[gmx-users] Problems building gromacs 4.6

[Mark Abraham]

On Fri, Jun 7, 2013 at 5:22 PM, Colin Bannister <
Colin.Bannister at nottingham.ac.uk> wrote:

> Mark,
>
> Thanks for the response.  We use MKL for a lot of applications, and have
> had no problems, except with gromacs.


Because of the historical pain of specifying the MKL link configuration,
the ease of use of FFTW, and the very small benefit MKL has seemed to
provide, it's never been a priority for GROMACS. The icc -mkl flag makes
life much easier, and future GROMACS versions will use it.

 I wonder if there is a bug in cmake, as surely getting a segmentation
> fault can never be right ?
>

I've never seen CMake segfault.

If anyone else has any ideas, I'd be grateful for them - I will try a build
> with FFTW3 for now.
>

Other than try a new CMake, no.

Mark


>
> Colin
>
> -----Original Message-----
> From: gmx-users-bounces at gromacs.org [mailto:gmx-users-bounces at gromacs.org]
> On Behalf Of Mark Abraham
> Sent: 07 June 2013 15:44
> To: Discussion list for GROMACS users
> Subject: Re: [gmx-users] Problems building gromacs 4.6
>
> On Fri, Jun 7, 2013 at 12:20 PM, colinbannister <
> colin.bannister at nottingham.ac.uk> wrote:
>
> > I am trying to build Gromacs 4.6 on a Linux machine, running
> > Scientific Linux 6.2.  I am doing an MPI build with Intel-MPI, and
> > Intel compilers, and MKL.
> >
> > Commands I issue to set environment, and run cmake (available as
> > cmake28)
> > are:
> >
> > [cczcb at login02 gromacs-build-nocuda]$ module load intel/compiler/64
> > [cczcb at login02 gromacs-build-nocuda]$ module load intel-mpi/64/4.1.0
> > [cczcb at login02 gromacs-build-nocuda]$ module load intel/mkl/64
> > [cczcb at login02 gromacs-build-nocuda]$ export CC=mpicc
> > [cczcb at login02 gromacs-build-nocuda]$ export CXX=mpicxx
> > [cczcb at login02 gromacs-build-nocuda]$ mklvars_intel64.sh
> > [cczcb at login02 gromacs-build-nocuda]$ cmake28 ../gromacs-4.6
> > -DCMAKE_INSTALL_PREFIX=/home/cczcb/gromacs-install -DGMX_MPI=ON
> > -DGMX_FFT_LIBRARY=mkl
> >
> > -DMKL_INCLUDE_DIR=/cm/shared/apps/intel/composer_xe/2011_sp1.11.339/mk
> > l/include
> >
> > -DMKL_LIBRARIES=/cm/shared/apps/intel/composer_xe/2011_sp1.11.339/mkl/
> > lib/intel64/libmkl_intel_lp64.so;/cm/shared/apps/intel/composer_xe/201
> > 1_sp1.11.339/mkl/lib/intel64/libmkl_sequential.so;/cm/shared/apps/inte
> > l/composer_xe/2011_sp1.11.339/mkl/lib/intel64/libmkl_core.so
> >
> > This completes, but gives Segmentation fault messages at the end.
> > Output attached as 'cmake28.out'
> >
> > When I type 'make', it gets to 53% then gives a large number of
> > 'undefined reference' errors, see 'make.out'.
> >
> > I have also tried only putting the libmkl_core.so library path as the
> > setting for MKL_LIBRARIES.  This avoids the Segmentation faults, but
> > again results in many undefined references, as you might expect.
> >
> > Any idea what could be happening here ?  Is Cmake messing up in
> > parsing the command in some way ?  Is there a different approach I
> > could use to build Gromacs ?
> >
>
> All that sounds very weird, and makes me wonder whether an MKL test
> program would compile on your system. Since the value of MKL is very
> limited, I would suggest compiling with FFTW for now.
>
> GROMACS 4.6.2 streamlined the MKL build process greatly for icc >= 11, but
> I would suggest waiting until GROMACS 4.6.3, because of a performance
> regression with MPI builds that we are working on fixing.
>
> Cheers,
>
> Mark
>
>
> >
> > Any assistance much appreciated!
> >
> > Colin
> >
> > cmake28.out
> > <http://gromacs.5086.x6.nabble.com/file/n5008889/cmake28.out>
> > make.out <http://gromacs.5086.x6.nabble.com/file/n5008889/make.out>
> >
> >
> >
> >
> > --
> > View this message in context:
> > http://gromacs.5086.x6.nabble.com/Problems-building-gromacs-4-6-tp5008
> > 889.html Sent from the GROMACS Users Forum mailing list archive at
> > Nabble.com.
> > --
> > gmx-users mailing list    gmx-users at gromacs.org
> > http://lists.gromacs.org/mailman/listinfo/gmx-users
> > * Please search the archive at
> > http://www.gromacs.org/Support/Mailing_Lists/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/Support/Mailing_Lists
> >
> --
> gmx-users mailing list    gmx-users at gromacs.org
> http://lists.gromacs.org/mailman/listinfo/gmx-users
> * Please search the archive at
> http://www.gromacs.org/Support/Mailing_Lists/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/Support/Mailing_Lists
> This message and any attachment are intended solely for the addressee and
> may contain confidential information. If you have received this message in
> error, please send it back to me, and immediately delete it.   Please do
> not use, copy or disclose the information contained in this message or in
> any attachment.  Any views or opinions expressed by the author of this
> email do not necessarily reflect the views of the University of Nottingham.
>
> This message has been checked for viruses but the contents of an attachment
> may still contain software viruses which could damage your computer
> system, you are advised to perform your own checks. Email communications
> with the University of Nottingham may be monitored as permitted by UK
> legislation.
>
>
>
>
> --
> gmx-users mailing list    gmx-users at gromacs.org
> http://lists.gromacs.org/mailman/listinfo/gmx-users
> * Please search the archive at
> http://www.gromacs.org/Support/Mailing_Lists/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/Support/Mailing_Lists
>