[Pw_forum] Compiling with intel fortran, MKL and MKL FFT wrappers
Axel Kohlmeyer
akohlmey at cmm.chem.upenn.edu
Tue Sep 30 15:36:31 CEST 2008
On Tue, 30 Sep 2008, vega wrote:
VL> Dear Serge,
VL>
VL> I'm very anxious to know how to link the intel FFTW in MKL to the QE.
just read the MKL documentation. it is explained in detail. RTFM.
but then again, it is a rather pointless exercise. FFTW works fine
and very reliable and MKL's fft (through fftw) is rarely faster and
crashes occasionally.
axel.
VL> >From your make.sys, I know you might link the intel FFTW to QE successfully.
VL>
VL> could you tell me which dir should be added to the LD_LIBRARY_PATH, or which
VL> flags should be added after the QE configure command?
VL>
VL> thank you
VL>
VL> vega
VL>
VL> =================================================================================
VL> Vega Lew (weijia liu)
VL> PH.D Candidate in Chemical Engineering
VL> State Key Laboratory of Materials-oriented Chemical Engineering
VL> College of Chemistry and Chemical Engineering
VL> Nanjing University of Technology, 210009, Nanjing, Jiangsu, China
VL>
VL> --------------------------------------------------
VL> From: "Serge Nakhmanson" <nakhmanson at anl.gov>
VL> Sent: Tuesday, September 30, 2008 7:48 AM
VL> To: "PWSCF Forum" <pw_forum at pwscf.org>
VL> Subject: [Pw_forum] Compiling with intel fortran, MKL and MKL FFT wrappers
VL>
VL> > Dear All,
VL> >
VL> > Here's a question to those of you who succeeded doing the Subj.
VL> >
VL> > Here's how MKL distribution looks like on my cluster:
VL> >
VL> > serafina {1323} > pwd
VL> > /usr/local/intel/mkl/10.0.4.023/interfaces
VL> > serafina {1324} > ls
VL> > blas95 fftw2xc fftw2x_cdft fftw2xf fftw3xc fftw3xf lapack95
VL> > serafina {1325} >
VL> >
VL> > We (me and our sysadmin) compiled all versions of the MKL FFT
VL> > wrappers, but we assumed that the parallel one (fftw2x_cdft)
VL> > is the one to be used when compiling Espresso.
VL> >
VL> > However, linking with that one produces a bunch of undefined
VL> > references in {espresso}/clib/fft_stick.c.
VL> >
VL> > When, instead, we link with fftw2xc (which is serial), everything
VL> > compiles and works just fine.
VL> >
VL> > My question is: is the latter (serial) library really the
VL> > one intended to be used with Espresso, or is it the parallel
VL> > one that should work but fails due to some function renaming
VL> > in the latest MKL version? Or is there something else that
VL> > could be wrong?
VL> >
VL> > Just in case, here are a few informative lines from our make.sys:
VL> >
VL> > DFLAGS = -D__INTEL -D__FFTW -D__MPI -D__PARA
VL> > IFLAGS
VL> > = -I../include -I/usr/local/intel/mkl/10.0.4.023/include/fftw
VL> > BLAS_LIBS
VL> > = -L/usr/local/intel/mkl/10.0.4.023/lib/em64t -lmkl_intel_lp64
VL> > -lmkl_sequential -lmkl_core
VL> > LAPACK_LIBS
VL> > = -L/usr/local/intel/mkl/10.0.4.023/lib/em64t -lmkl_intel_lp64
VL> > -lmkl_sequential -lmkl_core
VL> > #FFT_LIBS
VL> > = -L/usr/local/intel/mkl/10.0.4.023/lib/64 -lfftw2x_cdft_DOUBLE
VL> > FFT_LIBS = -L/usr/local/intel/mkl/10.0.4.023/lib/64 -lfftw2xc_intel
VL> >
VL> >
VL> > THX,
VL> >
VL> > Serge
VL> >
VL> >
VL> > --
VL> > *********************************************************
VL> > Serge M. Nakhmanson phone: (630) 252-5205
VL> > Assistant Scientist fax: (630) 252-4798
VL> > MSD-212, Rm. C-224
VL> > Argonne National Laboratory
VL> > 9700 S. Cass Ave.
VL> > Argonne, IL 60439
VL> > *********************************************************
VL> > _______________________________________________
VL> > Pw_forum mailing list
VL> > Pw_forum at pwscf.org
VL> > http://www.democritos.it/mailman/listinfo/pw_forum
VL> >
VL> _______________________________________________
VL> Pw_forum mailing list
VL> Pw_forum at pwscf.org
VL> http://www.democritos.it/mailman/listinfo/pw_forum
VL>
--
=======================================================================
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.
More information about the users
mailing list