[Wien] parallel compilation

Jonas Baltrusaitis jasius_1 at yahoo.com
Tue Sep 15 14:51:11 CEST 2009


Peter,

are you implying that it is possible to run one job on 2-4 cores in 'parallel' without any mpi and granularization? Could you elaborate on that?

Jonas

--- On Tue, 9/15/09, Peter Blaha <pblaha at theochem.tuwien.ac.at> wrote:

> From: Peter Blaha <pblaha at theochem.tuwien.ac.at>
> Subject: Re: [Wien] parallel compilation
> To: "A Mailing list for WIEN2k users" <wien at zeus.theochem.tuwien.ac.at>
> Date: Tuesday, September 15, 2009, 1:33 AM
> > if you do not select fine
> grained parallel, siteconfig_lapw does not give you options
> to tune parallel parameters.  does that mean that it is
> not compiling parallel at all?  if i do not do fine
> tuning, i can get a complete compilation. 
> 
> If one does NOT select "fine grain parallel"; siteconfig
> will not produce any
> mpi-executables. This is clearly the proper choice
> 
> i) for beginners
> ii) on "small" computers, where you may not have
> installed   mpi + fftw  ;
>     or have only a slow network between them.
> iii) when you intend to limit your unitcells to about 50
> atoms/cell.
> 
> (Still k-point parallelization is possible, and this is
> much more efficient
> than mpi for small systems).
> 
> If you select fine grain parallelization, also the
> mpi-executables will be
> produced. But you need:
> 
> i) a properly installed mpi (+ mpi-fortran compiler, which
> is usually "created"
>    when mpi is installed. (Usually MPF is
> NOT  ifort, but   mpif90  !!)
>    Note: mpi2 needs some small code changes
> in lapw0.F
> ii) a corresponding Scalapack+Blacs (also suitable for your
> compiler)
>     (usually it comes together with the mkl)
> iii) FFTW routines (compiled with the mpi option.
> iv) you need to properly specify the "installation place"
> (-Lxxxx) and the
>     name of the libraries (-lxxxx). Also make
> sure you understand 32-bit vs
>     64bit ("emt") issues.
> 
> In addition, this code is useful ONLY if you have unitcells
> with more than
> about 50 atoms !!!! and you have hardware with fast network
> (Infiniband),
> i.e. Gbit ethernet may work, but will be slow.
> 
> On a SINGLE Quadcore-computer, mpi-parallelization is
> probably NOT worth the effort
> (use  OMP_NUM_THREAD 2 (or 4))
> 
> 
> linuxif9:MPF:ifort        ???? 
> unlikely! (see above)
> 
> 
> 
> > # Linux PC system with IFC 10 compiler + mkl 10 (-ip
> is broken; -static does not give traceback-lines)
> > linuxif9:FC:ifortlinuxif9:MPF:ifort
> > linuxif9:CC:cclinuxif9:FOPT:-FR -mp1 -w
> -prec_div  -pc80 -pad -align -DINTEL_VML -traceback -O
> > 3 -xW
> > linuxif9:FPOPT:$(FOPT) -FR -mp1 -w -prec_div -pc80
> -pad -align -DINTEL_VML -traceback
> -I/opt/mpich2/includelinuxif9:LDFLAGS:  $(FOPT)
> -L/opt/intel/mkl/10.2.1.017/lib/em64t <http://10.2.1.017/lib/em64t> -pthread -i-static
> > linuxif9:R_LIBS:-L/opt/intel/mkl/10.2.1.017/lib/em64t
> <http://10.2.1.017/lib/em64t> -lmkl_lapack -lmkl
> -liomp5
> >  -lguide -lmkl_core
> > linuxif9:DPARALLEL:'-DParallel'
> > linuxif9:RP_LIBS:
> -L/opt/intel/mkl/10.2.1.017/lib/em64t <http://10.2.1.017/lib/em64t> -lmkl_lapack
> -lmkl_intel
> > _lp64 -lmkl_scalapack_lp64 -lmkl_blacs_lp64
> -lmkl_sequential -lmkl_intel_ilp64 -
> > lmkl_scalapack_ilp64  -L/opt/fftw-2.1.5/lib/
> -lfftw_mpi -lfftw -L/opt/mpich2/lib
> >  -lmpich -lfmpich       
>          
>    linuxif9:MPIRUN:mpiexec _EXEC_
> > 
> > i am running on a xeon, CentOS5 machine.  intel
> noncommercial.
> > 
> > Any information would be appreciated.
> > 
> > Thank you,
> > JD
> > 
> > On Mon, Sep 14, 2009 at 10:31 AM, Jeff DeReus <jdereus at gmail.com
> <mailto:jdereus at gmail.com>>
> wrote:
> > 
> >     fftw was compiled into
> /opt/fftw-2.1.5 with flags
> > 
> >       $ ./configure
> --prefix=/opt/fftw-2.1.5/ --enable-mpi --enable-threads
> > 
> >     Thank you,
> >     JD
> > 
> > 
> >     On Mon, Sep 14, 2009 at 10:27
> AM, Laurence Marks
> >     <L-marks at northwestern.edu
> <mailto:L-marks at northwestern.edu>>
> wrote:
> > 
> >         It looks like
> you did not compile fftw (or it is somewhere else).
> > 
> >         2009/9/14 Jeff
> DeReus <jdereus at gmail.com
> >         <mailto:jdereus at gmail.com>>:
> >          > Hello
> again.  i am having some issues compiling lapw0/1/2
> >         modules in
> >          >
> wein2k_09.  i am running on a CentOS 5.3 box. 
> intel
> >         non-commercial
> >          > compilers and
> mkl.
> >          >
> >          >
> >          >
> >          > here are my
> current parallel settings from siteconfig.
> >          >
> >          >   
> Current settings:
> >          >   
>   RP  RP_LIB(SCALAPACK+PBLAS):
> >     
>    -L/opt/intel/mkl/10.2.1.017/lib/em64t
> <http://10.2.1.017/lib/em64t>
> >          > -lmkl_lapack
> -lmkl_intel_lp64 -lmkl_scalapack_lp64
> >     
>    -lmkl_blacs_lp64
> >          >
> -lmkl_sequential -L/opt/fftw-2.1.5/lib/ -lfftw_mpi -lfftw
> >     
>    -L/opt/mpich2/lib
> >          > -lmpich
> >          >   
>   FP  FPOPT(par.comp.options): $(FOPT) -FR -mp1
> -w
> >         -prec_div -pc80
> -pad
> >          > -align
> -DINTEL_VML -traceback -I/opt/mpich2/include
> >          >
> -I/opt/fftw-2.1.5/fortran
> >          >   
>   MP  MPIRUN commando        :
> mpiexec _EXEC_
> >          >
> >          > all other
> modules compile correctly.  if i do not enable
> parallel
> >          > functionality,
> the compilation completes with no errors.
> >          >
> >          > when compiling
> /opt/wien2k/SRC_lapw0 it ends with these
> >         errors which i
> have
> >          > not been able
> to track down.
> >          >
> >          > fftw_para.o: In
> function `exec_fftw_para_':
> >          >
> fftw_para.F:(.text+0x77): undefined reference to
> >     
>    `fftwnd_f77_mpi_'
> >          >
> fftw_para.F:(.text+0xb2): undefined reference to
> >     
>    `fftwnd_f77_mpi_'
> >          > fftw_para.o: In
> function `init_fftw_para_':
> >          >
> fftw_para.F:(.text+0x101): undefined reference to
> >          >
> `fftw3d_f77_mpi_create_plan_'
> >          >
> fftw_para.F:(.text+0x129): undefined reference to
> >          >
> `fftw3d_f77_mpi_create_plan_'
> >          >
> fftw_para.F:(.text+0x14d): undefined reference to
> >          >
> `fftwnd_f77_mpi_local_sizes_'
> >          > make[1]: ***
> [lapw0_mpi] Error 1
> >          > make[1]:
> Leaving directory `/opt/wien2k/SRC_lapw0'
> >          > make: ***
> [para] Error 2
> >          >
> >          > however,
> /opt/wien2k/SRC_lapw1/2 end with these errors.
> >          >
> >          > <snip>
> >          >
> /opt/intel/mkl/10.2.1.017/lib/em64t/libmkl_intel_lp64.so
> >         <http://10.2.1.017/lib/em64t/libmkl_intel_lp64.so>:
> undefined
> >          > reference to
> `mkl_pdett__d_forward_trig_transform'
> >          >
> /opt/intel/mkl/10.2.1.017/lib/em64t/libmkl_intel_lp64.so
> >         <http://10.2.1.017/lib/em64t/libmkl_intel_lp64.so>:
> undefined
> >          > reference to
> `mkl_trans_mkl_comatcopy'
> >          >
> /opt/intel/mkl/10.2.1.017/lib/em64t/libmkl_intel_lp64.so
> >         <http://10.2.1.017/lib/em64t/libmkl_intel_lp64.so>:
> undefined
> >          > reference to
> `mkl_pdett_s_init_trig_transform'
> >          >
> /opt/intel/mkl/10.2.1.017/lib/em64t/libmkl_intel_lp64.so
> >         <http://10.2.1.017/lib/em64t/libmkl_intel_lp64.so>:
> undefined
> >          > reference to
> `mkl_pdepl_s_commit_helmholtz_3d'
> >          >
> /opt/intel/mkl/10.2.1.017/lib/em64t/libmkl_intel_lp64.so
> >         <http://10.2.1.017/lib/em64t/libmkl_intel_lp64.so>:
> undefined
> >          > reference to
> `mkl_pdepl_d_commit_helmholtz_2d'
> >          >
> /opt/intel/mkl/10.2.1.017/lib/em64t/libmkl_intel_lp64.so
> >         <http://10.2.1.017/lib/em64t/libmkl_intel_lp64.so>:
> undefined
> >          > reference to
> `mkl_pdett__d_commit_trig_transform'
> >          >
> /opt/intel/mkl/10.2.1.017/lib/em64t/libmkl_intel_lp64.so
> >         <http://10.2.1.017/lib/em64t/libmkl_intel_lp64.so>:
> undefined
> >          > reference to
> `mkl_trans_mkl_comatadd'
> >          >
> /opt/intel/mkl/10.2.1.017/lib/em64t/libmkl_intel_lp64.so
> >         <http://10.2.1.017/lib/em64t/libmkl_intel_lp64.so>:
> undefined
> >          > reference to
> `mkl_serv_getversionstring_c'
> >          >
> /opt/intel/mkl/10.2.1.017/lib/em64t/libmkl_intel_lp64.so
> >         <http://10.2.1.017/lib/em64t/libmkl_intel_lp64.so>:
> undefined
> >          > reference to
> `mkl_spblas_lp64_mkl_dcsrsky'
> >          >
> /opt/intel/mkl/10.2.1.017/lib/em64t/libmkl_intel_lp64.so
> >         <http://10.2.1.017/lib/em64t/libmkl_intel_lp64.so>:
> undefined
> >          > reference to
> `mkl_pdett__d_init_trig_transform'
> >          > make[1]: ***
> [lapw2c_mpi] Error 1
> >          > make[1]:
> Leaving directory `/opt/wien2k/SRC_lapw2'
> >          > make: *** [cp]
> Error 2
> >          >
> >          >
> >          > i can find no
> mention of the fftw3d errors anywhere.
> >          >
> >          > my complete
> appropriate section of siteconfig
> >          >
> >          > # Linux PC
> system with IFC 10 compiler + mkl 10 (-ip is
> >         broken; -static
> does
> >          > not give
> traceback-lines)
> >          >
> linuxif9:FC:ifort
> >          >
> linuxif9:MPF:ifort
> >          > linuxif9:CC:cc
> >          >
> linuxif9:FOPT:-FR -mp1 -w -prec_div  -pc80 -pad -align
> >         -DINTEL_VML
> >          > -traceback -O3
> -xW
> >          >
> linuxif9:FPOPT:$(FOPT) -FR -mp1 -w -prec_div -pc80 -pad
> >         -align
> -DINTEL_VML
> >          > -traceback
> -I/opt/mpich2/include
> >          >
> linuxif9:LDFLAGS:  $(FOPT)
> >     
>    -L/opt/intel/mkl/10.2.1.017/lib/em64t
> >         <http://10.2.1.017/lib/em64t> -pthread
> >          > -i-static
> >          >
> linuxif9:R_LIBS:-L/opt/intel/mkl/10.2.1.017/lib/em64t
> >         <http://10.2.1.017/lib/em64t> -lmkl_lapack -lmkl
> >          > -liomp5 -lguide
> -lmkl_core
> >          >
> linuxif9:DPARALLEL:'-DParallel'
> >          >
> linuxif9:RP_LIBS: -L/opt/intel/mkl/10.2.1.017/lib/em64t
> >         <http://10.2.1.017/lib/em64t> -lmkl_lapack
> >          >
> -lmkl_intel_lp64 -lmkl_scalapack_lp64 -lmkl_blacs_lp64
> >     
>    -lmkl_sequential
> >          >
> -L/opt/fftw-2.1.5/lib/ -lfftw_mpi -lfftw
> >          >
> linuxif9:MPIRUN:mpiexec _EXEC_
> >          >
> >          >
> >          > any assistance
> in this matter would be appreciated.
> >          >
> >          > Thank you,
> >          > JD
> >          >
> >          >
> _______________________________________________
> >          > Wien mailing
> list
> >          > Wien at zeus.theochem.tuwien.ac.at
> >         <mailto:Wien at zeus.theochem.tuwien.ac.at>
> >          > http://zeus.theochem.tuwien.ac.at/mailman/listinfo/wien
> >          >
> >          >
> > 
> > 
> > 
> >         --
> >         Laurence Marks
> >         Department of
> Materials Science and Engineering
> >         MSE Rm 2036 Cook
> Hall
> >         2220 N Campus
> Drive
> >         Northwestern
> University
> >         Evanston, IL
> 60208, USA
> >         Tel: (847)
> 491-3996 Fax: (847) 491-7820
> >         email: L-marks
> at northwestern dot edu
> >         Web:
> www.numis.northwestern.edu <http://www.numis.northwestern.edu>
> >         Chair,
> Commission on Electron Crystallography of IUCR
> >     
>    www.numis.northwestern.edu/ <http://www.numis.northwestern.edu/>
> >         Electron
> crystallography is the branch of science that uses electron
> >         scattering and
> imaging to study the structure of matter.
> >     
>    _______________________________________________
> >         Wien mailing
> list
> >         Wien at zeus.theochem.tuwien.ac.at
> >         <mailto: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