Gromacs installation patched with plumed

User discussions
1

GROMACS version:2020.2
GROMACS modification: Yes (plumed)

Dear all,

I’m trying to install GROMACS 2020.2 patched by plumed, to run targeted MD with MPI.
These commands were executed successfully;

tar zxvf gromacs-2020.2.tar.gz
cd gromacs-2020.2
plumed patch -p (and I selected 4)

picture_pc_7a3ffbd74c7c945623fec6b394f92f6e
picture_pc_7a3ffbd74c7c945623fec6b394f92f6e800×130 20.2 KB

mkdir build
cd build

After that I tried this command;

…/…/cmake/cmake-3.9.6-Linux-x86_64/bin/cmake … -DCMAKE_C_COMPILER=/misc/home/ikuma/openmpi-4.0.5/bin/mpicc -DCMAKE_CXX_COMPILER=/misc/home/ikuma/openmpi-4.0.5/bin/mpicxx -DGMX_MPI=on -DCMAKE_INSTALL_PREFIX=/misc/home/ikuma/gromacs-2020.2/ -DGMX_FFT_LIBRARY=fftw3 -DFFTWF_LIBRARIES=/misc/home/ikuma/fftw-3.3.8/lib/libfftw3f_mpi.so -DFFTWF_INCLUDE_DIR=/misc/home/ikuma/fftw-3.3.8/include/

And got an error message;

Could not find fftw3f library named libfftw3f, please specify its location in CMAKE_PREFIX_PATH or FFTWF_LIBRARY by hand (e.g. -DFFTWF_LIBRARY=’/path/to/libfftw3f.so’)
CMake Error at cmake/gmxManageFFTLibraries.cmake:92 (MESSAGE):
Cannot find FFTW 3 (with correct precision - libfftw3f for mixed-precision
GROMACS or libfftw3 for double-precision GROMACS). Either choose the right
precision, choose another FFT(W) library (-DGMX_FFT_LIBRARY), enable the
advanced option to let GROMACS build FFTW 3 for you
(-DGMX_BUILD_OWN_FFTW=ON), or use the really slow GROMACS built-in fftpack
library (-DGMX_FFT_LIBRARY=fftpack).
Call Stack (most recent call first):
CMakeLists.txt:715 (include)

– Configuring incomplete, errors occurred!
See also “/misc/home/ikuma/gromacs-2020.2/build/CMakeFiles/CMakeOutput.log”.
See also “/misc/home/ikuma/gromacs-2020.2/build/CMakeFiles/CMakeError.log”.

Following the error message , I tried;

…/…/cmake/cmake-3.9.6-Linux-x86_64/bin/cmake … -DCMAKE_C_COMPILER=/misc/home/ikuma/openmpi-4.0.5/bin/mpicc -DCMAKE_CXX_COMPILER=/misc/home/ikuma/openmpi-4.0.5/bin/mpicxx -DGMX_MPI=on -DCMAKE_INSTALL_PREFIX=/misc/home/ikuma/gromacs-2020.2/ -DGMX_FFT_LIBRARY=fftw3 -DFFTWF_LIBRARIY=/misc/home/ikuma/fftw-3
.3.8/lib/libfftw3f_mpi.so -DFFTWF_INCLUDE_DIR=/misc/home/ikuma/fftw-3.3.8/include/

where the sentence -DGMX_FFT_LIBRARIES was changed to -DGMX_FFT_LIBRARY.

And I got;

CMake Error at cmake/FindFFTW.cmake:105 (message):
Could not find fftwf_plan_many_[r2c|c2r] in
/usr/local/fftw/3.3.8/lib/libfftw3f_mpi.so, take a look at the error
message in /misc/home/ikuma/gromacs-2020.2/build/CMakeFiles/CMakeError.log
to find out what went wrong. If you are using a static lib (.a) make sure
you have specified all dependencies of fftw3f in FFTWF_LIBRARY by hand
(e.g. -DFFTWF_LIBRARY=’/path/to/libfftw3f.so;/path/to/libm.so’) !
Call Stack (most recent call first):
cmake/gmxManageFFTLibraries.cmake:89 (find_package)
CMakeLists.txt:715 (include)

But the directory “/usr/local/” needs root , which I don’t have.

what I’d like to know is ;

・how to solve this problem (meaning how to configure GROMACS2020-2 with libfftw3f_mpi.so), if it is possible, without root

or
・any other method to install GROMACS with PLUMED+MPI

Could someone help me?

Thank you,
kimu handexyin

2

BS"D

You still seem to have a typo in your path specification:
DFFTWF_LIBRARIY
instead of
DFFTWF_LIBRARY

Harry

3

There is no need to compile FFTW with threading or MPI support, but it does no harm.

4

Dear all,
Thank you for your advises. I tried the following command;

…/…/cmake/cmake-3.9.6-Linux-x86_64/bin/cmake … -DCMAKE_C_COMPILER=/misc/home/ikuma/openmpi-4.0.5/bin/mpicc -DCMAKE_CXX_COMPILER=/misc/home/ikuma/openmpi-4.0.5/bin/mpicxx -DGMX_MPI=on -DCMAKE_INSTALL_PREFIX=/misc/home/ikuma/gromacs-2020.2/ -DGMX_BUILD_OWN_FFTW=ON

-DGMX_BUILD_FFTW_OWN is added according to your advise that fftw compiling does not affect on threading/MPI support.

Configuration and command “make” were done successfully, and then I tried “make check” , and got some messages like this;

NOTE: The number of threads is not equal to the number of (logical) cores
and the -pin option is set to auto: will not pin threads to cores.
This can lead to significant performance degradation.
Consider using -pin on (and -pinoffset in case you run multiple jobs).
starting mdrun ‘spc-and-methanol’
2 steps, 0.0 ps.
Determining Verlet buffer for a tolerance of 0.005 kJ/mol/ps at 298 K
Calculated rlist for 1x1 atom pair-list as 1.025 nm, buffer size 0.025 nm
Set rlist, assuming 4x4 atom pair-list, to 1.022 nm, buffer size 0.022 nm
Note that mdrun will redetermine rlist based on the actual pair-list setup
This run will generate roughly 0 Mb of data

Writing final coordinates.

           Core t (s)   Wall t (s)        (%)
   Time:        0.159        0.080      199.9
             (ns/day)    (hour/ns)

Performance: 3.260 7.362
[ OK ] MdrunCanWrite/NptTrajectories.WithDifferentPcoupl/2 (84 ms)
[----------] 3 tests from MdrunCanWrite/NptTrajectories (252 ms total)

[----------] Global test environment tear-down
[==========] 84 tests from 21 test cases ran. (18489 ms total)
[ PASSED ] 83 tests.
[ FAILED ] 1 test, listed below:
[ FAILED ] MdrunTest.WritesHelp

1 FAILED TEST
YOU HAVE 12 DISABLED TESTS

and like this separated form the above;

98% tests passed, 1 tests failed out of 52

Label Time Summary:
GTest = 131.87 sec (52 tests)
IntegrationTest = 45.68 sec (9 tests)
MpiTest = 85.27 sec (8 tests)
SlowTest = 70.40 sec (2 tests)
UnitTest = 15.79 sec (41 tests)

Total Test time (real) = 131.94 sec

The following tests FAILED:
43 - MdrunTests (Failed)
Errors while running CTest
make[3]: *** [CMakeFiles/run-ctest-nophys] Error 8
make[2]: *** [CMakeFiles/run-ctest-nophys.dir/all] Error 2
make[1]: *** [CMakeFiles/check.dir/rule] Error 2
make: *** [check] Error 2

do you think I can continue the installation? Or, should I fix something?
If you know how to solve it or encounter this type of error before, please advise me.

Best Regards,
kimu

5

This test fails, because PLUMED adds the new command line option -plumed to gmx mdrun but doesn’t update the expected output for the test (which is basically running gmx mdrun -h) nor is the plumed patch disabling the test.

I think this is rather an issue to be brought up with the PLUMED developers, however I would also really appreciate some guidance from the GROMACS team on how to (easily) disable/skip/ignore specific tests that I have determined to be non-critical.