[gmx-developers] Release 4.6 fails to compile

[Justin A. Lemkul]

On 6/6/12 8:03 PM, Christoph Junghans wrote:
>>
>> Hi All,
>>
>> I just wanted to point out that based on the latest change to the 4.6
>> release branch (free energy, etc) there is a small bug that breaks
>> compilation for mdrun-gpu.  The problem comes from the changes to idef.h -
>> the introduction of A and B states for bonded parameters causes the
>> following:
>>
>> Building C object src/gmxlib/CMakeFiles/gmx.dir/invblock.c.o
>> In file included from
>> /home/jalemkul/ATHENA/software/openmm-2.0/include/OpenMM.h:43,
>>                  from
>> /home/jalemkul/ATHENA/software/git/release-4-6/src/kernel/openmm_wrapper.cpp:59:
>> /home/jalemkul/ATHENA/software/openmm-2.0/include/openmm/CustomHbondForce.h:
>> In constructor ‘OpenMM::CustomHbondForce::GroupInfo::GroupInfo(int, int,
>> int, const std::vector<double, std::allocator<double>  >&)’:
>> /home/jalemkul/ATHENA/software/openmm-2.0/include/openmm/CustomHbondForce.h:439:
>> warning: ‘OpenMM::CustomHbondForce::GroupInfo::p3’ will be initialized after
>> /home/jalemkul/ATHENA/software/openmm-2.0/include/openmm/CustomHbondForce.h:438:
>> warning:   ‘std::vector<double, std::allocator<double>  >
>> OpenMM::CustomHbondForce::GroupInfo::parameters’
>> /home/jalemkul/ATHENA/software/openmm-2.0/include/openmm/CustomHbondForce.h:442:
>> warning:   when initialized here
>> [  7%] [  7%] Building C object src/gmxlib/CMakeFiles/gmx.dir/macros.c.o
>> /home/jalemkul/ATHENA/software/git/release-4-6/src/kernel/openmm_wrapper.cpp:
>> In function ‘void* openmm_init(FILE*, const char*, t_inputrec*, gmx_mtop_t*,
>> gmx_localtop_t*, t_mdatoms*, t_forcerec*, t_state*)’:
>> /home/jalemkul/ATHENA/software/git/release-4-6/src/kernel/openmm_wrapper.cpp:929:
>> error: ‘struct t_iparams::<anonymous>’ has no member named ‘r13’
>> /home/jalemkul/ATHENA/software/git/release-4-6/src/kernel/openmm_wrapper.cpp:929:
>> error: ‘struct t_iparams::<anonymous>’ has no member named ‘kUB’
>> /home/jalemkul/ATHENA/software/git/release-4-6/src/kernel/openmm_wrapper.cpp:932:
>> error: ‘struct t_iparams::<anonymous>’ has no member named ‘theta’
>> /home/jalemkul/ATHENA/software/git/release-4-6/src/kernel/openmm_wrapper.cpp:932:
>> error: ‘struct t_iparams::<anonymous>’ has no member named ‘ktheta’
>>
>> Adding 'A' to the end of each of these terms solves the problem, though that
>> is a bit of a hack, but would work assuming free energy features are not
>> intended to be run on GPU.  A more elegant solution should probably be
>> found, but I didn't care about elegance when I was trying to compile ;)
>>
>> I just thought I should point this out so that perhaps a fix will be put in
>> place.  I can file this on redmine if necessary so no one forgets.
> Can you push a patch to gerrit?
>

I certainly can, I just wasn't sure what the implications might be.  If the 
above assumption about free energy not mattering on GPU, then it seems to be 
straightforward.

I'll submit the patch; if there are objections they can be voiced on gerrit.  I 
suppose that's where such things should normally occur anyway.

-Justin

> Related question, do we want to keep OpenMM support in 4.6? OpenMM-2.0
> is more than obsolete, they have released 4.1 last month and porting
> Gromacs to it will need some time. I have unsuccessfully tried porting
> Gromacs to OpenMM-3.0 at some point. And the new hybrid kernels will
> use cuda, so GPUs will be supported anyway.
>
> Christoph
>
> 2012/6/6 Justin A. Lemkul<jalemkul at vt.edu>:
>>
>> --
>> ========================================
>>
>> Justin A. Lemkul, Ph.D.
>> Research Scientist
>> Department of Biochemistry
>> Virginia Tech
>> Blacksburg, VA
>> jalemkul[at]vt.edu | (540) 231-9080
>> http://www.bevanlab.biochem.vt.edu/Pages/Personal/justin
>>
>> ========================================
>> --
>> gmx-developers mailing list
>> gmx-developers at gromacs.org
>> http://lists.gromacs.org/mailman/listinfo/gmx-developers
>> Please don't post (un)subscribe requests to the list. Use the www interface
>> or send it to gmx-developers-request at gromacs.org.
>
>
>

-- 
========================================

Justin A. Lemkul, Ph.D.
Research Scientist
Department of Biochemistry
Virginia Tech
Blacksburg, VA
jalemkul[at]vt.edu | (540) 231-9080
http://www.bevanlab.biochem.vt.edu/Pages/Personal/justin

========================================