Error in installing VOTCA v1.4.1 on cluster

308 views
Skip to first unread message

Salman Bin Kashif

unread,
Apr 24, 2020, 8:59:03 PM4/24/20
to votca
I am trying to Install Votca on cluster. In have loaded the required modules there. However, I am getting an error in terms of finding the package.I used the following command:

cmake -DCMAKE_INSTALL_PREFIX=${prefix} -DWITH_FFTW=OFF -DWITH_SQLITE3=OFF ..

The error message I get is:

-- Boost version: 1.53.0
-- Found the following Boost libraries:
--   program_options
--   filesystem
--   system
-- checking for module 'gsl'
--   package 'gsl' not found
-- Could NOT find GSL (missing:  GSL_LIBRARY GSL_INCLUDE_DIR) 
-- Intel(R) MKL could not be found.
-- 
-- The following OPTIONAL packages have been found:

 * PkgConfig
 * EXPAT
 * Doxygen

-- The following REQUIRED packages have been found:

 * Threads
 * Boost (required version >= 1.39.0)

-- The following OPTIONAL packages have not been found:

 * GSL
 * MKL

-- Could NOT find TXT2TAGS (missing:  TXT2TAGS_EXECUTABLE) 
txt2tags not found, help cmake to find it by setting TXT2TAGS_EXECUTABLE
-- Boost version: 1.53.0
-- Found the following Boost libraries:
--   program_options
--   filesystem
--   system
-- checking for module 'libvotca_tools'
--   package 'libvotca_tools' not found
CMake Error at /usr/share/cmake/Modules/FindPackageHandleStandardArgs.cmake:108 (message):
  Could NOT find SQLITE3 (missing: SQLITE3_INCLUDE_DIR)
Call Stack (most recent call first):
  /usr/share/cmake/Modules/FindPackageHandleStandardArgs.cmake:315 (_FPHSA_FAILURE_MESSAGE)
  csg/CMakeModules/FindSQLITE3.cmake:34 (find_package_handle_standard_args)
  csg/CMakeModules/FindVOTCA_TOOLS.cmake:32 (find_package)
  csg/CMakeLists.txt:83 (find_package)


-- Configuring incomplete, errors occurred!
See also "/home/sbinkas/votca/build/CMakeFiles/CMakeOutput.log".
See also "/home/sbinkas/votca/build/CMakeFiles/CMakeError.log".


Christoph Junghans

unread,
Apr 24, 2020, 9:10:33 PM4/24/20
to vo...@googlegroups.com
Votca-1.4.1 was released back in Sept. 2017 and we don't really
support such old versions anymore.
If you don't have a strong reason to use v1.4.1 try to use Votca-1.6.
see https://github.com/votca/votca/blob/master/README.md
Usually WITH_SQLITE3=OFF should disable the search for sqlite3,
something funky is going on here as L32 in v1.4.1 isn't actually
checking for sqlite3:
https://github.com/votca/csg/blob/v1.4.1/CMakeModules/FindVOTCA_CSG.cmake
Please check you are using the correct version of csg.

Christoph
>
>
> -- Configuring incomplete, errors occurred!
> See also "/home/sbinkas/votca/build/CMakeFiles/CMakeOutput.log".
> See also "/home/sbinkas/votca/build/CMakeFiles/CMakeError.log".
>
>
> --
> Join us on Slack: https://join.slack.com/t/votca/signup
> ---
> You received this message because you are subscribed to the Google Groups "votca" group.
> To unsubscribe from this group and stop receiving emails from it, send an email to votca+un...@googlegroups.com.
> To view this discussion on the web visit https://groups.google.com/d/msgid/votca/187cdc4c-6526-4716-8532-d60341aec7a1%40googlegroups.com.



--
Christoph Junghans
Web: http://www.compphys.de

Salman Bin Kashif

unread,
Apr 24, 2020, 11:00:58 PM4/24/20
to vo...@googlegroups.com
Thank you, Christoph. I now tried installing v 1.6. I am trying on local machine for now since it is easier to install all the dependencies.

Initially, my cmake version was old which was 3.5.1 and on installing votca it was throwing an error that minimum 3.10 version is required for cmake. Hence I manually installed cmake version 3.16.5

I verified cmake installation by checking the cmake --version
 
However, on installing the votca back I got the following error:

Could not find a configuration file for package "Eigen3" that is compatible
  with requested version "3.3.0".

  The following configuration files were considered but not accepted:

    /usr/lib/cmake/eigen3/Eigen3Config.cmake, version: unknown

Would I have to install specific cmake version?


Christoph Junghans

unread,
Apr 25, 2020, 12:00:43 AM4/25/20
to vo...@googlegroups.com


On Fri, Apr 24, 2020, 21:00 Salman Bin Kashif <sbkas...@gmail.com> wrote:
Thank you, Christoph. I now tried installing v 1.6. I am trying on local machine for now since it is easier to install all the dependencies.

Initially, my cmake version was old which was 3.5.1 and on installing votca it was throwing an error that minimum 3.10 version is required for cmake. Hence I manually installed cmake version 3.16.5

I verified cmake installation by checking the cmake --version
 
However, on installing the votca back I got the following error:

Could not find a configuration file for package "Eigen3" that is compatible
  with requested version "3.3.0".

  The following configuration files were considered but not accepted:

    /usr/lib/cmake/eigen3/Eigen3Config.cmake, version: unknown
This error means the Eigen3 you have installed is too old. Eigen-3.3.0 got released back in Nov. 2016, so that isn't very new either.

What version do you have installed?

Christoph

Salman Bin Kashif

unread,
Apr 25, 2020, 12:07:31 AM4/25/20
to vo...@googlegroups.com
Thanks again for answering my questions.

"What version do you have installed?" 

For eigen3, I have installed the version libeigen3-dev which is mentioned in the dependencies section. The version information for eigen3 is as follows:

libeigen3-dev is already the newest version (3.3~beta1-2)


-Salman

 

Christoph Junghans

unread,
Apr 25, 2020, 9:10:57 AM4/25/20
to vo...@googlegroups.com
On Fri, Apr 24, 2020 at 10:07 PM Salman Bin Kashif <sbkas...@gmail.com> wrote:
>
> Thanks again for answering my questions.
>
> "What version do you have installed?"
>
> For eigen3, I have installed the version libeigen3-dev which is mentioned in the dependencies section. The version information for eigen3 is as follows:
>
> libeigen3-dev is already the newest version (3.3~beta1-2)
Yes, and 3.3-beta1 is older than 3.3.0, so you need a newer version of eigen3.

What Linux distribution are you using?

Christoph
> To view this discussion on the web visit https://groups.google.com/d/msgid/votca/CAFwyQHwpcxE%3D%3DZK%3D2Udw9fSFpn3d2rkCk1COVSpY5MaDQ_TpCw%40mail.gmail.com.

Salman Bin Kashif

unread,
Apr 25, 2020, 9:15:41 AM4/25/20
to vo...@googlegroups.com
Okay. I will install a newer version manually and apt-get consider the  3.3-beta1 version to be the latest.

My linux version is:

Ubuntu 16.04.4 LTS 

Christoph Junghans

unread,
Apr 25, 2020, 9:41:41 AM4/25/20
to vo...@googlegroups.com
On Sat, Apr 25, 2020 at 7:15 AM Salman Bin Kashif <sbkas...@gmail.com> wrote:
>
> Okay. I will install a newer version manually and apt-get consider the 3.3-beta1 version to be the latest.
That might help:
https://askubuntu.com/questions/1068232/how-to-update-libeigen3-to-version-3-3-on-16-04-xenial

>
> My linux version is:
>
> Ubuntu 16.04.4 LTS
Hmm, as the name suggests that Ubuntu is from 2016 and we will only
support the latest LTS version, which currently is 18.04 (until later
this month when 20.04 LTS is released). We just don't have the
manpower to support anything but the latest versions of most Linux
distributions.

Plus, the gromacs version in 16.04 might be also too old to be used
with VOTCA-1.6.

Christoph
> To view this discussion on the web visit https://groups.google.com/d/msgid/votca/CAFwyQHzzXw%2ByX-ggrOJpXS7p1y0NoWpYrT658Bky%2B9_UW3Kneg%40mail.gmail.com.

Salman Bin Kashif

unread,
Apr 25, 2020, 10:53:37 AM4/25/20
to vo...@googlegroups.com
Thanks very much for being patient with me.

I had installed eigen3 easier which incidentally happen to be from the same source you pointed out. Eigen is being read properly I assume.

-- Found Eigen3: /home/sbinkas/software/eigen-eigen-67e894c6cd8f/build_dir

However, there are two more issues:

-- Could NOT find TXT2TAGS (missing: TXT2TAGS_EXECUTABLE)
txt2tags not found, help cmake to find it by setting TXT2TAGS_EXECUTABLE


I have provided the path to executable txt2tags file in /home/software/txt2tags-2.6/txt2tags but it still is not fould. However, the error was not fatal due to it.

The error was fatal due to not having been able to read c++ version in gromacs library. It is reaching to the libgromacs.so file though. 

The error message is as follows:

-- Looking for c++ gmx_version in /home/sbinkas/software/gromacs-2018.3/lib64/libgromacs.so
-- Looking for c++ gmx_version in /home/sbinkas/software/gromacs-2018.3/lib64/libgromacs.so - not found
CMake Error at csg/CMakeModules/FindGROMACS.cmake:57 (message):
  Could not find a suitable gromacs library.  gmx_version is not defined in
  the gromacs library, that is very very strange, take a look at the error
  message in /home/sbinkas/software/votca/build/CMakeFiles/CMakeError.log to
  find out what was going wrong.  This most likely means that your gromacs
  version is too old, we need at least gromacs 2016!

My version is 2018.3

I have used this command for cmake:

cmake -DBUILD_CSGAPPS=ON -DCMAKE_INSTALL_PREFIX=${prefix} -DWITH_GMX=ON ..

Providing the path to include  directory and library also makes no difference.

However, I would note a caveat. The gromacs version I installed manually on cluster was installed with -DGMX_MPI=OFF.

But, the dependencies requirement suggests gromacs-openmpi. Can that be a reason?

PS: I am talking about installation on cluster now. On local machine Votca package got installed using apt-get but as expected, the simulations were super slow.

Thanks for all your help again. 

Christoph Junghans

unread,
Apr 25, 2020, 12:12:57 PM4/25/20
to vo...@googlegroups.com
On Sat, Apr 25, 2020 at 8:53 AM Salman Bin Kashif <sbkas...@gmail.com> wrote:
>
> Thanks very much for being patient with me.
>
> I had installed eigen3 easier which incidentally happen to be from the same source you pointed out. Eigen is being read properly I assume.
>
> -- Found Eigen3: /home/sbinkas/software/eigen-eigen-67e894c6cd8f/build_dir
>
> However, there are two more issues:
>
> -- Could NOT find TXT2TAGS (missing: TXT2TAGS_EXECUTABLE)
> txt2tags not found, help cmake to find it by setting TXT2TAGS_EXECUTABLE
>
>
> I have provided the path to executable txt2tags file in /home/software/txt2tags-2.6/txt2tags but it still is not fould. However, the error was not fatal due to it.
Don't worry this one is optional.

>
> The error was fatal due to not having been able to read c++ version in gromacs library. It is reaching to the libgromacs.so file though.
>
> The error message is as follows:
>
> -- Looking for c++ gmx_version in /home/sbinkas/software/gromacs-2018.3/lib64/libgromacs.so
> -- Looking for c++ gmx_version in /home/sbinkas/software/gromacs-2018.3/lib64/libgromacs.so - not found
> CMake Error at csg/CMakeModules/FindGROMACS.cmake:57 (message):
> Could not find a suitable gromacs library. gmx_version is not defined in
> the gromacs library, that is very very strange, take a look at the error
> message in /home/sbinkas/software/votca/build/CMakeFiles/CMakeError.log to
> find out what was going wrong. This most likely means that your gromacs
> version is too old, we need at least gromacs 2016!
>
> My version is 2018.3
>
> I have used this command for cmake:
>
> cmake -DBUILD_CSGAPPS=ON -DCMAKE_INSTALL_PREFIX=${prefix} -DWITH_GMX=ON ..
>
> Providing the path to include directory and library also makes no difference.
>
> However, I would note a caveat. The gromacs version I installed manually on cluster was installed with -DGMX_MPI=OFF.
>
> But, the dependencies requirement suggests gromacs-openmpi. Can that be a reason?
Votca needs a non-mpi version, so that should be fine. You only need
gromacs with mpi at runtime to do parallel simulations.
Have a look at the end of
/home/sbinkas/software/votca/build/CMakeFiles/CMakeError.log to see
what is wrong with your libgromacs.


> PS: I am talking about installation on cluster now. On local machine Votca package got installed using apt-get but as expected, the simulations were super slow.
Usually you can speed up the gromacs simulation part by running on all cores.

Christoph
> To view this discussion on the web visit https://groups.google.com/d/msgid/votca/CAFwyQHy91W4YdUGKiYrispvdjrfSQhSFHKrWMubs4X8uDGexTQ%40mail.gmail.com.

Salman Bin Kashif

unread,
Apr 25, 2020, 12:26:54 PM4/25/20
to vo...@googlegroups.com
This is the error associated with libgromacs:

gmake[1]: *** No rule to make target `/home/sbinkas/gromacs/lib/libgromacs.so', needed by `cmTC_b3327    '.  Stop.
101 gmake[1]: Leaving directory `/home/sbinkas/software/votca/build/CMakeFiles/CMakeTmp'
102 gmake: *** [cmTC_b3327/fast] Error 2

Christoph Junghans

unread,
Apr 25, 2020, 12:31:48 PM4/25/20
to vo...@googlegroups.com
On Sat, Apr 25, 2020 at 10:26 AM Salman Bin Kashif <sbkas...@gmail.com> wrote:
>
> This is the error associated with libgromacs:
>
> gmake[1]: *** No rule to make target `/home/sbinkas/gromacs/lib/libgromacs.so', needed by `cmTC_b3327 '. Stop.
> 101 gmake[1]: Leaving directory `/home/sbinkas/software/votca/build/CMakeFiles/CMakeTmp'
> 102 gmake: *** [cmTC_b3327/fast] Error 2
In the CMake output it says, it is using
/home/sbinkas/software/gromacs-2018.3/lib64/libgromacs.so, but now it
is looking for
`/home/sbinkas/gromacs/lib/libgromacs.so, you might want to help CMake
by setting -DGROMACS_LIBRARY=/home/sbinkas/software/gromacs-2018.3/lib64/libgromacs.so.
> To view this discussion on the web visit https://groups.google.com/d/msgid/votca/CAFwyQHzq6TrBChPWMEdpKKid8CoLLbb555weSia_SHzx4tX%3D5A%40mail.gmail.com.

Salman Bin Kashif

unread,
Apr 25, 2020, 12:52:15 PM4/25/20
to vo...@googlegroups.com
Thank you. This suggestion works. The cmake command did not throw any fatal error this time.

However, I would like to discuss some caveats:

1. Would TXT2TAGS not being found be an issue in later steps? Like when running the simulation? It does suggest optional. But I assume it may be used for generating xml files during run.

2. Warning with run time library: 

  Cannot generate a safe runtime search path for target votca_tools because
  files in some directories may conflict with libraries in implicit
  directories:

    runtime library [libexpat.so.1] in /usr/lib64 may be hidden by files in:
      /home/sbinkas/software/lib 

I kind of get where is it coming from. This /home/sbinkas/software/lib got added while installing one of the dependencies. I don't remember which one to be exact, but possibly it was from eigen3. So a conflict may have arisen. Should I worry about it?

Again, I very much appreciate your help.

-Salman


Output:
 

- The following OPTIONAL packages have been found:

 * Git
 * FFTW3
 * UnixCommands
 * PkgConfig
 * HDF5

 * Doxygen

-- The following REQUIRED packages have been found:

 * EXPAT
 * Threads
 * Boost (required version >= 1.53.0)
 * Eigen3 (required version >= 3.3.0)
 * GROMACS (required version >= 2016)


-- The following OPTIONAL packages have not been found:

 * MKL
 * TXT2TAGS
 * CLANG_FORMAT (required version >= 7.0.1)

-- Configuring done
CMake Warning at tools/src/libtools/CMakeLists.txt:5 (add_library):
  Cannot generate a safe runtime search path for target votca_tools because
  files in some directories may conflict with libraries in implicit
  directories:

    runtime library [libexpat.so.1] in /usr/lib64 may be hidden by files in:
      /home/sbinkas/software/lib

  Some of these libraries may not be found correctly.


-- Generating done
-- Build files have been written to: /home/sbinkas/software/votca/build

Christoph Junghans

unread,
Apr 25, 2020, 1:03:45 PM4/25/20
to vo...@googlegroups.com
On Sat, Apr 25, 2020 at 10:52 AM Salman Bin Kashif <sbkas...@gmail.com> wrote:
>
> Thank you. This suggestion works. The cmake command did not throw any fatal error this time.
>
> However, I would like to discuss some caveats:
>
> 1. Would TXT2TAGS not being found be an issue in later steps? Like when running the simulation? It does suggest optional. But I assume it may be used for generating xml files during run.
txt2tags is only used to build the manpage (documentation), so no issue here.
>
> 2. Warning with run time library:
>
> Cannot generate a safe runtime search path for target votca_tools because
> files in some directories may conflict with libraries in implicit
> directories:
>
> runtime library [libexpat.so.1] in /usr/lib64 may be hidden by files in:
> /home/sbinkas/software/lib
This error basically means there are two libexpat in your LD_LIBRARY_PATH.
You tell CMake to explicitly use the one from
/home/sbinkas/software/lib by add
-DCMAKE_PREFIX_PATH=/home/sbinkas/software.

Christoph
> To view this discussion on the web visit https://groups.google.com/d/msgid/votca/CAFwyQHy4vWM9Ceaw5scKKeY-fJd0kMBPYziKsLw6qarkNS%3DR9w%40mail.gmail.com.

Salman Bin Kashif

unread,
Apr 25, 2020, 1:25:14 PM4/25/20
to vo...@googlegroups.com
Thank you. That works too. Next error was encountered in make -j5 command.

/home/sbinkas/software/include/eigen3/unsupported/Eigen/FFT:75:21: fatal error: fftw3.h: No such file or directory

I didn't find fftw.h in the entire eigen folder though. I am not sure which path should I provide then.

Christoph Junghans

unread,
Apr 25, 2020, 1:41:56 PM4/25/20
to vo...@googlegroups.com
On Sat, Apr 25, 2020 at 11:25 Salman Bin Kashif <sbkas...@gmail.com> wrote:
Thank you. That works too. Next error was encountered in make -j5 command.

/home/sbinkas/software/include/eigen3/unsupported/Eigen/FFT:75:21: fatal error: fftw3.h: No such file or directory
it seems the fftw3 dev package is missing!

Can you post the output of your CMake execution, what fftw3 found?

Christoph 

Salman Bin Kashif

unread,
Apr 25, 2020, 1:45:20 PM4/25/20
to vo...@googlegroups.com
I have attached the printed output from cmake command herewith

notes.txt

Christoph Junghans

unread,
Apr 25, 2020, 1:59:24 PM4/25/20
to vo...@googlegroups.com
On Sat, Apr 25, 2020 at 11:45 Salman Bin Kashif <sbkas...@gmail.com> wrote:
I have attached the printed output from cmake command herewith
thanks, please do me a favor and also attach the output of “make VERBOSE=1”.

Salman Bin Kashif

unread,
Apr 25, 2020, 2:09:54 PM4/25/20
to vo...@googlegroups.com
Following is the output of make VERBOSE-1

[sbinkas@node0170 build]$ make VERBOSE=1
/software/cmake/3.10.0/bin/cmake -H/home/sbinkas/software/votca -B/home/sbinkas/software/votca/build --check-build-system CMakeFiles/Makefile.cmake 0
/software/cmake/3.10.0/bin/cmake -E cmake_progress_start /home/sbinkas/software/votca/build/CMakeFiles /home/sbinkas/software/votca/build/CMakeFiles/progress.marks
make -f CMakeFiles/Makefile2 all
make[1]: Entering directory `/home/sbinkas/software/votca/build'
make -f tools/share/man/CMakeFiles/votca_tools_manpages.dir/build.make tools/share/man/CMakeFiles/votca_tools_manpages.dir/depend
make[2]: Entering directory `/home/sbinkas/software/votca/build'
cd /home/sbinkas/software/votca/build && /software/cmake/3.10.0/bin/cmake -E cmake_depends "Unix Makefiles" /home/sbinkas/software/votca /home/sbinkas/software/votca/tools/share/man /home/sbinkas/software/votca/build /home/sbinkas/software/votca/build/tools/share/man /home/sbinkas/software/votca/build/tools/share/man/CMakeFiles/votca_tools_manpages.dir/DependInfo.cmake --color=
make[2]: Leaving directory `/home/sbinkas/software/votca/build'
make -f tools/share/man/CMakeFiles/votca_tools_manpages.dir/build.make tools/share/man/CMakeFiles/votca_tools_manpages.dir/build
make[2]: Entering directory `/home/sbinkas/software/votca/build'
make[2]: Nothing to be done for `tools/share/man/CMakeFiles/votca_tools_manpages.dir/build'.
make[2]: Leaving directory `/home/sbinkas/software/votca/build'
[  0%] Built target votca_tools_manpages
make -f tools/src/libtools/CMakeFiles/gitversion.dir/build.make tools/src/libtools/CMakeFiles/gitversion.dir/depend
make[2]: Entering directory `/home/sbinkas/software/votca/build'
cd /home/sbinkas/software/votca/build && /software/cmake/3.10.0/bin/cmake -E cmake_depends "Unix Makefiles" /home/sbinkas/software/votca /home/sbinkas/software/votca/tools/src/libtools /home/sbinkas/software/votca/build /home/sbinkas/software/votca/build/tools/src/libtools /home/sbinkas/software/votca/build/tools/src/libtools/CMakeFiles/gitversion.dir/DependInfo.cmake --color=
make[2]: Leaving directory `/home/sbinkas/software/votca/build'
make -f tools/src/libtools/CMakeFiles/gitversion.dir/build.make tools/src/libtools/CMakeFiles/gitversion.dir/build
make[2]: Entering directory `/home/sbinkas/software/votca/build'
cd /home/sbinkas/software/votca/build/tools/src/libtools && /software/cmake/3.10.0/bin/cmake -DTOP_SOURCE_DIR="/home/sbinkas/software/votca" -DGIT_EXECUTABLE="/bin/git" -P /home/sbinkas/software/votca/tools/CMakeModules/gitversion.cmake
Current git revision is 98e27cd
make[2]: Leaving directory `/home/sbinkas/software/votca/build'
[  0%] Built target gitversion
make -f tools/src/libtools/CMakeFiles/votca_tools.dir/build.make tools/src/libtools/CMakeFiles/votca_tools.dir/depend
make[2]: Entering directory `/home/sbinkas/software/votca/build'
cd /home/sbinkas/software/votca/build && /software/cmake/3.10.0/bin/cmake -E cmake_depends "Unix Makefiles" /home/sbinkas/software/votca /home/sbinkas/software/votca/tools/src/libtools /home/sbinkas/software/votca/build /home/sbinkas/software/votca/build/tools/src/libtools /home/sbinkas/software/votca/build/tools/src/libtools/CMakeFiles/votca_tools.dir/DependInfo.cmake --color=
make[2]: Leaving directory `/home/sbinkas/software/votca/build'
make -f tools/src/libtools/CMakeFiles/votca_tools.dir/build.make tools/src/libtools/CMakeFiles/votca_tools.dir/build
make[2]: Entering directory `/home/sbinkas/software/votca/build'
make[2]: Nothing to be done for `tools/src/libtools/CMakeFiles/votca_tools.dir/build'.
make[2]: Leaving directory `/home/sbinkas/software/votca/build'
[ 27%] Built target votca_tools
make -f tools/src/tools/CMakeFiles/votca_property.dir/build.make tools/src/tools/CMakeFiles/votca_property.dir/depend
make[2]: Entering directory `/home/sbinkas/software/votca/build'
cd /home/sbinkas/software/votca/build && /software/cmake/3.10.0/bin/cmake -E cmake_depends "Unix Makefiles" /home/sbinkas/software/votca /home/sbinkas/software/votca/tools/src/tools /home/sbinkas/software/votca/build /home/sbinkas/software/votca/build/tools/src/tools /home/sbinkas/software/votca/build/tools/src/tools/CMakeFiles/votca_property.dir/DependInfo.cmake --color=
make[2]: Leaving directory `/home/sbinkas/software/votca/build'
make -f tools/src/tools/CMakeFiles/votca_property.dir/build.make tools/src/tools/CMakeFiles/votca_property.dir/build
make[2]: Entering directory `/home/sbinkas/software/votca/build'
[ 28%] Building CXX object tools/src/tools/CMakeFiles/votca_property.dir/votca_property.cc.o
cd /home/sbinkas/software/votca/build/tools/src/tools && /software/gcc/5.4.0/bin/g++   -I/home/sbinkas/software/votca/tools/include -I/home/sbinkas/software/votca/build/tools/include -isystem /home/sbinkas/software/include/eigen3  -Wall -Wextra -Wpedantic -Wshadow -Wconversion -O3 -DNDEBUG   -std=c++14 -o CMakeFiles/votca_property.dir/votca_property.cc.o -c /home/sbinkas/software/votca/tools/src/tools/votca_property.cc
In file included from /home/sbinkas/software/votca/tools/include/votca/tools/eigen.h:52:0,
                 from /home/sbinkas/software/votca/tools/include/votca/tools/property.h:21,
                 from /home/sbinkas/software/votca/tools/include/votca/tools/application.h:21,
                 from /home/sbinkas/software/votca/tools/src/tools/votca_property.cc:22:

/home/sbinkas/software/include/eigen3/unsupported/Eigen/FFT:75:21: fatal error: fftw3.h: No such file or directory
compilation terminated.
make[2]: *** [tools/src/tools/CMakeFiles/votca_property.dir/votca_property.cc.o] Error 1
make[2]: Leaving directory `/home/sbinkas/software/votca/build'
make[1]: *** [tools/src/tools/CMakeFiles/votca_property.dir/all] Error 2
make[1]: Leaving directory `/home/sbinkas/software/votca/build'
make: *** [all] Error 2

Christoph Junghans

unread,
Apr 25, 2020, 3:44:04 PM4/25/20
to vo...@googlegroups.com
Ok, I found it, that is actually a bug in VOTCA, I made a fix here:
https://github.com/votca/tools/pull/234
alternatively, you can disable fftw, with -DCMAKE_DISABLE_FIND_PACKAGE_FFTW3=ON

Christoph
> To view this discussion on the web visit https://groups.google.com/d/msgid/votca/CAFwyQHyaSDj_y-zCi2h-13P2TtheEAs9TW3J5-XgGNf5awj%3Dxg%40mail.gmail.com.

Salman Bin Kashif

unread,
Apr 25, 2020, 3:46:34 PM4/25/20
to vo...@googlegroups.com
Thanks so much. I wiil try using the disable package option first.

Salman Bin Kashif

unread,
Apr 25, 2020, 4:03:50 PM4/25/20
to vo...@googlegroups.com
Thank you. I tried the seI guess there may be another issue. I used the following command:

cmake -DBUILD_CSGAPPS=ON -DCMAKE_INSTALL_PREFIX=${prefix} -DWITH_GMX=ON -DGROMACS_INCLUDE_DIR=/home/sbinkas/software/gromacs-2018.3/include -DGROMACS_LIBRARY=/home/sbinkas/software/gromacs-2018.3/lib64/libgromacs.so -DCMAKE_DISABLE_FIND_PACKAGE_FFTW3=ON -DCMAKE_PREFIX_PATH=/home/sbinkas/software ..

cmake command runs fine.

make -j5 throws the error

Error looks like:

../libtools/libvotca_tools.so.6: undefined reference to `boost::program_options::option_description::format_name[abi:cxx11][ 66%] Built target votca_csg

make: *** [all] Error 2

Am I missing something?

In the meanwhile, I will try with option 1.

Salman Bin Kashif

unread,
Apr 25, 2020, 4:30:24 PM4/25/20
to vo...@googlegroups.com
I am still getting errors in the make -j5 command. I am not sure which branch should I pull the changes from. The modules changed in the above mentioned pull doesn't seem to have changed in my local copy. I am pulling from stable branch.

Initially, I have cloned for version v1.6

Christoph Junghans

unread,
Apr 25, 2020, 4:54:48 PM4/25/20
to vo...@googlegroups.com
On Sat, Apr 25, 2020 at 2:30 PM Salman Bin Kashif <sbkas...@gmail.com> wrote:
>
> I am still getting errors in the make -j5 command. I am not sure which branch should I pull the changes from. The modules changed in the above mentioned pull doesn't seem to have changed in my local copy. I am pulling from stable branch.
If haven't merge the patch yet, try to go into the csg folder and do:
$ git remote update
$ git checkout for/stable/fftw_fix
to change the code.

Christoph
> To view this discussion on the web visit https://groups.google.com/d/msgid/votca/CAFwyQHzhzLokv_xK6rMEcbpTjOgXX1pTrSGq-bQAs8QQu_uKog%40mail.gmail.com.

Christoph Junghans

unread,
Apr 25, 2020, 4:58:13 PM4/25/20
to vo...@googlegroups.com
On Sat, Apr 25, 2020 at 2:03 PM Salman Bin Kashif <sbkas...@gmail.com> wrote:
>
> Thank you. I tried the seI guess there may be another issue. I used the following command:
>
> cmake -DBUILD_CSGAPPS=ON -DCMAKE_INSTALL_PREFIX=${prefix} -DWITH_GMX=ON -DGROMACS_INCLUDE_DIR=/home/sbinkas/software/gromacs-2018.3/include -DGROMACS_LIBRARY=/home/sbinkas/software/gromacs-2018.3/lib64/libgromacs.so -DCMAKE_DISABLE_FIND_PACKAGE_FFTW3=ON -DCMAKE_PREFIX_PATH=/home/sbinkas/software ..
>
> cmake command runs fine.
>
> make -j5 throws the error
>
> Error looks like:
>
> ../libtools/libvotca_tools.so.6: undefined reference to `boost::program_options::option_description::format_name[abi:cxx11][ 66%] Built target votca_csg
This error means your boost was compiled with a different compiler and
possibly has no C++-11 support enabled.

Christoph
> To view this discussion on the web visit https://groups.google.com/d/msgid/votca/CAFwyQHwt0yXzKSOfqzkxmHsda6x2H5bkUrNqQm-5PyGUNX2Q6w%40mail.gmail.com.

Salman Bin Kashif

unread,
Apr 25, 2020, 5:36:05 PM4/25/20
to vo...@googlegroups.com
Thanks. The version I had installed manually or one available in my cluster doesn't seem to work. Would you suggest any specific boost installation?

Christoph Junghans

unread,
Apr 25, 2020, 6:31:19 PM4/25/20
to vo...@googlegroups.com
On Sat, Apr 25, 2020 at 15:36 Salman Bin Kashif <sbkas...@gmail.com> wrote:
Thanks. The version I had installed manually or one available in my cluster doesn't seem to work. Would you suggest any specific boost installation?
Try to install votca and its deps with Spack:
Christoph 

Salman Bin Kashif

unread,
Apr 25, 2020, 6:49:13 PM4/25/20
to vo...@googlegroups.com
Hey Christoph,

Installing a different version of boost as well as hd5 seemed to have worked in eliminating erorr (hd5 associated error showed up later). The software got installed till final step (make install) without throwing any fatal error precisely at the same point when your message was received. Would you recommend spack option right away or if the current installation throw errors during running the tutorials?

I would like to note one thing:

In navigating to csg, git checkout for/stable/fftw_fix was not being shown as available options. Available options were master, stable and various versions. I then noticed that you made changes in the tools folder from the pull link you sent. I hence tried git checkout fpr/stable/fftw_fix and it took me to a different branch, and I also manually checked for the changes in the modules you mentioned. Hence, it looks like the changes have to be made in the tools folder. Still, I want to confirm this change with you to make sure that it doesn't come to bite me later.

I have another question:

Which file would we need to source for loading VOTCA, like GMXRC of Gromacs. Or would we need to do something else to make use of VOTCA functions.

Thanks again for helping me out.





Christoph Junghans

unread,
Apr 25, 2020, 6:57:21 PM4/25/20
to vo...@googlegroups.com
On Sat, Apr 25, 2020 at 16:49 Salman Bin Kashif <sbkas...@gmail.com> wrote:
Hey Christoph,

Installing a different version of boost as well as hd5 seemed to have worked in eliminating erorr (hd5 associated error showed up later). The software got installed till final step (make install) without throwing any fatal error precisely at the same point when your message was received. Would you recommend spack option right away or if the current installation throw errors during running the tutorials?

I would like to note one thing:

In navigating to csg, git checkout for/stable/fftw_fix was not being shown as available options. Available options were master, stable and various versions. I then noticed that you made changes in the tools folder from the pull link you sent. I hence tried git checkout fpr/stable/fftw_fix and it took me to a different branch, and I also manually checked for the changes in the modules you mentioned. Hence, it looks like the changes have to be made in the tools folder. Still, I want to confirm this change with you to make sure that it doesn't come to bite me later.
Not sure what happened here, but don’t worrying it will be part of stable soon! And if the build worked you should be good to go, FFTW is only used for calculate correlation functions, if you build without it you get a slightly slower version from Eigen3.


I have another question:

Which file would we need to source for loading VOTCA, like GMXRC of Gromacs. Or would we need to do something else to make use of VOTCA functions.
You need to source GMXRC and VOTCARC.

Christoph 

Salman Bin Kashif

unread,
Apr 26, 2020, 4:20:34 PM4/26/20
to vo...@googlegroups.com
I am getting an error on running csg map

csg_map: error while loading shared libraries: libboost_program_options.so.1.55.0: cannot open shared object file: No such file or directory

I have installed a boost version manually in the folder /home/sbinkas/

I have sourced VOTCARC.bash

Christoph Junghans

unread,
Apr 26, 2020, 4:33:56 PM4/26/20
to vo...@googlegroups.com
On Sun, Apr 26, 2020 at 2:20 PM Salman Bin Kashif <sbkas...@gmail.com> wrote:
>
> I am getting an error on running csg map
>
> csg_map: error while loading shared libraries: libboost_program_options.so.1.55.0: cannot open shared object file: No such file or directory
>
> I have installed a boost version manually in the folder /home/sbinkas/
Try to add the path where libboost_program_options.so.1.55.0.so lives
to your LD_LIBRARY_PATH, e.g.
export LD_LIBRARY_PATH="/home/sbinkas/boost/lib64:${LD_LIBRARY_PATH}"
(you might want to add that line to your ~/.bashrc, so you don't have
to do it each time you login.

Alternatively, you can build VOTCA with -DENABLE_RPATH_INJECT=ON.

Christoph
> To view this discussion on the web visit https://groups.google.com/d/msgid/votca/CAFwyQHziUznLvB%3DpBhAM_SnEn2rTCXdtLuP5Ag4aL5AS6Cfq%3Dw%40mail.gmail.com.

Salman Bin Kashif

unread,
Apr 26, 2020, 5:07:29 PM4/26/20
to vo...@googlegroups.com
Ok. I will try reinstalling with second option as providing path doesn't seem to work.

Secondly, I has installed votca using spack. But I am not sure where to source it from. Can you let me know the probable file location. I can try that first

Salman Bin Kashif

unread,
Apr 26, 2020, 5:17:30 PM4/26/20
to vo...@googlegroups.com
From the spack installation following is the load error

[sbinkas@node1078 ~]$ spack load votca-csg@1.6
==> Error: votca-csg@1.6 matches multiple packages.
  Matching packages:
    ybythll votca-csg@1.6%g...@5.4.0 arch=linux-ol7-haswell
    ols2s7o votca-csg@1.6%g...@5.4.0 arch=linux-ol7-haswell
  Use a more specific spec.

Christoph Junghans

unread,
Apr 26, 2020, 5:23:58 PM4/26/20
to vo...@googlegroups.com
On Sun, Apr 26, 2020 at 15:07 Salman Bin Kashif <sbkas...@gmail.com> wrote:
Ok. I will try reinstalling with second option as providing path doesn't seem to work.
make sure you use the actual location in LD_LIBRARY_PATH if you are using tcsh instead of bash you well need to use setenv instead of the export command. You can also add the path to your VOTRARC.


Secondly, I has installed votca using spack. But I am not sure where to source it from. Can you let me know the probable file location. I can try that first
Just run “ spack find -p votca-csg” and it will show the path.

Christoph 

Salman Bin Kashif

unread,
Apr 26, 2020, 5:51:07 PM4/26/20
to vo...@googlegroups.com
Thanks. spack version seems to work fine.

However, I have one more question from the spce ibi tutorial. I have run the rus.sh and Extarct_Energies.sh from atomistic folder successfully which means that it has a reference trajectory.vi 

On running the run.sh file the simulation throws an error that Group cur-off scheme cannot be used with OpenMP threads.

If I set scheme to Verlet, it is asking to change from Coulomb-type User and  vdw-type user to cut-off or PME which might not be right as parameters should be generated from tabulated values.

I am not sure what is wrong in implementation that it is asling for openmp

I have sourced

spack load votca-tools@1.6
spack load votca-xtp@1.6



Christoph Junghans

unread,
Apr 26, 2020, 7:00:16 PM4/26/20
to vo...@googlegroups.com
On Sun, Apr 26, 2020 at 3:51 PM Salman Bin Kashif <sbkas...@gmail.com> wrote:
>
> Thanks. spack version seems to work fine.
>
> However, I have one more question from the spce ibi tutorial. I have run the rus.sh and Extarct_Energies.sh from atomistic folder successfully which means that it has a reference trajectory.vi
>
> On running the run.sh file the simulation throws an error that Group cur-off scheme cannot be used with OpenMP threads.
>
> If I set scheme to Verlet, it is asking to change from Coulomb-type User and vdw-type user to cut-off or PME which might not be right as parameters should be generated from tabulated values.

Ah, I think I know what this is.

Can you see what gromacs version spack installed?
(spack find -p gromacs)

If is it 2020 or newer, you will have rebuild votca-csg with gromacs-2019.
$ spack uninstall votca-csg
$ spack install votca-csg@1.6 ^gromacs@2019.6

Background is that gromacs-2020 has no support for tabulated
interactions (see https://gitlab.com/gromacs/gromacs/-/issues/1347),
which you need for coarse-graining. Our cmake prints out a warning,
but spack makes it easy to overlook it. Hopefully tabulated
interactions will be back in gromacs-2021, but it won't hurt to
comment on the gromacs isse.

Christoph

Christoph
> To view this discussion on the web visit https://groups.google.com/d/msgid/votca/CAFwyQHzd_%3DedFkPH7i2mOaaLT_bpeGqrA2PRMDmMstxvNMBA%2BA%40mail.gmail.com.

Christoph Junghans

unread,
Apr 26, 2020, 7:14:21 PM4/26/20
to vo...@googlegroups.com
On Sun, Apr 26, 2020 at 17:00 Christoph Junghans <jung...@votca.org> wrote:
On Sun, Apr 26, 2020 at 3:51 PM Salman Bin Kashif <sbkas...@gmail.com> wrote:
>
> Thanks. spack version seems to work fine.
>
> However, I have one more question from the spce ibi tutorial. I have run the rus.sh and Extarct_Energies.sh from atomistic folder successfully which means that it has a reference trajectory.vi
>
> On running the run.sh file the simulation throws an error that Group cur-off scheme cannot be used with OpenMP threads.
>
> If I set scheme to Verlet, it is asking to change from Coulomb-type User and  vdw-type user to cut-off or PME which might not be right as parameters should be generated from tabulated values.

Ah, I think I know what this is.

Can you see what gromacs version spack installed?
(spack find -p gromacs)

If is it 2020 or newer, you will have rebuild votca-csg with gromacs-2019.
$ spack uninstall votca-csg
$ spack install votca-csg@1.6 ^gromacs@2019.6
Should be automatic in spack soon.

Salman Bin Kashif

unread,
Apr 27, 2020, 7:26:33 AM4/27/20
to vo...@googlegroups.com
Thanks again. In the csg-tutorials force matching, I am not sure what tool or file to use to calculate the final RDF. I see the CG-CG potential file but not sure where to calculate the RDF from.

In IBI, a CG-CG dist file is created at each step and that can be used to compare RDF with atomistic, right?

Christoph Junghans

unread,
Apr 27, 2020, 7:41:00 AM4/27/20
to vo...@googlegroups.com
On Mon, Apr 27, 2020 at 05:26 Salman Bin Kashif <sbkas...@gmail.com> wrote:
Thanks again. In the csg-tutorials force matching, I am not sure what tool or file to use to calculate the final RDF. I see the CG-CG potential file but not sure where to calculate the RDF from.
You will have to run a course-grained simulation with potential and calculate the red from that.


In IBI, a CG-CG dist file is created at each step and that can be used to compare RDF with atomistic, right?
correct!

You can use the files to run a simulation with the force-matched potential.

Christoph 

Salman Bin Kashif

unread,
Apr 27, 2020, 7:50:06 AM4/27/20
to vo...@googlegroups.com
Thanks. Sorry for my ignorance as I am new to this package. I am not sure if there is already a script to run simulation with CG potential. 

If there is not, I am assuming the following can work:

Use the mdp file from IBI, and supply the table generated in FM as user defined potential using appropriate GROMACS syntax?



Salman Bin Kashif

unread,
Apr 27, 2020, 7:53:03 AM4/27/20
to vo...@googlegroups.com
Maybe this is the command for running CSG potential:

csg_call --options fmatch.xml --ia-name CG-CG --ia-type non-bonded convert_potential gromacs --clean input.pot table_CG_CG.xvg # calls convert_potential gromacs. Unsampled regions for distances smaller than the min value are extrapolated.

Christoph Junghans

unread,
Apr 27, 2020, 8:11:57 AM4/27/20
to vo...@googlegroups.com


On Mon, Apr 27, 2020, 05:50 Salman Bin Kashif <sbkas...@gmail.com> wrote:
Thanks. Sorry for my ignorance as I am new to this package. I am not sure if there is already a script to run simulation with CG potential. 

If there is not, I am assuming the following can work:

Use the mdp file from IBI, and supply the table generated in FM as user defined potential using appropriate GROMACS syntax?
Yes, that is correct! Also use the gro file from IBI. You might want to increase the number of a MD steps in the mdp a bit, as you can effort a long simulation than a single iteration.

Christoph



Christoph Junghans

unread,
Apr 27, 2020, 8:15:47 AM4/27/20
to vo...@googlegroups.com


On Mon, Apr 27, 2020, 05:53 Salman Bin Kashif <sbkas...@gmail.com> wrote:
Maybe this is the command for running CSG potential:

csg_call --options fmatch.xml --ia-name CG-CG --ia-type non-bonded convert_potential gromacs --clean input.pot table_CG_CG.xvg # calls convert_potential gromacs. Unsampled regions for distances smaller than the min value are extrapolated.
The core function of this command is "convert_potential gromacs", which will convert "input.pot" (a VOTCA style potential) to table_CG_CG.xvg (a gromacs style table).
All the other things are just options.
(see "csg_call --help")

Christoph

Salman Bin Kashif

unread,
Apr 27, 2020, 8:54:20 AM4/27/20
to vo...@googlegroups.com
Alright. That makes sense.

I encountered another an issue in running re tutorial for spce using the default files. The error message suggests: 

Updating parameters
an error occurred:
Hessian NOT a positive definite!
This can be a result of poor initial guess or ill-suited CG potential settings or poor CG sampling.

The potential file to modify will be cg.param.init, right? Or is  there something else I may be missing?




Christoph Junghans

unread,
Apr 27, 2020, 9:12:06 AM4/27/20
to vo...@googlegroups.com
On Mon, Apr 27, 2020 at 6:54 AM Salman Bin Kashif <sbkas...@gmail.com> wrote:
>
> Alright. That makes sense.
>
> I encountered another an issue in running re tutorial for spce using the default files. The error message suggests:
>
> Updating parameters
> an error occurred:
> Hessian NOT a positive definite!
> This can be a result of poor initial guess or ill-suited CG potential settings or poor CG sampling.
>
> The potential file to modify will be cg.param.init, right? Or is there something else I may be missing?
The error message says it all, it could be cg.param.init or you might
need to run longer iterations.
> To view this discussion on the web visit https://groups.google.com/d/msgid/votca/CAFwyQHzYi_tVpGiuZFaZO%3DV0M9gKF6HnCDZ74saZRESvxMV73A%40mail.gmail.com.

Christoph Junghans

unread,
Apr 27, 2020, 10:08:37 AM4/27/20
to vo...@googlegroups.com
On Mon, Apr 27, 2020 at 7:11 AM Christoph Junghans <jung...@votca.org> wrote:
>
> On Mon, Apr 27, 2020 at 6:54 AM Salman Bin Kashif <sbkas...@gmail.com> wrote:
> >
> > Alright. That makes sense.
> >
> > I encountered another an issue in running re tutorial for spce using the default files. The error message suggests:
> >
> > Updating parameters
> > an error occurred:
> > Hessian NOT a positive definite!
> > This can be a result of poor initial guess or ill-suited CG potential settings or poor CG sampling.
> >
> > The potential file to modify will be cg.param.init, right? Or is there something else I may be missing?
> The error message says it all, it could be cg.param.init or you might
> need to run longer iterations.
Ok, I see what you mean now in the spce/re tutorial, let me have a look.

Christoph

Christoph Junghans

unread,
Apr 27, 2020, 2:36:48 PM4/27/20
to vo...@googlegroups.com
On Mon, Apr 27, 2020 at 8:08 AM Christoph Junghans <jung...@votca.org> wrote:
>
> On Mon, Apr 27, 2020 at 7:11 AM Christoph Junghans <jung...@votca.org> wrote:
> >
> > On Mon, Apr 27, 2020 at 6:54 AM Salman Bin Kashif <sbkas...@gmail.com> wrote:
> > >
> > > Alright. That makes sense.
> > >
> > > I encountered another an issue in running re tutorial for spce using the default files. The error message suggests:
> > >
> > > Updating parameters
> > > an error occurred:
> > > Hessian NOT a positive definite!
> > > This can be a result of poor initial guess or ill-suited CG potential settings or poor CG sampling.
> > >
> > > The potential file to modify will be cg.param.init, right? Or is there something else I may be missing?
> > The error message says it all, it could be cg.param.init or you might
> > need to run longer iterations.
> Ok, I see what you mean now in the spce/re tutorial, let me have a look.
https://github.com/votca/csg-tutorials/issues/81

Salman Bin Kashif

unread,
Apr 27, 2020, 2:44:55 PM4/27/20
to vo...@googlegroups.com
Thank you. Since I am using Spack installation, should I just update the votca-csg package?

Christoph Junghans

unread,
Apr 27, 2020, 2:47:04 PM4/27/20
to vo...@googlegroups.com
On Mon, Apr 27, 2020 at 12:44 PM Salman Bin Kashif <sbkas...@gmail.com> wrote:
>
> Thank you. Since I am using Spack installation, should I just update the votca-csg package?
Well, it is just an issue description, we are still working on the
fix. If you need the relative entropy right now, you will have to use
VOTCA 1.5.1
> To view this discussion on the web visit https://groups.google.com/d/msgid/votca/CAFwyQHwwbXp-KScHAESR-tf7k3dyOAxw6qv%2BUj_mzUtDmceEDw%40mail.gmail.com.

Christoph Junghans

unread,
May 27, 2020, 11:53:22 AM5/27/20
to vo...@googlegroups.com
On Mon, Apr 27, 2020 at 12:46 PM Christoph Junghans <jung...@votca.org> wrote:
>
> On Mon, Apr 27, 2020 at 12:44 PM Salman Bin Kashif <sbkas...@gmail.com> wrote:
> >
> > Thank you. Since I am using Spack installation, should I just update the votca-csg package?
> Well, it is just an issue description, we are still working on the
> fix. If you need the relative entropy right now, you will have to use
> VOTCA 1.5.1
The issue has been fixed and will be included in votca-1.6.1 until
that is released, you can use votca-csg@stable in spack.
Reply all
Reply to author
Forward
0 new messages