Scribd
Upload
489 views12 pages

WIEN2k Installation

WIEN2k runs on Linux platforms from PCs to supercomputers. It uses Fortran90 and is parallelized using k-point parallelization and MPI. Users must register, download the software, compile it for their system, and configure their environment and directories. Help is available through the user guide, website FAQ, and mailing list.

Uploaded by

Mohamed-Amine Talbi
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
489 views12 pages

WIEN2k Installation

WIEN2k runs on Linux platforms from PCs to supercomputers. It uses Fortran90 and is parallelized using k-point parallelization and MPI. Users must register, download the software, compile it for their system, and configure their environment and directories. Help is available through the user guide, website FAQ, and mailing list.

Uploaded by

Mohamed-Amine Talbi
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 12

WIEN2k- hardware/software

 WIEN2k runs on any Linux platform from PCs, Macs,


workstations, clusters to supercomputers
 Intel I7 quad (six)-core processors with fast memory bus (1.5-3
Gb/core, Gbit-network, SATA disks). 1000 € /PC,
 with a few such PCs you have a quite powerful cluster (k-parallel)

 60 - 100 atom / cell, requires 2-4 Gb RAM

 installation support for many platforms + compiler

 Fortran90 (dynamical allocation, modules)


 real/complex version (inversion)
 many individual modules, linked together with C-shell or perl-scripts

 web-based GUI – w2web (perl)

 f90 compiler (ifort, gfortran), BLAS-library (mkl, gotolib),


FFTW, perl5, ghostscript (+jpg), gnuplot(+png), Tcl/Tk
(Xcrysden), pdf-reader, www-browser, octave, opendx
Installation of WIEN2k
 Register via http://www.wien2k.at
 Create your $WIENROOT directory (e.g. ./WIEN2k )
 Download wien2k_13.tar and examples (executables)
 Uncompress and expand all files using:
 tar –xvf wien2k_12.tar

 gunzip *.gz

 ./expand_lapw

 This leads to the following directories:


 ./SRC (scripts, ug.ps)
 ./SRC_aim (programs)
 …

 SRC_templates (example inputs)


 …

 SRC_usersguide_html (HTML-version of UG)


 example_struct_files (examples)
 TiC
siteconfig_lapw
 ****************************************************
 * W I E N *
 * site configuration *
 ****************************************************
 S specify a system
 C specify compiler
 O specify compiler options, BLAS and LAPACK
 P configure Parallel execution
 D Dimension Parameters
 R Compile/Recompile
 U Update a package
 L Perl path (if not in /usr/bin/perl)
 Q Quit
D: define NMATMAX (adjust to your hardware/paging!):
NMATMAX=5000 256Mb (real) or 500Mb (complex)
NMATMAX=10000  1Gb (real) or 2Gb (complex)  80-100 atoms/unitcell
NUME=1000  number of eigenvalues (adjust to NMATMAX)
Compilation

 recommendation: Intels Fortran compiler (includes mkl)


free for non-commercial (but not for academic), www intel.com
 which ifort  tells you if you can use ifort and which version you have
 usually installed in /opt/intel/composerxe-2011…./bin/intel64 (ls ….)
 include ifortvars.csh and mklvars.csh in your .bashrc/.cshrc file:
 source /opt/intel/11.0/074/bin/ifortvars.csh intel64
 source /opt/intel/11.0/074/mkl/tools/environment/mklvarsem64t.csh
 ifort 12 (vers. 8.0 and early 12.x buggy, 9.x, 10.0, 11.x ok)
 for older versions dynamic linking recommended (depends on ifort version,
requires system and compiler libraries at runtime, needs $LD_LIBRARY_PATH)
 IA32 bit, IA64 bit (Itanium) or Intel64 (em64t) -version
 mkl-library: names change with every version, see:
http://software.intel.com/en-us/articles/intel-mkl-link-line-advisor
 9.x: -L/opt/intel/mkl/lib -lmkl_lapack –lmkl_em64t -lmkl_core (libmkl_core.so)
 10.0: -L/opt/intel/mkl/lib -lmkl_lapack –lmkl
 compiler/linker options depend on compiler version + Linux-version !!
 -FR (free format) -lguide –lpthread -pthread
compilation
 gfortran + gotolib, acml-lib, ATLAS-BLAS
 -static linking possible

 siteconfig has support for various ifort versions and gfortran

 it does NOT make sense to invest in new hardware but use a


„free“ compiler
userconfig_lapw
 Every user should run userconfig_lapw (setup of environment)
 support for tcsh and bash (requires .cshrc or .bashrc)
 sets PATH to $WIENROOT, sets variables and aliases

 $WIENROOT, $SCRATCH, $EDITOR, $PDFREADER, $STRUCTEDIT_PATH


 pslapw: ps –ef | grep lapw
 lsi: ls –als *.in* lso: ls -als *.output*
 lss: *.scf* lsc: *.clm*
 limit stacksize unlimited (otherwise: “segmentation fault”)

 $OMP_NUM_THREADS (for mkl+multi-core); $LD_LIBRARY_PATH


w2web
 w2web: acts as webserver on a userdefined (high) port.
 define user/password and port. (http://host.domain.xx:5000)
 behind firewall create a „ssh-tunnel“: ssh -fNL 2000:host:2000 user@host

 ~/.w2web/hostname/conf/w2web.conf: (configuration file)

 deny=*.*.*.*
 allow=128.130.134.* 128.130.142.10
 define execution types: NAME=commands (eg.: batch=batch < %f)
k-point Parallelization (lapw1+lapw2)

 very efficient parallelization even on loosely coupled PCs (slow network):


 common NFS filesystem (files must be accessible with the same path on
all machines; use /host1 as data-directory on host1)
 rsh/ssh without password (.rhosts; private/public keys)

 ssh-keygen –t rsa
 append .ssh/authorized_keys on remote host with id_rsa.pub of local host
 .machines file:
 1:host1 (speed:hostname)
 2:host2
 granularity:1 (1:10k+20k; 3: 3+6+3+6+3+6+rest load balancing,
not with $SCRATCH, -it
 extrafine:1 (rest in junks of 1 k)
 testpara (tests distribution); run_lapw –p
 case must fit into memory of one PC !
 high NFS load: use local $SCRATCH directory (only with commensurate k-
points/hosts)
 $OMP_NUM_THREADS (parallel diag. with mkl on multi-core CPU)
Flow of parallel execution

lapw1para lapw2para
fine-grain mpi-parallelization

 for bigger cases (> 50 atoms) and more than 4 cores


 fast network (Gbit, Myrinet, Infiniband, shared memory machines)

 mpi (you need to know which mpi is installed (mpich-1.2, open-mpi, intel-mpi,…)
 mpif90 or mpiifort

 scalapack (included in ifort 11)


 llibmkl_blacs_lp64.a or ibmkl_blacs_openmpi_lp64.a or libmkl_blacs_intelmpi_lp64.a
 FFTW (v. 2 or 3 ; mpi and sequ. version needed, -DFFTW2/3 in Makefiles)
 .machines file:
 1:host1:4 host2:4 8 mpi-parallel jobs on host1 and host2
 lapw0:host1:4 host2:4 8 parallel jobs; atom-loops only + fft !!!

 simultaneous k-point and


mpi-parallelization possible

 BN/Rh(111) nanomesh:
cell with 1100 atoms
 NMAT=45000-80000; 64 cpus, 2h / iteration; scales to at least 512 cores
WIEN2k_13.1

 always use latest version (bug fixes, improved performance,


new features)

 eventually: use prebuilt executables from our website !!


Getting help

 *_lapw –h „help switch“ of all WIEN2k-scripts


 help_lapw:
 opens usersguide.pdf; Use ^f keyword to search for an item („index“)
 html-version of the UG: ($WIENROOT/SRC_usersguide/usersguide.html)
 http://www.wien2k.at/reg_user
 FAQ page with answers to common questions
 Update information: When you think the program has an error, please
check newest version
 Textbook section: DFT and the family of LAPW methods by S.Cottenier

 Mailing-list:

 subscribe to the list (always use the same email)


 full text search of the „digest“ (your questions may have been answered
before)
 posting questions: Provide sufficient information, locate your problem
(case.dayfile, *.error, case.scf, case.outputX).
 „My calculation crashed. Please help.“ This will most likely not be answered.

You might also like