[Wien] segmentation fault

[Chandrika]

Hi Maurizio,
In bash.rc (or equivalent) insert: ulimit -s unlimited.
In lapw0para and x_lapw insert: limit stacksize unlimited.
For R_LIB options you need to give the full path of the mkl compiler library. In the L option full path of fortran lib is needed. 
In LD_LIBRARY_PATH did you add path for the fortran library?
Best wishes,
Chandrika
  ----- Original Message ----- 
  From: Maurizio Biasini 
  To: A Mailing list for WIEN2k users 
  Sent: Wednesday, August 15, 2007 12:38 AM
  Subject: [Wien] segmentation fault



  Subject: Segmentation fault with ifort compiler


  Dear WIEN2k users
  I  downloaded the
   wien2k_07.2 on a new linux system
  linux version
  2.6.18.2-34-default 

  The fortran compiler selected in ./siteconfig  was 
       J    Linux (Intel ifort 8 or 9 compiler + mkl 8.0 )

  However the version I downloaded was a more recent one
  compiler 10.0.023
  the  pertaining lapack library  is 9.1.021
  /opt/intel/mkl/9.1.021/lib 
  _serial/em64t/libmkl_lapack.a

  The compiler options  in ./siteconfig were 
     Current settings:
       O   Compiler options:        -FR -mp1 -w -prec_div -pc80 -pad -ip -DINTEL_VML
       L   Linker Flags:            $(FOPT) -L/opt/intel/mkl/9.1.021/lib/em64t -lpthread
       P   Preprocessor flags       '-DParallel' 
       R   R_LIB (LAPACK+BLAS):     -lmkl_lapack -lmkl_em64t -lvml -pthread

  I followed the message from ./siteconfig 
  "Please note: The present ifort 8 or 9 versions cannot produce static executables
  therefore dynamic linking is done. You must have a proper LD_LIBRARY_PATH 
  on all your machines"
  I.e.   I set 
  export LD_LIBRARY_PATH=/opt/intel/mkl/9.1.021/lib/em64t

  After these steps the compilation did succeed.
  For example, these are first lines of the Makefile of lapw2 
  .SUFFIXES:      .F
  .SUFFIXES:      .F90
  SHELL = /bin/sh
  FC = ifort
  MPF = ifort
  CC = cc
  FOPT =  -FR -mp1 -w -prec_div -pc80 -pad -ip -DINTEL_VML
  FPOPT =  -FR -mp1 -w -prec_div -pc80 -pad -ip -DINTEL_VML 
  DParallel = '-DParallel'
  FGEN = $(PARALLEL)
  LDFLAGS = $(FOPT) -L/opt/intel/mkl/9.1.021/lib/em64t -lpthread
  R_LIBS = -lmkl_lapack -lmkl_em64t -lvml -pthread
  C_LIBS = $(R_LIBS)
  RP_LIBS = $(R_LIBS) -L /usr/local/SCALAPACK -L /usr/local/BLACS/LIB -lpblas -lredist -ltools -lscalapack -lfblacs -lblacs -lmpi 
  CP_LIBS = $(RP_LIBS)
  DESTDIR = .


  I  am running  an antiferromagnetic case; I could provide further information if needed
  System URu2Si2
  I first run  a spin polarized case (without the clmcopy option) WITHOUT SPIN ORBIT 
  It was successuful.  I got convergence and total moment consistent with AF 
  spin moment:   U1 1.4 mu_B  and  in U2 = -1.4 mu_B
  However when I tried to add spin orbit I got a message 
  "segmentation fault" in lapw2 
  I add that  I tried on another computer the same case , with the same input files , with s.o. without ever getting any "segmentation fault" message.

  this is the *.dayfile output
  >   lapw0       (11:35:06) 5.780u 0.140s 0:06.02 98.3%  0+0k 0+0io 0pf+0w
  >   lapw1  -up          (11:35:12) 1.440u 0.136s 0:01.81 86.7%  0+0k 0+0io 0pf+0w
  >   lapw1  -dn          (11:35:14) 1.444u 0.104s 0:01.54 100.0% 0+0k 0+0io 0pf+0w 
  >   lapwso -up          (11:35:15) 1.432u 0.076s 0:01.83 81.9%  0+0k 0+0io 31pf+0w
  >   lapw2 -c -up -so    (11:35:17) Segmentation fault

  I  obviously suspect that the reason is the compiler. Any help from people who used this compiler are strongly appreciated. 
  Thank you for help
  please reply at 
  biasini at physics.ucr.edu

  I include case.struct, case.in1 , case.in2, case.inso in a zip files
  best wishes

  Maurizio Biasini
  H  522 Athens str 92507 Riverside CA
  tel +1 951 7877826
  University of California at Riverside 
  Riverside  92521 CA (USA)
  tel +1 951 827 2919
  FAX +1 951 827 4529

  e-mail biasini at physics.ucr.edu
-- 
Maurizio Biasini, PhD
University of California at Riverside
3401 Watkins Drive 
Physics Department
Riverside  92521 CA (USA)
tel +1 951 827 2919
FAX +1 951 827 4529
E-MAil biasini at physics.ucr.edu 

Permanent address 
ENEA via Don Fiammelli 2 
40129 Bologna Italy tel +39 051 6098155 FAX +39 051 6098062
e-mail biasini at risc990.bologna.enea.it

Home page	 http://diaf3.bologna.enea.it/biasini/home.html





------------------------------------------------------------------------------


  _______________________________________________
  Wien mailing list
  Wien at zeus.theochem.tuwien.ac.at
  http://zeus.theochem.tuwien.ac.at/mailman/listinfo/wien
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://zeus.theochem.tuwien.ac.at/pipermail/wien/attachments/20070816/e8863b97/attachment.html