libint error for HFX (error: CP2K and libint were compiled with different LIBINT_MAX_AM)

Alfio Lazzaro alfio.... at gmail.com
Mon May 7 08:43:00 UTC 2018


Ciao Daniele,
The problem is that (as the message says) CP2K was compiled with a given 
set of Angular momentum values for LIBINT, for instance:

-D__LIBINT_MAX_AM=6 -D__LIBDERIV_MAX_AM1=5

while the package LIBINT was compiled with a different set of values 
(usually higher values). Therefore, this is not a bug on CP2K, rather a 
compilation error on the debian packages.
I don't know about the debian package, my assumption is that libint is 
shared between several packages, which can use different AM values.
You should better refer to Debian people on how to solve the problem...

Alfio






Il giorno giovedì 3 maggio 2018 19:20:33 UTC+2, Daniele Fontanari ha 
scritto:
>
> Hello everyone,
>
> On my university cluster I have CP2K installed (by the admins using the 
> packaged version for debian I think, I'm having problems to do it by 
> source) and If I try to run the test 
> "QS/regtest-hybrid-1/H2O-hybrid-pbe0.inp" I get the error in the title (and 
> the same if try to generate a simple input using pbe0)
>
> I can provide more informations if you tell me which command to run, for 
> the moment:
>
> $dpkg-query -l | grep -e "libint" -e "cp2k"
> ii  cp2k                                   
> 4.1-1+b1                          amd64        Ab Initio Molecular Dynamics
> ii  cp2k-data                              
> 4.1-1                             all          Ab Initio Molecular Dynamics 
> (data files)
> ii  libint1                                
> 1.1.6-2+b1                        amd64        Evaluate the integrals in 
> modern atomic and molecular theory
>
> and the header in the output file reads:
>
>  DBCSR| Multiplication 
> driver                                               BLAS
>  DBCSR| Multrec recursion 
> limit                                              512
>  DBCSR| Multiplication stack 
> size                                           1000
>  DBCSR| Maximum elements for images                                    
> UNLIMITED
>  DBCSR| Multiplicative factor virtual 
> images                                   1
>  DBCSR| Randmat seed                                                    
> 12341313
>  DBCSR| Multiplication size 
> stacks                                             3
>  DBCSR| Number of 3D layers                                               
> SINGLE
>  DBCSR| Use MPI memory 
> allocation                                              T
>  DBCSR| Use RMA 
> algorithm                                                      F
>  DBCSR| Use Communication 
> thread                                               T
>  DBCSR| Communication thread 
> load                                             87
> ...
> ...
> ...
>  CP2K| version string:                                          CP2K 
> version 4.1
>  CP2K| source code revision number:                                    
> svn:17462
>  CP2K| cp2kflags: libint fftw3 libxc elpa3 parallel mpi3 
> scalapack              
>  CP2K| is freely available from                            
> https://www.cp2k.org/
>  CP2K| Program compiled at                          Fri Mar 17 00:36:25 
> UTC 2017
>  CP2K| Program compiled on                                            
> x86-ubc-01
>  CP2K| Program compiled for                                
> Linux-x86_64-gfortran
>  CP2K| Data directory path                                       
> /usr/share/cp2k
>  CP2K| Input file name                                       
> H2O-hybrid-pbe0.inp
>
> I was wondering if the error is known/fixed (I really think that the 
> admins just installed cp2k without compiling it by hand through the package 
> manager) so it might be sufficient to re-download it or if there is a 
> possible workaround?
>
> Cheers
>
> Daniele 
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.cp2k.org/archives/cp2k-user/attachments/20180507/2dbb0a94/attachment.htm>


More information about the CP2K-user mailing list