unknow subsection ACID_HYDRONIUM_DISTANCE in metadynamics

[Matthias Krack]

Hi Jiasen

you can run your cp2k binary with the --xml flag, e.g. 
"Linux-x86-64-gfortran.popt --xml", which should generate a file 
"cp2k_input.xml". That file contains a full dump of the CP2K input file 
structure supported by that cp2k executable. Search (grep) for 
"ACID_HYDRONIUM" in that file.

The precompiled cp2k-6.1.ssmp binary is only OpenMP parallel but not MPI 
parallel and thus it does not run with mpirun/mpiexec. You can only employ 
multiple OpenMP threads, e.g. 4 threads using "export OMP_NUM_THREADS=4; 
cp2k-6.1.ssmp <input file>"

Matthias

On Tuesday, 10 July 2018 20:29:02 UTC+2, terr... at gmail.com wrote:
>
> HI all,
> I was trying to use the collective variable ACID_HYDRONIUM_DISTANCE as my 
> CV, but cp2k took this as an unknown subsection of COLVAR. Has anyone 
> successfully used this CV? or this is a problem originating from my 
> compilation, because i found that ACID_HYDRONIUM_SHELL is also not 
> recognized by my binary. Please kindly help me out of this, i appreciate 
> your help. I am running with cp2k-6.1. I am running on Ubuntu 14.04.4 LTS, 
> Here is the arch file i used for compilation on:
>
> Linux-x86-64-gfortran.popt
>
> # Tested with: GFortran 6.4, MPICH 3.2, LAPACK 3.5.0, ScaLAPACK 2.0.2
> CC         = gcc
> CPP        =
> FC         = mpif90
> LD         = mpif90
> AR         = ar -r
> MPI_LIBRARY_PATH   = /home/jiasen/lib
> LIBPATH    = /home/jiasen/lib
> FFTW_LIB   = /home/jiasen/lib/fftw-3.3.8/.libs
> FFTW_INC   = /home/jiasen/lib/fftw-3.3.8/api
> LIBINT_LIB = /home/jiasen/lib/libint-1.1.4/lib
> LIBINT_INC = /home/jiasen/lib/libint-1.1.4/include
> DFLAGS     = -D__FFTW3 -D__LIBINT -D__MPI_VERSION=1 -D__BLACS\
>              -D__LIBINT_MAX_AM=6 -D__LIBDERIV_MAX_AM1=5 -D__MAX_CONTR=4\
>              -D__parallel -D__SCALAPACK
> CPPFLAGS   = 
> FCFLAGS    = $(DFLAGS) -O2 -ffast-math -ffree-form -ffree-line-length-none\
>              -ftree-vectorize -funroll-loops\
>              -mtune=native\
>              -I$(FFTW_INC) -I$(LIBINT_INC) 
> LDFLAGS    = $(FCFLAGS) 
> LIBS       = $(MPI_LIBRARY_PATH)/libscalapack.so\
>              $(LIBPATH)/liblapack.so\
>              $(LIBPATH)/libblas.so\
>              $(FFTW_LIB)/libfftw3.so\
>              $(LIBINT_LIB)/libderiv.a\
>              $(LIBINT_LIB)/libint.a\
>              /home/jiasen/lib/libblacs-openmpi.so\
>
>
> (MAY BE i should open another topic for the following)
> Actually I would like to not compile cp2k-6.1 by myself were the 
> precompiled cp2k-6.1.ssmp could work for me on a CentOS cluster. I have 
> tried with openmpi-1.x.x and openmpi-2.x.x, but got the same problem. The 
> problem of using the precompiled .ssmp is that it runs the job N times 
> independently, instead of parallel. 
> the command I use is the following:
>
> mpirun -n N .ssmp -i input -o output
>
> if possible, please kindly guide me on how to run the precompiled 
> excutables on my cluster.
>
> Best Regards
> Jiasen Guo
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.cp2k.org/archives/cp2k-user/attachments/20180711/c0a383af/attachment.htm>