[Pw_forum] Configure script problem

[Axel Kohlmeyer]

On Tue, Feb 14, 2012 at 11:11 AM, Federico Vadell
<federico at clustering.com.ar> wrote:
>  Im sorry, i was wrong. The mkl liraries were found, but it did not seem to
> find the LAPACK and FFTW libs. However I carried on with the compilation,
> ended without error, the only thing i noticed was the following line in the
> output of the configure script:
>
> checking for library containing dspev... none required

yes, because MKL contains *both* BLAS and LAPACK
in one command line. no reason to repeat it, right?

> I believe dspev is a lapack function, but since it said it was not required
> I moved on to the examples.
> I tried example01 in /path/to/espresso/examples/example01
> The out put was
>  [root at cuda_1 example01]# ./run_example

arrrrggggghhhh!!!!

you are compiling and running a user level
application as root??? that is baaaaadd!
nobody should do anything as root unless
there is no other way. far too easy to mess
up a machine and destroy lot of hard work.

> /opt/espresso-PRACE/examples/example01 : starting
>
> This example shows how to use pw.x to calculate the total energy and
> the band structure of four simple systems: Si, Al, Cu, Ni.
>
>  executables directory: /opt/espresso-PRACE/bin
>  pseudo directory: /opt/espresso-PRACE/pseudo
>  temporary directory: /root/tmp
>  checking that needed directories and files exist... done
>
>  running pw.x as: /opt/espresso-PRACE/bin/pw.x
>  running bands.x as: /opt/espresso-PRACE/bin/bands.x
>
>  cleaning /root/tmp... done
>  running the scf calculation for Si...2
> Error condition encountered during test: exit status = 2
> Aborting
>
> I checked the .out file of the example and found this error:
> %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
>  from cdiaghg : error # 21
>  diagonalization (ZHEGV*) failed
>  %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
>
>  stopping ...

there are selection of possible reasons,
miscompiled code  due to bad compilers
or overoptimization, a problem with the library,
memory corruption...

i would search the mailing list archive for similar
error messages and see what the outcome was.

axel.

>
> Thanks for your answer,
>
> Federico.-
>
> ----------------mensaje original-----------------
> De: "Axel Kohlmeyer" akohlmey at gmail.com
> Para: "Federico Vadell" federico at clustering.com.ar
> CC: pw_forum at pwscf.org
> Fecha: Tue, 14 Feb 2012 10:25:10 -0500
> -------------------------------------------------
>
>
>> On Tue, Feb 14, 2012 at 9:59 AM, Federico Vadell
>> federico at clustering.com.ar wrote:
>>> Dear all,
>>>
>>> I m having trouble with the configure script since it does not seem to
> find
>>> the mkl libs. Im using espresso-PRACE and intel composer XE 2011 sp1. Ive
>>> tried this line:
>>>
>>> ./configure  --enable-openmp --enable-cuda
>>> LIBDIRS="/opt/intel/composer_xe_2011_sp1.7.256/mkl/lib/intel64/"
>>> --prefix=/opt/espresso-PRACE
>>>
>>> where /opt/intel/composer_xe_2011_sp1.7.256/mkl/lib/intel64 has
>>>
>>> libaux_ia32_gfortran.a          libmkl_core.a
>>> libmkl_p4n.so
>>> libaux_ia32_gnu.a               libmkl_core.so
>>> libmkl_pgi_thread.a
>>> libaux_ia32_intel.a             libmkl_def.so
>>> libmkl_pgi_thread.so
>>> libaux_intel64_gfortran.a       libmkl_gf_ilp64.a
>>> libmkl_rt.so
>>> libaux_intel64_gnu.a            libmkl_gf_ilp64.so
>>> libmkl_scalapack_ilp64.a
>>> libaux_intel64_intel.a          libmkl_gf_lp64.a
>>> libmkl_scalapack_ilp64.so
>>> libmkl_avx.so                   libmkl_gf_lp64.so
>>> libmkl_scalapack_lp64.a
>>> libmkl_blacs_ilp64.a            libmkl_gnu_thread.a
>>> libmkl_scalapack_lp64.so
>>> libmkl_blacs_intelmpi_ilp64.a   libmkl_gnu_thread.so
>>> libmkl_sequential.a
>>> libmkl_blacs_intelmpi_ilp64.so  libmkl_intel_ilp64.a
>>> libmkl_sequential.so
>>> libmkl_blacs_intelmpi_lp64.a    libmkl_intel_ilp64.so
>>> libmkl_solver_ilp64.a
>>> libmkl_blacs_intelmpi_lp64.so   libmkl_intel_lp64.a
>>> libmkl_solver_ilp64_sequential.a
>>> libmkl_blacs_lp64.a             libmkl_intel_lp64.so
>>> libmkl_solver_lp64.a
>>> libmkl_blacs_openmpi_ilp64.a    libmkl_intel_sp2dp.a
>>> libmkl_solver_lp64_sequential.a
>>> libmkl_blacs_openmpi_lp64.a     libmkl_intel_sp2dp.so
>>> libmkl_vml_avx.so
>>> libmkl_blacs_sgimpt_ilp64.a     libmkl_intel_thread.a
>>> libmkl_vml_def.so
>>> libmkl_blacs_sgimpt_lp64.a      libmkl_intel_thread.so
>>> libmkl_vml_mc2.so
>>> libmkl_blas95_ilp64.a           libmkl_lapack95_ilp64.a
>>> libmkl_vml_mc3.so
>>> libmkl_blas95_lp64.a            libmkl_lapack95_lp64.a
>>> libmkl_vml_mc.so
>>> libmkl_cdft_core.a              libmkl_mc3.so
>>> libmkl_vml_p4n.so
>>> libmkl_cdft_core.so             libmkl_mc.so             locale
>>>
>>> These are the last lines of the configure:
>>>
>>> The following libraries have been found:
>>>   BLAS_LIBS=/opt/espresso-PRACE/phiGEMM/lib/libphigemm.a
>>> -lmkl_intel_lp64  -lmkl_intel_thread -lmkl_core
>>>   LAPACK_LIBS=
>>>   FFT_LIBS=
>>>
>>> If anyone can help i'd be gratefull
>>
>> why do you think there is a problem?
>> it looks to me as if MKL *was* found.
>>
>> axel.
>>
>>
>>>
>>> Federico.-
>>>
>>>
>>> _______________________________________________
>>> Pw_forum mailing list
>>> Pw_forum at pwscf.org
>>> http://www.democritos.it/mailman/listinfo/pw_forum
>>>
>>
>>
>>
>> --
>> Dr. Axel Kohlmeyer
>> akohlmey at gmail.com  http://goo.gl/1wk0
>>
>> College of Science and Technology
>> Temple University, Philadelphia PA, USA.
>>
>>
>
>
> _______________________________________________
> Pw_forum mailing list
> Pw_forum at pwscf.org
> http://www.democritos.it/mailman/listinfo/pw_forum



-- 
Dr. Axel Kohlmeyer
akohlmey at gmail.com  http://goo.gl/1wk0

College of Science and Technology
Temple University, Philadelphia PA, USA.