[mvapich-discuss] Mvapich building errors
Lorenzo Varriale
lorenzo.varriale at prismaspa.it
Wed Oct 17 09:04:39 EDT 2018
Hari,
Here’s the output from autopen.sh ;
[root at cn01 mvapich2-2.2rc1-gcc-485-cuda8]# ./autogen.sh
####################################
## Replicating confdb (and similar)
####################################
syncing 'src/mpl' --> 'src/pm/hydra/mpl'
syncing 'confdb' --> 'src/mpi/romio/confdb'
syncing 'confdb' --> 'src/mpl/confdb'
syncing 'confdb' --> 'src/pm/hydra/confdb'
syncing 'confdb' --> 'src/pm/hydra/mpl/confdb'
syncing 'confdb' --> 'test/mpi/confdb'
####################################
## Checking user environment
####################################
Verifying the location of autogen.sh... done
Checking if autotools are in the same location... yes, all in /usr
Checking for autoconf version... >= 2.67
Checking for automake version... bad automake installation
You either do not have automake in your path or it is too old (version
1.15 or higher required). You may be able to use
automake --version
Unfortunately, there is no standard format for the version output and
it changes between autotools versions. In addition, some versions of
autoconf choose among many versions and provide incorrect output).
[root at cn01 mvapich2-2.2rc1-gcc-485-cuda8]#
Yesterday we already stumbled in automake issues, hence we started some researches as per the yum available packages but, sadly, we still aren’t able to find a suitable version like the one suggested by autopen.
[root at cn01 mvapich2-2.2rc1-gcc-485-cuda8]# yum provides */automake
Plugin abilitati:fastestmirror, langpacks
Loading mirror speeds from cached hostfile
* base: ba.mirror.garr.it
* epel: epel.besthosting.ua
* extras: ba.mirror.garr.it
* updates: centos.mirror.ptisp.pt
automake-1.13.4-3.el7.noarch : A GNU tool for automatically creating Makefiles
Repo : base
Corrispondenza trovata in:
Nome file : /usr/bin/automake
1:bash-completion-extras-2.1-11.el7.noarch : Additional programmable completions for Bash
Repo : epel
Corrispondenza trovata in:
Nome file : /usr/share/bash-completion/completions/automake
automake-1.13.4-3.el7.noarch : A GNU tool for automatically creating Makefiles
Repo : @anaconda
Corrispondenza trovata in:
Nome file : /usr/bin/automake
[root at cn01 mvapich2-2.2rc1-gcc-485-cuda8]# yum search automake
Plugin abilitati:fastestmirror, langpacks
Loading mirror speeds from cached hostfile
* base: ba.mirror.garr.it
* epel: epel.besthosting.ua
* extras: ba.mirror.garr.it
* updates: centos.mirror.ptisp.pt
=================================================================== N/S matched: automake ====================================================================
automake.noarch : A GNU tool for automatically creating Makefiles
Ricerca effettuata solamente in nome e descrizione breve, usare "search all" per cercare in tutti i campi.
[root at cn01 mvapich2-2.2rc1-gcc-485-cuda8]# yum info automake
Plugin abilitati:fastestmirror, langpacks
Loading mirror speeds from cached hostfile
* base: ba.mirror.garr.it
* epel: epel.besthosting.ua
* extras: ba.mirror.garr.it
* updates: centos.mirror.ptisp.pt
Pacchetti installati
Nome : automake
Arch : noarch
Versione : 1.13.4
Rilascio : 3.el7
Dimensione : 1.7 M
Repo : installed
Dal repo : anaconda
Sommario : A GNU tool for automatically creating Makefiles
URL : http://www.gnu.org/software/automake/
Licenza : GPLv2+ and GFDL and Public Domain and MIT
Descrizione : Automake is a tool for automatically generating `Makefile.in'
: files compliant with the GNU Coding Standards.
:
: You should install Automake if you are developing software and would
: like to use its ability to automatically generate GNU standard
: Makefiles.
[root at cn01 mvapich2-2.2rc1-gcc-485-cuda8]#
How can we go further in investigation? Shall we download a newer version of automake from a different repo?
Best,
Lorenzo.
> Il giorno 17 ott 2018, alle ore 13:23, Subramoni, Hari <subramoni.1 at osu.edu> ha scritto:
>
> Hi, Lorenzo.
>
> Can you please try the following series of commands?
>
> ./autogen.sh
> ./configure …
> make
> make install
>
> These are warnings seen during the autogen step. We had fixed this with the 2.3GA release. Can you please let us know if you still see the errors after you do these steps?
>
> On a different note, if you are using the release tarball from our website, you need not do the autogen.sh (or the autoreconf -if) step. You can directly start from configure.
>
> Best Regards,
> Hari.
>
> From: mvapich-discuss-bounces at cse.ohio-state.edu On Behalf Of Lorenzo Varriale
> Sent: Wednesday, October 17, 2018 3:54 AM
> To: mvapich-discuss at cse.ohio-state.edu <mvapich-discuss at mailman.cse.ohio-state.edu>
> Subject: [mvapich-discuss] Mvapich building errors
>
> Greetings Team,
>
> We’re currently having issues in building several versions of mvapich on our cluster, the more we try, the less we achieve. Actually, we are at a loss, any suggestion is very appreciated.
>
> Following, some details about errors and environment we have on our cluster:
>
> MVAPICH2 V2.2rc1, 2.2, 2.3
>
> ##modules loaded
>
> gcc 4.8.5
> cuda 8.0
>
> ###command issues
>
> autoreconf -f -i
>
> ./configure --prefix=/home/share/opt/mvapich2-2.2-gcc-485-cuda8 F77=gfortran FC=gfortran CPP=cpp CC=gcc CXX=c++ LDFLAGS="-L/lib -L/home/share/nvidia/cuda-8.0/lib64" --enable-cuda=basic --with-cuda=/home/share/nvidia/cuda-8.0
>
> Make output error:
>
> [root at cn01 mvapich2-2.2rc1-gcc-485-cuda8]# make
> ( cd . && rm -f .err unusederr.txt ; rm -rf .tmp )
> ( cd . && \
> ./maint/extracterrmsgs -careful=unusederr.txt \
> -skip=src/util/multichannel/mpi.c \
> `cat maint/errmsgdirs` > .tmp 2>.err )
> ( cd . && if test -s .err ; then rm -f .tmp ; cat .err ; exit 1 ; fi )
> Warning: attempt to redefine **rdmacmmaxport. Duplicate ignored.
> Warning: attempt to redefine **rdmacmmaxport %d %d %d. Duplicate ignored.
> Warning: attempt to redefine **rdmacmminport. Duplicate ignored.
> Warning: attempt to redefine **rdmacmminport %d %d %d. Duplicate ignored.
> Warning: attempt to redefine **rdmacminvalidport. Duplicate ignored.
> Warning: attempt to redefine **rdmacminvalidport %d. Duplicate ignored.
> Warning: attempt to redefine **rdmacmportrange. Duplicate ignored.
> Warning: attempt to redefine **rdmacmportrange %d %d. Duplicate ignored.
> Warning: attempt to redefine **activeports. Duplicate ignored.
> Warning: attempt to redefine **activeports %d. Duplicate ignored.
> Warning: attempt to redefine **ibv_get_device_name. Duplicate ignored.
> Warning: attempt to redefine **onesidedcomps. Duplicate ignored.
> Warning: attempt to redefine **portquery. Duplicate ignored.
> Warning: attempt to redefine **portquery %d. Duplicate ignored.
> Warning: attempt to redefine **rdmacmmaxport. Duplicate ignored.
> Warning: attempt to redefine **rdmacmmaxport %d %d %d. Duplicate ignored.
> Warning: attempt to redefine **rdmacmminport. Duplicate ignored.
> Warning: attempt to redefine **rdmacmminport %d %d %d. Duplicate ignored.
> Warning: attempt to redefine **rdmacminvalidport. Duplicate ignored.
> Warning: attempt to redefine **rdmacminvalidport %d. Duplicate ignored.
> Warning: attempt to redefine **rdmacmportrange. Duplicate ignored.
> Warning: attempt to redefine **rdmacmportrange %d %d. Duplicate ignored.
> Warning: attempt to redefine **umadgetca. Duplicate ignored.
> Warning: attempt to redefine **umadgetrate. Duplicate ignored.
> Warning: attempt to redefine **umadinit. Duplicate ignored.
> make: *** [src/mpi/errhan/defmsg.h] Error 1
>
> Other system infos:
>
> [root at cn01 mvapich2-2.2rc1-gcc-485-cuda8]# uname -a
> Linux cn01 3.10.0-514.el7.x86_64 #1 SMP Tue Nov 22 16:42:41 UTC 2016 x86_64 x86_64 x86_64 GNU/Linux
>
> [root at cn01 mvapich2-2.2rc1-gcc-485-cuda8]# automake --version
> automake (GNU automake) 1.13.4
>
> [root at cn01 mvapich2-2.2rc1-gcc-485-cuda8]# rpm -qa | grep -i ofed
> mlnx-ofa_kernel-3.4-OFED.3.4.2.0.0.1.g30039f7.rhel7u3.x86_64
> knem-mlnx-1.1.2.90mlnx1-OFED.3.3.1.5.5.1.ga659703.rhel7u3.x86_64
> kmod-isert-1.7-OFED.3.4.1.0.6.1.g30039f7.rhel7u3.x86_64
> ofed-scripts-3.4-OFED.3.4.2.0.0.x86_64
> libibverbs-utils-1.2.1mlnx1-OFED.3.4.0.1.4.34200.x86_64
> libmlx5-devel-1.2.1mlnx1-OFED.3.4.1.0.0.34200.x86_64
> librdmacm-utils-1.1.0mlnx-OFED.3.4.0.0.4.34200.x86_64
> dapl-2.1.9mlnx-OFED.3.3.0.0.6.34100.x86_64
> kmod-mlnx-ofa_kernel-3.4-OFED.3.4.2.0.0.1.g30039f7.rhel7u3.x86_64
> kmod-knem-mlnx-1.1.2.90mlnx1-OFED.3.3.1.5.5.1.ga659703.rhel7u3.x86_64
> kmod-srp-1.6.1-OFED.3.4.1.0.6.1.g30039f7.rhel7u3.x86_64
> kmod-mlnx-sdp-3.4-OFED.3.4.1.0.6.1.g30039f7.rhel7u3.x86_64
> libibverbs-1.2.1mlnx1-OFED.3.4.0.1.4.34200.x86_64
> libibverbs-devel-static-1.2.1mlnx1-OFED.3.4.0.1.4.34200.x86_64
> libmlx4-1.2.1mlnx1-OFED.3.4.0.0.4.34200.x86_64
> libmlx5-1.2.1mlnx1-OFED.3.4.1.0.0.34200.x86_64
> libibcm-1.0.5mlnx2-OFED.3.4.0.0.4.34100.x86_64
> librdmacm-1.1.0mlnx-OFED.3.4.0.0.4.34200.x86_64
> librdmacm-devel-1.1.0mlnx-OFED.3.4.0.0.4.34200.x86_64
> libsdp-devel-1.1.108-OFED.3.0.8.gfbd01df.34100.x86_64
> dapl-devel-2.1.9mlnx-OFED.3.3.0.0.6.34100.x86_64
> dapl-utils-2.1.9mlnx-OFED.3.3.0.0.6.34100.x86_64
> mlnxofed-docs-3.4-2.0.0.0.noarch
> mlnx-ofa_kernel-devel-3.4-OFED.3.4.2.0.0.1.g30039f7.rhel7u3.x86_64
> kmod-iser-1.8.1-OFED.3.4.1.0.6.1.g30039f7.rhel7u3.x86_64
> kmod-mlnx-rds-3.4-OFED.3.4.1.0.6.1.g30039f7.rhel7u3.x86_64
> libibverbs-devel-1.2.1mlnx1-OFED.3.4.0.1.4.34200.x86_64
> libmlx4-devel-1.2.1mlnx1-OFED.3.4.0.0.4.34200.x86_64
> libibcm-devel-1.0.5mlnx2-OFED.3.4.0.0.4.34100.x86_64
> ibacm-1.2.1mlnx1-OFED.3.4.0.1.5.34100.x86_64
> libsdp-1.1.108-OFED.3.0.8.gfbd01df.34100.x86_64
> dapl-devel-static-2.1.9mlnx-OFED.3.3.0.0.6.34100.x86_64
>
>
> Thanks in advance,
>
> Lorenzo Varriale.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.cse.ohio-state.edu/pipermail/mvapich-discuss/attachments/20181017/6c47eea9/attachment-0001.html>
More information about the mvapich-discuss
mailing list