 11:53 AM, Maria Clelia Milletti
>> >>>>>> <mmilletti.emich.edu> wrote:
>> >>>>>>> I do get
>> >>>>>>> Hello
>> >>>>>>> Hello
>> >>>>>>>
>> >>>>>>> Maria C. Milletti
>> >>>>>>> Professor of Chemistry
>> >>>>>>> 503A Science Complex
>> >>>>>>> Eastern Michigan University
>> >>>>>>> (734) 487-1183
>> >>>>>>>
>> >>>>>>> On Tue, Sep 19, 2017 at 10:42 AM, Daniel Roe <
>> daniel.r.roe.gmail.com
>> >>>>>> wrote:
>> >>>>>>>> The errors in your log have to do with MPI, not with Amber itself.
>> >> Are
>> >>>>>>>> you able to use MPI to run a program at all? E.g.
>> >>>>>>>>
>> >>>>>>>> mpirun -n 2 echo Hello
>> >>>>>>>>
>> >>>>>>>> If the output is not just:
>> >>>>>>>>
>> >>>>>>>> Hello
>> >>>>>>>> Hello
>> >>>>>>>>
>> >>>>>>>> then there is something wrong with the MPI installation on your
>> >>>> machine.
>> >>>>>>>> -Dan
>> >>>>>>>>
>> >>>>>>>> On Tue, Sep 19, 2017 at 10:35 AM, Maria Clelia Milletti
>> >>>>>>>> <mmilletti.emich.edu> wrote:
>> >>>>>>>>> Thank you, that worked.
>> >>>>>>>>> I am now trying to install the parallel version on a different
>> >>>>>> machine,
>> >>>>>>>>> older but also a 12-core MacPro running OS 12.10.6
>> >>>>>>>>> The ./configure -mpi gnu in $AMBERHOME is successful, but when I
>> >>>> 'make
>> >>>>>>>>> test', every test fails. I have attached the diff and log files.
>> >>>>>>>>> Any help is greatly appreciated.
>> >>>>>>>>>
>> >>>>>>>>> Maria C. Milletti
>> >>>>>>>>> Professor of Chemistry
>> >>>>>>>>> 503A Science Complex
>> >>>>>>>>> Eastern Michigan University
>> >>>>>>>>> (734) 487-1183
>> >>>>>>>>>
>> >>>>>>>>> On Mon, Sep 18, 2017 at 4:33 PM, David A Case <
>> >>>> david.case.rutgers.edu
>> >>>>>>>>> wrote:
>> >>>>>>>>>
>> >>>>>>>>>> On Mon, Sep 18, 2017, Maria Clelia Milletti wrote:
>> >>>>>>>>>>
>> >>>>>>>>>>> I'm having a problem installing the parallel version of
>> >>>>>>>>>>> Amber16/AmberTools17 on a 12-core Intel XeonE5 MacPro (OS
>> >> 10.12.6).
>> >>>>>>>>>>> The serial installation was successful. I then downloaded
>> >> mpich-3.2
>> >>>>>>>> and
>> >>>>>>>>>>> installed it from the AmberTools/src directory using the
>> command
>> >>>>>>>>>>> ./configure_mpich gnu
>> >>>>>>>>>>> When I try to 'make test' I get the error
>> >>>>>>>>>>> No rule to make target 'test'.
>> >>>>>>>>>> What directory are you in when you issue this command? Note
>> that,
>> >>>>>> after
>> >>>>>>>>>> running configure_mpich, you need to go back to $AMBERHOME,
>> re-run
>> >>>>>>>>>> configure
>> >>>>>>>>>> with the "-mpi" option, then type "make install" followed by
>> "make
>> >>>>>>>> test".
>> >>>>>>>>>> ....dac
>> >>>>>>>>>>
>> >>>>>>>>>>
>> >>>>>>>>>> _______________________________________________
>> >>>>>>>>>> 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
>> >>>>>>>>>
>> >>>>>>>> --
>> >>>>>>>> -------------------------
>> >>>>>>>> Daniel R. Roe
>> >>>>>>>> Laboratory of Computational Biology
>> >>>>>>>> National Institutes of Health, NHLBI
>> >>>>>>>> 5635 Fishers Ln, Rm T900
>> >>>>>>>> Rockville MD, 20852
>> >>>>>>>> https://www.lobos.nih.gov/lcb
>> >>>>>>>>
>> >>>>>>>> _______________________________________________
>> >>>>>>>> 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
>> >>>>>> --
>> >>>>>> -------------------------
>> >>>>>> Daniel R. Roe
>> >>>>>> Laboratory of Computational Biology
>> >>>>>> National Institutes of Health, NHLBI
>> >>>>>> 5635 Fishers Ln, Rm T900
>> >>>>>> Rockville MD, 20852
>> >>>>>> https://www.lobos.nih.gov/lcb
>> >>>>>>
>> >>>>>> _______________________________________________
>> >>>>>> 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
>> >> _______________________________________________
>> >> 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
>



-- 
-------------------------
Daniel R. Roe
Laboratory of Computational Biology
National Institutes of Health, NHLBI
5635 Fishers Ln, Rm T900
Rockville MD, 20852
https://www.lobos.nih.gov/lcb
_______________________________________________
AMBER mailing list
AMBER.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber
Received on Thu Sep 21 2017 - 08:30:02 PDT
Custom Search