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

From: Ray Luo <rluo.uci.edu>
Date: Sat, 5 Mar 2016 11:16:02 -0800

I guess most of the time, the dependence was satisfied ... Did this
happen in the middle of making sander or something else?

Ray
--
Ray Luo, Ph.D.
Professor
Biochemistry, Molecular Biophysics, Chemical Physics,
Chemical and Biomedical Engineering
University of California, Irvine, CA 92697-3900
On Fri, Mar 4, 2016 at 11:36 PM, Hai Nguyen <nhai.qn.gmail.com> wrote:
> 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
_______________________________________________
AMBER-Developers mailing list
AMBER-Developers.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber-developers
Received on Sat Mar 05 2016 - 11:30:03 PST
Custom Search