<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.01 Transitional//EN">
<html>
  <head>
    <meta content="text/html; charset=ISO-8859-1"
      http-equiv="Content-Type">
  </head>
  <body bgcolor="#ffffff" text="#000000">
    Now everything works! Thanks for such a quick fix.<br>
    <br>
    On 05.04.2012 15:53, Matthew Knepley wrote:
    <blockquote
cite="mid:CAMYG4Gk9qiUqgZOyk1cgyanG2rzWSe6W-Z4L3Tj2ZKr2Wdq2FQ@mail.gmail.com"
      type="cite">On Thu, Apr 5, 2012 at 8:05 AM, Barry Smith <span
        dir="ltr">&lt;<a moz-do-not-send="true"
          href="mailto:bsmith@mcs.anl.gov">bsmith@mcs.anl.gov</a>&gt;</span>
      wrote:<br>
      <div class="gmail_quote">
        <blockquote class="gmail_quote" style="margin: 0pt 0pt 0pt
          0.8ex; border-left: 1px solid rgb(204, 204, 204);
          padding-left: 1ex;">
          <br>
          &nbsp; Matt,<br>
          <br>
          &nbsp; &nbsp;It doesn't look like that link line is even linking against
          the MKL libraries?</blockquote>
        <div><br>
        </div>
        <div>Dang, this is a bug with specifying --with-blas-lapack-lib,
          since it all gets dumped in</div>
        <div>self.lapackLib, then self.blasLib is empty. Pushed a fix to
          BuildSystem. Testing now.</div>
        <div><br>
        </div>
        <div>&nbsp; &nbsp;Matt</div>
        <div>&nbsp;</div>
        <blockquote class="gmail_quote" style="margin: 0pt 0pt 0pt
          0.8ex; border-left: 1px solid rgb(204, 204, 204);
          padding-left: 1ex;"><span class="HOEnZb"><font color="#888888"><br>
              &nbsp; Barry<br>
            </font></span>
          <div class="HOEnZb">
            <div class="h5"><br>
              On Apr 5, 2012, at 7:54 AM, Alexander Grayver wrote:<br>
              <br>
              &gt; On 05.04.2012 14:42, Matthew Knepley wrote:<br>
              &gt;&gt; On Thu, Apr 5, 2012 at 7:22 AM, Alexander Grayver
              &lt;<a moz-do-not-send="true"
                href="mailto:agrayver@gfz-potsdam.de">agrayver@gfz-potsdam.de</a>&gt;
              wrote:<br>
              &gt;&gt; On 05.04.2012 14:13, Jed Brown wrote:<br>
              &gt;&gt;&gt; On Thu, Apr 5, 2012 at 04:54, Alexander
              Grayver &lt;<a moz-do-not-send="true"
                href="mailto:agrayver@gfz-potsdam.de">agrayver@gfz-potsdam.de</a>&gt;
              wrote:<br>
              &gt;&gt;&gt; The inconsistency is still there. I pulled
              petsc-dev today and came across with the same problem:<br>
              &gt;&gt;&gt;<br>
              &gt;&gt;&gt; src/vec/vec/impls/seq/bvec2.c<br>
              &gt;&gt;&gt; 239: &nbsp; &nbsp;*z =
              PetscRealPart(BLASdot_(&amp;bn,xx,&amp;one,xx,&amp;one));<br>
              &gt;&gt;&gt; 240: &nbsp; &nbsp;*z = PetscSqrtReal(*z);<br>
              &gt;&gt;&gt;<br>
              &gt;&gt;&gt; This doesn't work with Intel MKL. However in
              one of the mailing list discussions it was said there is
              now check for that.<br>
              &gt;&gt;&gt; This check didn't work for me apparently. Any
              ideas why?<br>
              &gt;&gt;&gt;<br>
              &gt;&gt;&gt; As always, we need configure.log.<br>
              &gt;&gt;&gt;<br>
              &gt;&gt; Please find it attached.<br>
              &gt;&gt;<br>
              &gt;&gt; Please ALWAYS ALWAYS ALWAYS send configure.log to
              <a moz-do-not-send="true"
                href="mailto:petsc-maint@mcs.anl.gov">petsc-maint@mcs.anl.gov</a>.
              They are too<br>
              &gt;&gt; big for the public list.<br>
              &gt;<br>
              &gt; Sorry, next time I will.<br>
              &gt;<br>
              &gt;&gt;<br>
              &gt;&gt; Second, your problem is that we get a link error
              tying to test this so it looks like<br>
              &gt;&gt; your BLAS have no complex implementations:<br>
              &gt;<br>
              &gt; It does have zdotc. Old petsc-dev works with it fine
              (after I patched vec implementaton myself to get around
              this issue).<br>
              &gt;<br>
              &gt; home&gt; nm
              /opt/intel/Compiler/11.1/072/mkl/lib/em64t/libmkl_intel_lp64.a
              | grep zdotc<br>
              &gt; cblas_zdotci_lp64.o:<br>
              &gt; 0000000000000000 T cblas_zdotci_sub<br>
              &gt; cblas_zdotc_lp64.o:<br>
              &gt; 0000000000000000 T cblas_zdotc_sub<br>
              &gt; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp;U zdotc<br>
              &gt; _zdotci_lp64.o:<br>
              &gt; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp;U mkl_blas_zdotci<br>
              &gt; 0000000000000000 T zdotci<br>
              &gt; 0000000000000000 T zdotci_<br>
              &gt; _zdotc_lp64.o:<br>
              &gt; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp;U mkl_blas_zdotc<br>
              &gt; 0000000000000000 T zdotc<br>
              &gt; 0000000000000000 T zdotc_<br>
              &gt; home&gt;<br>
              &gt;<br>
              &gt;&gt;<br>
              &gt;&gt; sh: /opt/mpi/intel/openmpi-1.4.2/bin/mpicc &nbsp;-o
              /tmp/petsc-8AEO1f/config.packages.BlasLapack/conftest &nbsp;
              &nbsp;-wd1572 -Qoption,cpp,--extended_float_type -g
              /tmp/petsc-8AEO1f/config.packages.BlasLapack/conftest.o
              &nbsp;-ldl -L/opt/mpi/intel/openmpi-1.4.2/lib -lmpi -lopen-rte
              -lopen-pal -lnsl -lutil -L/opt/mpi/intel/openmpi-1.4.2/lib
              -L/opt/intel/Compiler/11.1/072/ipp/em64t/lib
              -L/opt/intel/Compiler/11.1/072/mkl/interfaces/fftw3xf
              -L//opt/intel/Compiler/11.1/072/lib/intel64
              -L/usr/lib64/gcc/x86_64-suse-linux/4.3
              -L/opt/intel/Compiler/11.1/072/ipp/em64t/lib
              -L/opt/intel/Compiler/11.1/072/mkl/interfaces/fftw3xf
              -L/usr/x86_64-suse-linux/lib -limf -lsvml -lipgo -ldecimal
              -lgcc_s -lirc -lpthread -lirc_s
              -L/opt/mpi/intel/openmpi-1.4.2/lib
              -L/opt/intel/Compiler/11.1/072/ipp/em64t/lib
              -L/opt/intel/Compiler/11.1/072/mkl/interfaces/fftw3xf
              -L//opt/intel/Compiler/11.1/072/lib/intel64
              -L/usr/lib64/gcc/x86_64-suse-linux/4.3
              -L/opt/intel/Compiler/11.1/072/ipp/em64t/lib
              -L/opt/intel/Compiler/11.1/072/mkl/interfaces/fftw3xf
              -L/usr/x86_64-suse-linux/lib -ldl<br>
              &gt;&gt; Executing: /opt/mpi/intel/openmpi-1.4.2/bin/mpicc
              &nbsp;-o /tmp/petsc-8AEO1f/config.packages.BlasLapack/conftest
              &nbsp; &nbsp;-wd1572 -Qoption,cpp,--extended_float_type -g
              /tmp/petsc-8AEO1f/config.packages.BlasLapack/conftest.o
              &nbsp;-ldl -L/opt/mpi/intel/openmpi-1.4.2/lib -lmpi -lopen-rte
              -lopen-pal -lnsl -lutil -L/opt/mpi/intel/openmpi-1.4.2/lib
              -L/opt/intel/Compiler/11.1/072/ipp/em64t/lib
              -L/opt/intel/Compiler/11.1/072/mkl/interfaces/fftw3xf
              -L//opt/intel/Compiler/11.1/072/lib/intel64
              -L/usr/lib64/gcc/x86_64-suse-linux/4.3
              -L/opt/intel/Compiler/11.1/072/ipp/em64t/lib
              -L/opt/intel/Compiler/11.1/072/mkl/interfaces/fftw3xf
              -L/usr/x86_64-suse-linux/lib -limf -lsvml -lipgo -ldecimal
              -lgcc_s -lirc -lpthread -lirc_s
              -L/opt/mpi/intel/openmpi-1.4.2/lib
              -L/opt/intel/Compiler/11.1/072/ipp/em64t/lib
              -L/opt/intel/Compiler/11.1/072/mkl/interfaces/fftw3xf
              -L//opt/intel/Compiler/11.1/072/lib/intel64
              -L/usr/lib64/gcc/x86_64-suse-linux/4.3
              -L/opt/intel/Compiler/11.1/072/ipp/em64t/lib
              -L/opt/intel/Compiler/11.1/072/mkl/interfaces/fftw3xf
              -L/usr/x86_64-suse-linux/lib -ldl<br>
              &gt;&gt; sh:<br>
              &gt;&gt; Possible ERROR while running linker:
              /tmp/petsc-8AEO1f/config.packages.BlasLapack/conftest.o:
              In function `main':<br>
              &gt;&gt;
              /tmp/petsc-8AEO1f/config.packages.BlasLapack/conftest.c:14:
              undefined reference to `zdotc_'<br>
              &gt;&gt; &nbsp;output: ret = 256<br>
              &gt;&gt; error message =
              {/tmp/petsc-8AEO1f/config.packages.BlasLapack/conftest.o:
              In function `main':<br>
              &gt;&gt;
              /tmp/petsc-8AEO1f/config.packages.BlasLapack/conftest.c:14:
              undefined reference to `zdotc_'<br>
              &gt;&gt; }<br>
              &gt;&gt;<br>
              &gt;&gt;<br>
              &gt;&gt; &nbsp; &nbsp; Matt<br>
              &gt;&gt;<br>
              &gt;&gt; In include/petscblaslapack_uscore.h (and other
              places), you will find the following. So we need to know
              why PETSC_COMPLEX_DOT_RESULT_ARGUMENT was not found by
              configure.<br>
              &gt;&gt;<br>
              &gt;&gt; This I can see...<br>
              &gt;&gt;<br>
              &gt;&gt;<br>
              &gt;&gt;&gt;<br>
              &gt;&gt;&gt; #ifdef PETSC_COMPLEX_DOT_RESULT_ARGUMENT<br>
              &gt;&gt;&gt; EXTERN_C_BEGIN<br>
              &gt;&gt;&gt; extern void zdotc_(PetscScalar *,const
              PetscBLASInt*,const PetscScalar*,const PetscBLASInt*,const
              PetscScalar*,const PetscBLASInt*);<br>
              &gt;&gt;&gt; PETSC_STATIC_INLINE PetscScalar
              BLASdot_(const PetscBLASInt *n,const PetscScalar *x,const
              PetscBLASInt *sx,const PetscScalar *y,const PetscBLASInt
              *sy) {<br>
              &gt;&gt;&gt; &nbsp; PetscScalar tmpz;<br>
              &gt;&gt;&gt; &nbsp; zdotc_(&amp;tmpz,n,x,sx,y,sy);<br>
              &gt;&gt;&gt; &nbsp; return tmpz;<br>
              &gt;&gt;&gt; }<br>
              &gt;&gt;&gt; EXTERN_C_END<br>
              &gt;&gt;&gt; #else<br>
              &gt;&gt;&gt; # &nbsp;define BLASdot_ &nbsp; &nbsp; zdotc_<br>
              &gt;&gt;&gt; #endif<br>
              &gt;&gt;<br>
              &gt;&gt;<br>
              &gt;&gt;<br>
              &gt;&gt; --<br>
              &gt;&gt; Regards,<br>
              &gt;&gt; Alexander<br>
              &gt;&gt;<br>
              &gt;&gt;<br>
              &gt;&gt;<br>
              &gt;&gt;<br>
              &gt;&gt; --<br>
              &gt;&gt; What most experimenters take for granted before
              they begin their experiments is infinitely more
              interesting than any results to which their experiments
              lead.<br>
              &gt;&gt; -- Norbert Wiener<br>
              &gt;<br>
              &gt;<br>
              &gt; --<br>
              &gt; Regards,<br>
              &gt; Alexander<br>
              &gt;<br>
              <br>
            </div>
          </div>
        </blockquote>
      </div>
      <br>
      <br clear="all">
      <div><br>
      </div>
      -- <br>
      What most experimenters take for granted before they begin their
      experiments is infinitely more interesting than any results to
      which their experiments lead.<br>
      -- Norbert Wiener<br>
    </blockquote>
    <br>
    <br>
    <pre class="moz-signature" cols="72">-- 
Regards,
Alexander</pre>
  </body>
</html>