[Wien] Regarding non-availability of shared library fftw3

Gavin Abo gsabo at crimson.ua.edu
Mon Apr 24 14:10:47 CEST 2017


Sounds like the compilation is fine, but your system environment doesn't 
know the location of your fftw3 installation.

Search the mailing list [ 
http://www.mail-archive.com/wien@zeus.theochem.tuwien.ac.at/maillist.html 
], you should find things there that can help.

As I mentioned before, I don't recommend installing as root unless you 
know what your doing [ 
http://www.mail-archive.com/wien@zeus.theochem.tuwien.ac.at/msg10594.html ].

Usually, that involves setting an environmental variable in .bashrc, 
setting a location in a conf file, or build statically [ 
http://www.mail-archive.com/wien@zeus.theochem.tuwien.ac.at/msg11752.html 
, 
http://www.mail-archive.com/wien@zeus.theochem.tuwien.ac.at/msg08016.html ].

Somethings, it needs set in the mpirun command [ 
http://www.mail-archive.com/wien@zeus.theochem.tuwien.ac.at/msg10003.html 
] or job script [ For example, -V may be needed to pass the environment 
variables for a job: 
https://crc.ibest.uidaho.edu/help/Tutorials/Submitting_a_job_using_qsub.html 
].

I haven't used Rocks.  So I'm not familiar with how to set it for your 
system.  I suggest you check your Rocks documentation, or search or ask 
on the Rocks mailing list, because it looks like you might need to set 
/usr/local/fftw-3.3.6-pl2/mpi/.libs/ in a script in your /etc/profile.d/ 
[ 
https://lists.sdsc.edu/pipermail/npaci-rocks-discussion/2015-January/066630.html 
, 
https://lists.sdsc.edu/pipermail/npaci-rocks-discussion/2006-December/022795.html 
].

On 4/24/2017 1:37 AM, Krisna Swaroop Sharma wrote:
> Dear Prof. Blaha
> This is to bring to your notice that we are trying to set up a cluster 
> with five i5 desktops ( one server and 4 nodes) on ROCKS, so that 
> meaningful computations for bigger cells/ molecules could be taken up 
> on WIEN2K. All these systems are qudra-core machines. Hard disk on 
> server of 1TB and on nodes 500GB. The RAM on server and nodes is 4GB 
> each. We were able to successfully install ROCKS software and it is 
> working fine.
> We have also installed *WIEN2K_16.1* on the system as per guidelines 
> given in the User Guide and giving appropriate options during 
> compilation. WIEN2K was successfully compiled with no compilation 
> error, but when we checked it for TiC on server, then during 
> initialization it stops at 'd-start', giving the message shared 
> library 'fftw3' could not open.
>
>
> Commandline: x_dstart
>
>
> Program input is:””
>
>
> */root/Desktop/WIEN2k16/dstart: error while loading shared libraries: 
> libfftw3.so.3: cannot open shared object file: No such file or directory
> *
>
> *
> *
>
> **
>
> *0.000u0.000s0:00.00 0.0%0+0k0+0io0pf+0w*
>
> *
> *
>
> **
>
> *error :command/root/Desktop/WIEN2k16/dstart dstart.def failed*
>
>
>
> We understand that some mistake might have been committed during 
> compilation, but we are not able to find it out , in spite of several 
> attempts by  using different options.
> We are using icc, ifort, and f90 provided by the respective 
> manufacturers only a few days back.
>
> *COMPLIER:-**Intel® Parallel Studio XE Cluster Edition for Linux 2017
> *
>
> *
> *
>
> *mpi verson : mpiifort*
>
>
>  The compiler options and, parallel configuration are as follows :
>
> Current Settings
>
> O
>
> 	
>
> Complier options
>
> 	
>
> -01 –FR –mp1 –w –prec_div –pc80 –pad –ip –DINTEL_VML –traceback 
> –assume buffered_io –I$(MKLROOT)/include
>
> F
>
> 	
>
> FFTW options
>
> 	
>
> -DFFTW3 –I/usr/local/fftw-3.3.6-pl2/mpi.libs//include
>
> X
>
> 	
>
> LIBX options
>
> 	
>
> E
>
> 	
>
> ELPA options
>
> 	
>
> L
>
> 	
>
> Linker Flags
>
> 	
>
> $(FOPT) –L$(MKLROOT)/lib/$(MKL_TARGET_ARCH) -pthread
>
> P
>
> 	
>
> Preprocessor flags
>
> 	
>
> ‘-DParallel’
>
> R
>
> 	
>
> R_LIB(LAPACK+BLAS)
>
> 	
>
> -lmkl_lapack95_lp64 –lmkl_intel_lp64 –lmkl_intel_thread –lmkl_core 
> –openmp -lpthread
>
> FL
>
> 	
>
> FFTW LIBS
>
> 	
>
> -lfftw3_mpi –lfftw3 –L/usr/local/fftw-3.3.6-pl2/mpi/.libs/
>
> XL
>
> 	
>
> LIBX LIBS
>
> 	
>
> EL
>
> 	
>
> ELPA LIBS
>
> 	
>
> S
>
> 	
>
> Save and Quit
>
> 	
>
> Q
>
> 	
>
> Quit and abandon changes
>
> 	
>
>
>  After WIEN2K works well, we shall try submission of jobs through 
> script. I will be grateful if some useful tips are given to solve the 
> problem, so that WIEN2K could be successfully installed in parallel 
> version.
> Thanking you and with warm regards.
> Dr. K. S. Sharma
> Professor of Physics
> The IIS University, Jaipur (India)
> <sharma.krishnaswaroop at gmail.com <mailto:sharma.krishnaswaroop at gmail.com>>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://zeus.theochem.tuwien.ac.at/pipermail/wien/attachments/20170424/5e97be1c/attachment.html>


More information about the Wien mailing list