[Pw_forum] /opt/intel/Compiler/11.0/069/lib/intel64/libimf.so: warning: warning: feupdateenv is not implemented and will always fail

[Axel Kohlmeyer]

On Wed, 19 Nov 2008, vega lew wrote:

VL> 
VL> Dear sir,
VL>  
VL> > oh my. Yet another version of the Intel compiler, with new and> better bugs
VL> which compiler was best and stable for QE? 9.1? 9.0? and what about the mkl?

what on earth makes you think that the compiler is the culprit
of your problems. as paolo suggested, you have to provide _proof_
(to yourself mostly) of what is going wrong. 

there are a bunch of issues with intel compilers that have been
discussed here many, many times. if one takes them all into account
there is no harm in using intel. various versions have been working
find for me on a large number of machine and also for codes that are
much more demanding in terms of correctness of the implementation of
fortran95 than Q-E. the key issue is to know what you are doing
and to not push the compiler too far (i.e. reduce the optimization
level and avoid excessive IPO or SSE optimizations). moreover, there
are additional complications with MKL 10 due to the way it multi-threads
across available processors unless OMP_NUM_THREADS is set to 1 on
all machines or one uses the proper linking sequence. all of that
has been discussed here many times and is well documented in the
intel provided documentation.

my suspicion is more that you are running either out of stack memory
or in general out of memory, but that is something that you can
monitor easily (and information of i am using x atoms, or y cutoff
does not help, there are many more parameters that determine how
much memory a calculation requires) and test systematically. since 
you don't provide crucial details that allow any qualified statements 
you'll have to learn to live with no proper help.
my prediction is that randomly swapping compilers or MPI libraries
will only increase your (and our) confusion, but not provide
any proper resolution.

cheers,
   axel.



VL>  
VL> if intel was not good enough. Do you think gfortran or other compilers would be better for QE on intel computers?
VL>  
VL> vega
VL> Vega Lew (weijia liu)PH.D Candidate in Chemical EngineeringState Key Laboratory of Materials-oriented Chemical EngineeringCollege of Chemistry and Chemical EngineeringNanjing University of Technology, 210009, Nanjing, Jiangsu, China> Date: Tue, 18 Nov 2008 18:02:22 +0100> From: giannozz at democritos.it> To: pw_forum at pwscf.org> Subject: Re: [Pw_forum] /opt/intel/Compiler/11.0/069/lib/intel64/libimf.so: warning: warning: feupdateenv is not implemented and will always fail> > vega lew wrote:> > > I recompile the QE 4.0.3 with intel compiler 11.0.> > oh my. Yet another version of the Intel compiler, with new and> better bugs> > > I found there was an warning during the "make all" process.> > the warning was like this,> > "/opt/intel/Compiler/11.0/069/lib/intel64/libimf.so: warning: warning: > > feupdateenv is not implemented and will always fail"> > Do you think the warning will have a bad effect on the QE?> > no, no effect at all> > Paolo> -- > Paolo Giannozzi, Democritos and!
  University of Udine, Italy> _______________________________________________> Pw_forum mailing list> Pw_forum at pwscf.org> http://www.democritos.it/mailman/listinfo/pw_forum
VL> _________________________________________________________________
VL> Invite your mail contacts to join your friends list with Windows Live Spaces. It's easy!
VL> http://spaces.live.com/spacesapi.aspx?wx_action=create&wx_url=/friends.aspx&mkt=en-us

-- 
=======================================================================
Axel Kohlmeyer   akohlmey at cmm.chem.upenn.edu   http://www.cmm.upenn.edu
   Center for Molecular Modeling   --   University of Pennsylvania
Department of Chemistry, 231 S.34th Street, Philadelphia, PA 19104-6323
tel: 1-215-898-1582,  fax: 1-215-573-6233,  office-tel: 1-215-898-5425
=======================================================================
If you make something idiot-proof, the universe creates a better idiot.