Re: [AMBER-Developers] Sander parallel build broken by PBSA updates

From: Hai Nguyen <nhai.qn.gmail.com>
Date: Sat, 5 Mar 2016 02:36:50 -0500

FYI: I've just tried "*make install -j24"* in another machine and got error

"Fatal Error: Can't rename module file 'pbtimer_module.mod0' to
'pbtimer_module.mod': No such file or directory
make[2]: *** [timer.o] Error 1
make[2]: *** Waiting for unfinished jobs....
make[2]: Leaving directory
`/gpfs/gpfs/project1/dacase-001/haichit/amber/amber/AmberTools/src/pbsa'
make[1]: *** [serial] Error 2
make[1]: Leaving directory
`/gpfs/gpfs/project1/dacase-001/haichit/amber/amber/AmberTools/src'
make: *** [install] Error 2
"

info: gcc 4.4.7, Linux, 24 cores/node; use "*git clean -fdx .*" then
"*./configure
-noX11 gnu*" then "*make install -j24*"; git
commit: 570e427eaf02a99d81e6bf1c36d7d3a51a27c529 (today)

*However*, this error seems happened randomly, sometimes the install
finished without any errors.

Hai

On Thu, Feb 18, 2016 at 3:54 PM, Hai Nguyen <nhai.qn.gmail.com> wrote:

> I just tried "make -j 8 install" for serial build and ok with my machine
> too.
>
> Hai
>
> On Thu, Feb 18, 2016 at 3:36 PM, Ray Luo <rluo.uci.edu> wrote:
>
>> Gerald,
>>
>> Your fix also worked on my side! Thanks a lot!
>>
>> I used "make -j 8 install" though I skipped python install in both the
>> serial and parallel tests.
>>
>> All the best,
>> Ray
>> --
>> Ray Luo, Ph.D.
>> Professor
>> Biochemistry, Molecular Biophysics, Chemical Physics,
>> Chemical and Biomedical Engineering
>> University of California, Irvine, CA 92697-3900
>>
>>
>> On Thu, Feb 18, 2016 at 11:23 AM, Ray Luo <rluo.uci.edu> wrote:
>> > Thanks, I'm downloading your fix ...
>> >
>> > Also, the python-related installation seems to mess up the dependence
>> > issue whether I use single or multi-thread make ... I don't know why
>> > my changes in pbsa and sander have caused this to break.
>> >
>> > All the best,
>> > Ray
>> > --
>> > Ray Luo, Ph.D.
>> > Professor
>> > Biochemistry, Molecular Biophysics, Chemical Physics,
>> > Chemical and Biomedical Engineering
>> > University of California, Irvine, CA 92697-3900
>> >
>> >
>> > On Thu, Feb 18, 2016 at 11:08 AM, Gerald Monard
>> > <Gerald.Monard.univ-lorraine.fr> wrote:
>> >> Hi,
>> >>
>> >> I think that I found a fix. The problem was in the makedepend program
>> in
>> >> pbsa/ that had not been updated. That should work now. Can anybody
>> test?
>> >>
>> >> Best,
>> >>
>> >> Gerald.
>> >>
>> >> On 02/18/2016 07:01 PM, Ross Walker wrote:
>> >>> I second this - I tried to disable PBSA building for now but of
>> course it is linked into Sander etc so we'd disabling Sander as well.
>> >>>
>> >>> I'd suggest spending a few hours trying to fix it and if you can't
>> the checkin should be reversed until this is fixed.
>> >>>
>> >>> Serial builds really aren't a viable option these days - try doing a
>> serial build on a Xeon Phi and you'll understand. :-(
>> >>>
>> >>> All the best
>> >>> Ross
>> >>>
>> >>>> On Feb 18, 2016, at 9:13 AM, Daniel Roe <daniel.r.roe.gmail.com>
>> wrote:
>> >>>>
>> >>>> On Wed, Feb 17, 2016 at 6:04 PM, Ray Luo, Ph.D. <ray.luo.uci.edu>
>> wrote:
>> >>>>> Ross,
>> >>>>>
>> >>>>> Please pull again to see whether it fixed the build problem on your
>> >>>>> side. Looks like multi-thread make does not work ... Too many
>> >>>>> dependence issues ...
>> >>>>
>> >>>> The multi-threaded build definitely doesn't work for me when I hit
>> pbsa stuff:
>> >>>>
>> >>>> $ make -j6 install
>> >>>> ...
>> >>>> make[2]: Entering directory
>> '/home/droe/Amber/GIT/amber/AmberTools/src/pbsa'
>> >>>> gcc -c -O3 -mtune=native -fPIC ...
>> >>>> ...
>> >>>> Makefile:238: recipe for target 'pb_p3m.LIBPBSA.o' failed
>> >>>> Makefile:238: recipe for target 'pb_fdfrc.LIBPBSA.o' failed
>> >>>>
>> >>>> Amber is pretty large nowadays, and not being able to build in
>> >>>> parallel is a real drawback. We can currently get around this by
>> >>>> adding a '.NOTPARALLEL' target to the pbsa Makefile (or forcing -j1),
>> >>>> but ideally I think the multi-threaded build should be fixed.
>> >>>>
>> >>>> -Dan
>> >>>>
>> >>>>>
>> >>>>> All the best,
>> >>>>> Ray
>> >>>>> --
>> >>>>> Ray Luo, Ph.D.
>> >>>>> Professor
>> >>>>> Biochemistry, Molecular Biophysics, Chemical Physics,
>> >>>>> Chemical and Biomedical Engineering
>> >>>>> University of California, Irvine, CA 92697-3900
>> >>>>>
>> >>>>>
>> >>>>> On Wed, Feb 17, 2016 at 12:59 PM, Ray Luo, Ph.D. <ray.luo.uci.edu>
>> wrote:
>> >>>>>> Okay, I think I've fixed it. Don't know why the same makefile
>> works in
>> >>>>>> amber15 but doesn't work in amber16. I'm testing all the
>> >>>>>> sander/pbsa/nab combs to make sure all can build without
>> interruption
>> >>>>>> and also pass the tests.
>> >>>>>>
>> >>>>>> Ray
>> >>>>>> --
>> >>>>>> Ray Luo, Ph.D.
>> >>>>>> Professor
>> >>>>>> Biochemistry, Molecular Biophysics, Chemical Physics,
>> >>>>>> Chemical and Biomedical Engineering
>> >>>>>> University of California, Irvine, CA 92697-3900
>> >>>>>>
>> >>>>>>
>> >>>>>> On Tue, Feb 16, 2016 at 10:34 PM, Ray Luo, Ph.D. <ray.luo.uci.edu>
>> wrote:
>> >>>>>>> Finally I can reproduce your problem in building the MPI sander
>> ...
>> >>>>>>>
>> >>>>>>> Will get it fixed tomorrow morning ...
>> >>>>>>>
>> >>>>>>> Ray
>> >>>>>>> --
>> >>>>>>> Ray Luo, Ph.D.
>> >>>>>>> Professor
>> >>>>>>> Biochemistry, Molecular Biophysics, Chemical Physics,
>> >>>>>>> Chemical and Biomedical Engineering
>> >>>>>>> University of California, Irvine, CA 92697-3900
>> >>>>>>>
>> >>>>>>>
>> >>>>>>> On Tue, Feb 16, 2016 at 10:17 PM, Ray Luo, Ph.D. <ray.luo.uci.edu>
>> wrote:
>> >>>>>>>> Hi Ross,
>> >>>>>>>>
>> >>>>>>>> Looks like that there is no extra untracked file at least in all
>> the
>> >>>>>>>> src folders, though there were indeed two .swp files in the test
>> >>>>>>>> cases. So I think lack of "git clean -fdx" is not the problem.
>> I've
>> >>>>>>>> finished the serial build and is running "test.sander.BASIC".
>> Looking
>> >>>>>>>> good, so far.
>> >>>>>>>>
>> >>>>>>>> I'll do the MPI build next after another "git clean -fdx" ...
>> >>>>>>>>
>> >>>>>>>> However, I think one possible reason could be the use of
>> multi-thread
>> >>>>>>>> in make. There may be inter-dependent issues involved.
>> >>>>>>>>
>> >>>>>>>> I'll do some experiment with the MPI build to see whether this
>> is the cause.
>> >>>>>>>>
>> >>>>>>>> Ray
>> >>>>>>>> --
>> >>>>>>>> Ray Luo, Ph.D.
>> >>>>>>>> Professor
>> >>>>>>>> Biochemistry, Molecular Biophysics, Chemical Physics,
>> >>>>>>>> Chemical and Biomedical Engineering
>> >>>>>>>> University of California, Irvine, CA 92697-3900
>> >>>>>>>>
>> >>>>>>>>
>> >>>>>>>> On Tue, Feb 16, 2016 at 8:24 PM, Ross Walker <
>> ross.rosswalker.co.uk> wrote:
>> >>>>>>>>> Hi Ray,
>> >>>>>>>>>
>> >>>>>>>>> It's not about whether you compiled in that folder - it is
>> whether you maybe edited files in that folder that you never committed /
>> pushed. Try doing:
>> >>>>>>>>>
>> >>>>>>>>> git clean -f -d -x
>> >>>>>>>>> git status
>> >>>>>>>>>
>> >>>>>>>>> And see what it reports.
>> >>>>>>>>>
>> >>>>>>>>> All the best
>> >>>>>>>>> Ross
>> >>>>>>>>>
>> >>>>>>>>>> On Feb 16, 2016, at 23:04, Ray Luo <rluo.uci.edu> wrote:
>> >>>>>>>>>>
>> >>>>>>>>>> Hi Ross,
>> >>>>>>>>>>
>> >>>>>>>>>> Yes, I think my folder is clean since I have never compiled
>> anything
>> >>>>>>>>>> in my amber git folder. Every time after syncing with the
>> master
>> >>>>>>>>>> branch, I copy the whole master branch to a new folder for
>> compiling
>> >>>>>>>>>> and testing.
>> >>>>>>>>>>
>> >>>>>>>>>> I'm about to test this evening's synced version, but will do
>> "git
>> >>>>>>>>>> clean -fxd" before configuring/compiling. Maybe this is why.
>> >>>>>>>>>>
>> >>>>>>>>>> By the way, could you add my uci email to the mailing list as
>> well,
>> >>>>>>>>>> i.e. "ray.luo.uci.edu"?
>> >>>>>>>>>>
>> >>>>>>>>>> All the best,
>> >>>>>>>>>> Ray
>> >>>>>>>>>> --
>> >>>>>>>>>> Ray Luo, Ph.D.
>> >>>>>>>>>> Professor
>> >>>>>>>>>> Biochemistry, Molecular Biophysics, Chemical Physics,
>> >>>>>>>>>> Chemical and Biomedical Engineering
>> >>>>>>>>>> University of California, Irvine, CA 92697-3900
>> >>>>>>>>>>
>> >>>>>>>>>>
>> >>>>>>>>>> On Tue, Feb 16, 2016 at 7:51 PM, Ross Walker <
>> ross.rosswalker.co.uk> wrote:
>> >>>>>>>>>>> Hi Ray,
>> >>>>>>>>>>>
>> >>>>>>>>>>> If I do this - on my OSX laptop (GCC 4.9.3), MPICH version
>> 3.1.4
>> >>>>>>>>>>>
>> >>>>>>>>>>> git pull
>> >>>>>>>>>>> git clean -f -d -x
>> >>>>>>>>>>> git status
>> >>>>>>>>>>> On branch master
>> >>>>>>>>>>> Your branch is up-to-date with 'origin/master'.
>> >>>>>>>>>>> nothing to commit, working directory clean
>> >>>>>>>>>>>
>> >>>>>>>>>>> ./configure -mpi gnu
>> >>>>>>>>>>> make install
>> >>>>>>>>>>>
>> >>>>>>>>>>> ...
>> >>>>>>>>>>> ...
>> >>>>>>>>>>> dnrm2.F90:113.24:
>> >>>>>>>>>>>
>> >>>>>>>>>>> ASSIGN 110 TO NEXT
>> >>>>>>>>>>> 1
>> >>>>>>>>>>> Warning: Deleted feature: ASSIGN statement at (1)
>> >>>>>>>>>>> mpif90 -DBINTRAJ -DEMIL -DMPI -c -O3 -mtune=native -fPIC
>> -ffree-form -I/Users/rcw/amber/amber/include
>> -I/Users/rcw/amber/amber/include -o dcopy.o dcopy.F90
>> >>>>>>>>>>> mpif90 -DBINTRAJ -DEMIL -DMPI -c -O3 -mtune=native -fPIC
>> -ffree-form -I/Users/rcw/amber/amber/include
>> -I/Users/rcw/amber/amber/include -o pb_force.o pb_force.F90
>> >>>>>>>>>>> pb_force.F90:292.77:
>> >>>>>>>>>>>
>> >>>>>>>>>>> sa_init, sa_driver,
>> sa_free, sa_free_mb, &
>> >>>>>>>>>>>
>> 1
>> >>>>>>>>>>> Error: Symbol 'saslave_init' referenced at (1) not found in
>> module 'solvent_accessibility'
>> >>>>>>>>>>> make[3]: *** [pb_force.o] Error 1
>> >>>>>>>>>>> make[2]: *** [libpbsa] Error 2
>> >>>>>>>>>>> make[1]: *** [parallel] Error 2
>> >>>>>>>>>>> make: *** [install] Error 2
>> >>>>>>>>>>>
>> >>>>>>>>>>> Same problem. Are you sure you have the latest tree and it is
>> clean and doesn't have any modified files locally or files you forgot to
>> add?
>> >>>>>>>>>>>
>> >>>>>>>>>>> All the best
>> >>>>>>>>>>> Ross
>> >>>>>>>>>>>
>> >>>>>>>>>>>
>> >>>>>>>>>>>> On Feb 16, 2016, at 15:56, Ray Luo, Ph.D. <ray.luo.uci.edu>
>> wrote:
>> >>>>>>>>>>>>
>> >>>>>>>>>>>> Hi Ross,
>> >>>>>>>>>>>>
>> >>>>>>>>>>>> So far I can "configure -mpi gnu", "make install", and can
>> also pass
>> >>>>>>>>>>>> "make sander.BASIC.MPI". I'm running "make test.parallel.MM".
>> I'm
>> >>>>>>>>>>>> using "mpirun -np 4" for the mpi jobs. This is on Rocks
>> 6.1/Centos
>> >>>>>>>>>>>> 6.1. Which step was the problem?
>> >>>>>>>>>>>>
>> >>>>>>>>>>>> However I disabled "pbsa_mbfocus" in AmberTools/test/ since
>> I'm in the
>> >>>>>>>>>>>> process of removing this feature to accommodate incoming
>> CUDA. I also
>> >>>>>>>>>>>> need to update the test cases for sander/nab/mmpbsa ...
>> which will be
>> >>>>>>>>>>>> partially in today for sander/nab. I'm trying to get
>> mmpbsa.py to work
>> >>>>>>>>>>>> with the new python environment.
>> >>>>>>>>>>>>
>> >>>>>>>>>>>> All the best,
>> >>>>>>>>>>>> Ray
>> >>>>>>>>>>>> --
>> >>>>>>>>>>>> Ray Luo, Ph.D.
>> >>>>>>>>>>>> Professor
>> >>>>>>>>>>>> Biochemistry, Molecular Biophysics, Chemical Physics,
>> >>>>>>>>>>>> Chemical and Biomedical Engineering
>> >>>>>>>>>>>> University of California, Irvine, CA 92697-3900
>> >>>>>>>>>>>>
>> >>>>>>>>>>>>
>> >>>>>>>>>>>> On Tue, Feb 16, 2016 at 10:21 AM, Ray Luo, Ph.D. <
>> ray.luo.uci.edu> wrote:
>> >>>>>>>>>>>>> Hi Ross,
>> >>>>>>>>>>>>>
>> >>>>>>>>>>>>> I'm looking into this ... There was a major overhaul of the
>> code in
>> >>>>>>>>>>>>> addition to new features in the last check in.
>> >>>>>>>>>>>>>
>> >>>>>>>>>>>>> All the best,
>> >>>>>>>>>>>>>
>> >>>>>>>>>>>>> Ray
>> >>>>>>>>>>>>>
>> >>>>>>>>>>>>> --
>> >>>>>>>>>>>>> Ray Luo, Ph.D.
>> >>>>>>>>>>>>> Professor
>> >>>>>>>>>>>>> Biochemistry, Molecular Biophysics, Chemical Physics,
>> >>>>>>>>>>>>> Chemical and Biomedical Engineering
>> >>>>>>>>>>>>> University of California, Irvine, CA 92697-3900
>> >>>>>>>>>>>
>> >>>>>>>>>
>> >>>>>
>> >>>>> _______________________________________________
>> >>>>> AMBER-Developers mailing list
>> >>>>> AMBER-Developers.ambermd.org
>> >>>>> http://lists.ambermd.org/mailman/listinfo/amber-developers
>> >>>>
>> >>>>
>> >>>>
>> >>>> --
>> >>>> -------------------------
>> >>>> Daniel R. Roe, PhD
>> >>>> Department of Medicinal Chemistry
>> >>>> University of Utah
>> >>>> 30 South 2000 East, Room 307
>> >>>> Salt Lake City, UT 84112-5820
>> >>>> http://home.chpc.utah.edu/~cheatham/
>> >>>> (801) 587-9652
>> >>>> (801) 585-6208 (Fax)
>> >>>>
>> >>>> _______________________________________________
>> >>>> AMBER-Developers mailing list
>> >>>> AMBER-Developers.ambermd.org
>> >>>> http://lists.ambermd.org/mailman/listinfo/amber-developers
>> >>>
>> >>>
>> >>> _______________________________________________
>> >>> AMBER-Developers mailing list
>> >>> AMBER-Developers.ambermd.org
>> >>> http://lists.ambermd.org/mailman/listinfo/amber-developers
>> >>>
>> >>
>> >> --
>> >>
>> ____________________________________________________________________________
>> >>
>> >> Prof. Gerald MONARD
>> >> SRSMC, Université de Lorraine, CNRS
>> >> Boulevard des Aiguillettes B.P. 70239
>> >> F-54506 Vandoeuvre-les-Nancy, FRANCE
>> >>
>> >> e-mail : Gerald.Monard.univ-lorraine.fr
>> >> tel. : +33 (0)383.684.381
>> >> fax : +33 (0)383.684.371
>> >> web : http://www.monard.info
>> >>
>> >>
>> ____________________________________________________________________________
>> >>
>> >>
>> >> _______________________________________________
>> >> AMBER-Developers mailing list
>> >> AMBER-Developers.ambermd.org
>> >> http://lists.ambermd.org/mailman/listinfo/amber-developers
>>
>> _______________________________________________
>> AMBER-Developers mailing list
>> AMBER-Developers.ambermd.org
>> http://lists.ambermd.org/mailman/listinfo/amber-developers
>>
>
>
_______________________________________________
AMBER-Developers mailing list
AMBER-Developers.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber-developers
Received on Sat Mar 05 2016 - 00:00:02 PST
Custom Search