Hi Tim, the configure script of the vt-integration don't check whether MPI I/O is disabled, so the VT's MPI_File_* wrapper functions will be built. A quick solution for this problem should be to add the option "--with-contrib-vt-flags=--without-mpi-io" to the configure script. I'm working on a better solution...
To disable the vt-integration completely add the following option to the
configure script:
configure --enable-contrib-no-build=vt ...
Sorry again for this trouble!
Regards,
Matthias
On Di, 2008-01-29 at 07:37 -0500, Tim Prins wrote:
> We have run into another issue with the vt integration. Configuring with
> '--disable-mpi-io' results in:
>
> vt_mpiwrap.c:3565: error: `fh' undeclared (first use in this function)
> vt_mpiwrap.c:3565: error: `buf' undeclared (first use in this function)
> vt_mpiwrap.c:3565: error: `count' undeclared (first use in this function)
> vt_mpiwrap.c:3565: error: `datatype' undeclared (first use in this function)
> vt_mpiwrap.c:3565: error: `status' undeclared (first use in this function)
> vt_mpiwrap.c: At top level:
> vt_mpiwrap.c:3582: error: syntax error before "fh"
> vt_mpiwrap.c: In function `MPI_File_seek_shared':
> vt_mpiwrap.c:3596: error: `fh' undeclared (first use in this function)
> vt_mpiwrap.c:3596: error: `offset' undeclared (first use in this function)
> vt_mpiwrap.c:3596: error: `whence' undeclared (first use in this function)
> vt_mpiwrap.c: At top level:
> vt_mpiwrap.c:3613: error: syntax error before "fh"
> vt_mpiwrap.c: In function `MPI_File_write_ordered':
> vt_mpiwrap.c:3629: error: `fh' undeclared (first use in this function)
> vt_mpiwrap.c:3629: error: `buf' undeclared (first use in this function)
> vt_mpiwrap.c:3629: error: `count' undeclared (first use in this function)
> vt_mpiwrap.c:3629: error: `datatype' undeclared (first use in this function)
> vt_mpiwrap.c:3629: error: `status' undeclared (first use in this function)
> vt_mpiwrap.c: At top level:
> vt_mpiwrap.c:3646: error: syntax error before "fh"
> vt_mpiwrap.c: In function `MPI_File_write_shared':
> vt_mpiwrap.c:3662: error: `fh' undeclared (first use in this function)
> vt_mpiwrap.c:3662: error: `buf' undeclared (first use in this function)
> vt_mpiwrap.c:3662: error: `count' undeclared (first use in this function)
> vt_mpiwrap.c:3662: error: `datatype' undeclared (first use in this function)
> vt_mpiwrap.c:3662: error: `status' undeclared (first use in this function)
> rm -f libvt.a
> mv -f .deps/libvt_mpi_a-vt_user_region.Tpo
> .deps/libvt_mpi_a-vt_user_region.Po
> ar cru libvt.a vt_comp_gnu.o vt_memhook.o vt_memreg.o vt_iowrap.o
> vt_iowrap_helper.o
> rfg_regions.o rfg_filter.o rfg_groups.o vt_env.o vt_error.o vt_metmap.o
> vt_otf_gen.o vt_otf_sum.o
> vt_otf_trc.o vt_pform_linux.o vt_pomp_base.o vt_strdup.o vt_thrd.o
> vt_user_comment.o
> vt_user_count.o vt_user_region.o pomp_fwrapper_base.o
> mv -f .deps/libvt_mpi_a-vt_mpireq.Tpo .deps/libvt_mpi_a-vt_mpireq.Po
> vt_fmpiwrap.c: In function `vt_mpi_file_close_f':
> vt_fmpiwrap.c:2080: error: `MPI_File' undeclared (first use in this
> function)
> vt_fmpiwrap.c:2080: error: (Each undeclared identifier is reported only once
> vt_fmpiwrap.c:2080: error: for each function it appears in.)
> vt_fmpiwrap.c:2080: error: syntax error before "l_fh"
> vt_fmpiwrap.c:2081: error: `l_fh' undeclared (first use in this function)
> vt_fmpiwrap.c: In function `vt_mpi_file_open_f':
> vt_fmpiwrap.c:2095: error: `MPI_File' undeclared (first use in this
> function)
> vt_fmpiwrap.c:2095: error: syntax error before "l_fh"
> vt_fmpiwrap.c:2103: error: `l_fh' undeclared (first use in this function)
> make[5]: *** [libvt_mpi_a-vt_mpiwrap.o] Error 1
> make[5]: *** Waiting for unfinished jobs....
> ranlib libvt.a
> make[5]: *** Waiting for unfinished jobs....
> make[5]: *** [libvt_fmpi_a-vt_fmpiwrap.o] Error 1
> mv -f .deps/libvt_mpi_a-vt_iowrap.Tpo .deps/libvt_mpi_a-vt_iowrap.Po
> make[5]: Leaving directory
> `/san/homedirs/mpiteam/mtt-runs/odin/20080128-Nightly/pb_5/installs/VUuZ/src/openmpi-1.3a1r17289/ompi/contrib/vt/vt/vtlib'
> make[4]: *** [all-recursive] Error 1
> make[4]: Leaving directory
> `/san/homedirs/mpiteam/mtt-runs/odin/20080128-Nightly/pb_5/installs/VUuZ/src/openmpi-1.3a1r17289/ompi/contrib/vt/vt'
> make[3]: *** [all] Error 2
> make[3]: Leaving directory
> `/san/homedirs/mpiteam/mtt-runs/odin/20080128-Nightly/pb_5/installs/VUuZ/src/openmpi-1.3a1r17289/ompi/contrib/vt/vt'
> make[2]: *** [all-recursive] Error 1
> make[2]: Leaving directory
> `/san/homedirs/mpiteam/mtt-runs/odin/20080128-Nightly/pb_5/installs/VUuZ/src/openmpi-1.3a1r17289/ompi/contrib/vt'
> make[1]: *** [all-recursive] Error 1
> make[1]: Leaving directory
> `/san/homedirs/mpiteam/mtt-runs/odin/20080128-Nightly/pb_5/installs/VUuZ/src/openmpi-1.3a1r17289/ompi'
> make: *** [all-recursive] Error 1
>
>
> Sorry for the messy output, we use make -j 6.
>
> Thanks,
>
> Tim
>
> Andreas Knüpfer wrote:
> > On Tuesday 29 January 2008, George Bosilca wrote:
> >> Look like VT do not correctly compute dependencies. A static build
> >> will fails if libz.a is not installed on the system.
> >>
> >> /usr/bin/ld: cannot find -lz
> >> collect2: ld returned 1 exit status
> >> make[5]: *** [vtfilter] Error 1
> >>
> >> george.
> >>
> >
> > this issue should be fixed now, even without additional configure
> > parameters.
> >
> >
> >
> >
> > ------------------------------------------------------------------------
> >
> > _______________________________________________
> > devel mailing list
> > de...@open-mpi.org
> > http://www.open-mpi.org/mailman/listinfo.cgi/devel
>
> _______________________________________________
> devel mailing list
> de...@open-mpi.org
> http://www.open-mpi.org/mailman/listinfo.cgi/devel
--
Matthias Jurenz,
Center for Information Services and
High Performance Computing (ZIH), TU Dresden,
Willersbau A106, Zellescher Weg 12, 01062 Dresden
phone +49-351-463-31945, fax +49-351-463-37773
smime.p7s
Description: S/MIME cryptographic signature
