[Wien] Error: undefined symbol: __intel_cpu_indicator

Arunkumar S. Bommannavar abomm at hpcat.aps.anl.gov
Sat Mar 4 00:03:27 CET 2006


Peter,

Thank you very much for your time.

Intel support tells me the following:

"The __intel_cpu_indicator and __intel_cpu_indicator_init are defined in libimf.so and not libimf.a, with the new 9.0 compilers. So if you are building your application please don't use the switch -static or -i-static."

I'll keep you posted.

Regards
Arun


-----Original Message-----
From: wien-bounces at zeus.theochem.tuwien.ac.at [mailto:wien-bounces at zeus.theochem.tuwien.ac.at] On Behalf Of Peter Blaha
Sent: Wednesday, March 01, 2006 12:39 PM
To: A Mailing list for WIEN2k users
Subject: Re: [Wien] Error: undefined symbol: __intel_cpu_indicator

I just experimented today with the latest free ifort version (on an P4emt 
system with SUSE 9.3).

using -i-dynamic   and running lapw1c I got exactly the error message 
above.

It can be avoided using 
       -static-libcxa
              Links the Intel libcxa C++ library statically. This option 
              is the opposite of  -shared-libcxa.
(which is from the most recent "man ifort")

The ifort release notes also state, that depending on the memory-model 
(medium or large) -i-dynamic is necessary and -i-static would not work properly.
(Nice that it still links without error, but crashes later.

Thus I would recommend (on em64 systems) NOT to us -i-static.

Regards

> It seems not to be a subroutine that is in any of the static Intel libraries.
> the cpuid is usually set in libmkl_ia32 (old Pentium or Xeon), libmkl_i2p 
> (Itanium), or libmkl_em64t (new Pentium or Xeon).
> However, the subroutine may be defined with different name in the object file. 
> In this case I expect that the wrong dynamic libryrry may cause the error.
> 
> Anyway, I do not see why to link lapack dynamically but not blas, both can be 
> linked savely static.
> The only problem with dynamic linking is libpthread.
> 
> -static-libcxa is not longer an option if following the Intel compiler manual, 
> is it a regular Linux option ?
> 
> Ciao Gerhard
> 
> Am Mittwoch, 1. März 2006 17:34 schrieb Peter Blaha:
> > 
> > Another possible hint:    
> > 
> > replace the linker option   -i-static    by   -static-libcxa
> > 
> > 
> > > Subject: Re: [Wien] Error: undefined symbol: __intel_cpu_indicator 
> > > 
> > > That's one of the intel libraries which is not in your LD_LIBRARY_PATH
> > > (if 
> > > you have that). Do ldd /usr/local/wien2k/lapw1 ; you may have to go into
> > > a 
> > > csh and/or batch it to replicate the issue. -i-static during compilation
> > > 
> > > may help.
> > > 
> > > On Tue, 28 Feb 2006, Arunkumar S. Bommannavar wrote:
> > > 
> > > > Hello,
> > > >
> > > >
> > > >
> > > > My system is as follows:
> > > >
> > > >
> > > >
> > > > CPU : Intel(R) Xeon(TM) CPU 2.66GHz
> > > >
> > > > OS   : Linux Redhat 9.0
> > > >
> > > >
> > > >
> > > > Installed Intel Fortran 9.0 , MKL 8.0, and WIEN2k_06. I was able to
> > > > compile it and then tried TiC example. During "run SCF" stage,  I get
> > > > the following error message.
> > > >
> > > >
> > > >
> > > > LAPW0 END
> > > >
> > > > /usr/local/wien2k/lapw1: relocation error: /usr/local/wien2k/lapw1:
> > > > undefined symbol: __intel_cpu_indicator
> > > >
> > > >
> > > >
> > > > What am I doing wrong? Please let me know what other information I
> > > > should supply to identify the problem.
> > > >
> > > >
> > > >
> > > > Regards
> > > >
> > > > Arun S. Bommannavar
> > > >
> > > >
> > > >
> > > > Phone: (630) 252-0497
> > > >
> > > > e-mail : abomm at hpcat.aps.anl.gov
> > > >
> > > >
> > > >
> > > >
> > > 
> > > Note: if you have an old email address for me, please note that "nwu"
> > > has
> > > been changed to "northwestern".
> > > -----------------------------------------------
> > > Laurence Marks
> > > Department of Materials Science and Engineering
> > > MSE Rm 2036 Cook Hall
> > > 2220 N Campus Drive
> > > Northwestern University
> > > Evanston, IL 60201, USA
> > > Tel: (847) 491-3996 Fax: (847) 491-7820
> > > email: L-marks at northwestern dot edu
> > > http://www.numis.northwestern.edu
> > > -----------------------------------------------
> > > 
> > > _______________________________________________
> > > Wien mailing list
> > > Wien at zeus.theochem.tuwien.ac.at
> > > http://zeus.theochem.tuwien.ac.at/mailman/listinfo/wien
> > > _______________________________________________
> > > Wien mailing list
> > > Wien at zeus.theochem.tuwien.ac.at
> > > http://zeus.theochem.tuwien.ac.at/mailman/listinfo/wien
> > > 
> > 
> > 
> >                                       P.Blaha
> > --------------------------------------------------------------------------
> > Peter BLAHA, Inst.f. Materials Chemistry, TU Vienna, A-1060 Vienna
> > Phone: +43-1-58801-15671             FAX: +43-1-58801-15698
> > Email: blaha at theochem.tuwien.ac.at    WWW: 
> http://info.tuwien.ac.at/theochem/
> > --------------------------------------------------------------------------
> > _______________________________________________
> > Wien mailing list
> > Wien at zeus.theochem.tuwien.ac.at
> > http://zeus.theochem.tuwien.ac.at/mailman/listinfo/wien
> > 
> _______________________________________________
> Wien mailing list
> Wien at zeus.theochem.tuwien.ac.at
> http://zeus.theochem.tuwien.ac.at/mailman/listinfo/wien
> 


                                      P.Blaha
--------------------------------------------------------------------------
Peter BLAHA, Inst.f. Materials Chemistry, TU Vienna, A-1060 Vienna
Phone: +43-1-58801-15671             FAX: +43-1-58801-15698
Email: blaha at theochem.tuwien.ac.at    WWW: http://info.tuwien.ac.at/theochem/
--------------------------------------------------------------------------
_______________________________________________
Wien mailing list
Wien at zeus.theochem.tuwien.ac.at
http://zeus.theochem.tuwien.ac.at/mailman/listinfo/wien


More information about the Wien mailing list