[R-sig-hpc] Problems Installing Rmpi on openSUSE 11.1 x86_64
Martin Morgan
mtmorgan at fhcrc.org
Thu Jul 9 16:04:29 CEST 2009
Melwin <win at comets.de> writes:
> Thanks to Hao, Brian and Martin (offlist) for their suggestions.
>
> I tried them all:
> - setting LD_LIBRARY_PATH
> - using "--configure-args=--with-mpi=/path_to_lam"
> - using mpicc
> -> all without success when trying to install Rmpi (same error message
> as before)
>
> Since I cannot accept using OpenMPI using 100% CPU power in idle mode, I
> still need to get LAM working.
> Martin suggested modifiying "configure.ac":
>
> tar xzf Rmpi_0.5-7.tar.gz
> vim Rmpi/configure.ac
>
> I added
> after line 67: MPITYPE="LAM"
> after line 128:
> elif test -f ${MPI_ROOT}/lib64/libmpi.so; then #Till
> echo "Found libmpi in ${MPI_ROOT}/lib64"
> MPI_LIBS="-L${MPI_ROOT}/lib64 -lmpi"
> MPI_INCLUDE="${MPI_INCLUDE} ${MPI_INCLUDE}/64"
>
> followed by
> cd Rmpi && autoconf && cd ..
> R CMD INSTALL --clean Rmpi
>
> compiles without errors (see snippet 1 below).
> However, when tryinig to load Rmpi in R I get an error:
>
>> library("Rmpi")
> Error in dyn.load(file, DLLpath = DLLpath, ...) :
> kann shared library
> /home/francke/R/x86_64-unknown-linux-gnu-library/2.8/Rmpi/libs/Rmpi.so'
> nicht laden:
> /usr/lib64/libmpi.so.0: undefined symbol: lam_ssi_cr_did
> Fehler in library("Rmpi") : .First.lib fehlgeschlagen für 'Rmpi'
> Error in dyn.unload(file.path(libpath, "libs", paste("Rmpi",
> .Platform$dynlib.ext, :
> shared library
> /home/francke/R/x86_64-unknown-linux-gnu-library/2.8/Rmpi/libs/Rmpi.so'
> wurde nicht geladen
My guess is that you have successfully told R how to find LAM during
compile, but that your system needs to be configured so that R can
find LAM when it is running. I guess that
R CMD ldd /home/francke/R/x86_64-unknown-linux-gnu-library/2.8/Rmpi/libs/Rmpi.so
lists something like 'libmpi.so.0 => not found' (I'm just guessing at
the missing library name; there may be several). What you need to do
is to tell R where to look, e.g., temporarily with
LD_LIBRARY_PATH=/path/to/lib64 R
and more permanently by ldconfig or the like.
Martin
>
> Translation:
>> library("Rmpi")
> Error in dyn.load(file, DLLpath = DLLpath, ...) :
> could not load shared library
> /home/francke/R/x86_64-unknown-linux-gnu-library/2.8/Rmpi/libs/Rmpi.so':
> /usr/lib64/libmpi.so.0: undefined symbol: lam_ssi_cr_did
> Error in library("Rmpi") : .First.lib failed for 'Rmpi'
> Error in dyn.unload(file.path(libpath, "libs", paste("Rmpi",
> .Platform$dynlib.ext, :
> shared library
> /home/francke/R/x86_64-unknown-linux-gnu-library/2.8/Rmpi/libs/Rmpi.so'
> not loaded
>>
>
> "Rmpi.so" exists, so I dunno what's wrong. Any help is appreciated,
>
> Melwin
>
>
>
> ####snippet: installation of RMPI
> * Installing to library
> '/home/francke/R/x86_64-unknown-linux-gnu-library/2.8'
> * Installing *source* package 'Rmpi' ...
> checking for gcc... gcc
> checking for C compiler default output file name... a.out
> checking whether the C compiler works... yes
> checking whether we are cross compiling... no
> checking for suffix of executables...
> checking for suffix of object files... o
> checking whether we are using the GNU C compiler... yes
> checking whether gcc accepts -g... yes
> checking for gcc option to accept ISO C89... none needed
> I am here /usr
> Try to find mpi.h ...
> Found in /usr/include
> Try to find libmpi.so or libmpich.a
> Found libmpi in /usr/lib64
> Try to find liblam.so ...
> checking for main in -llam... no
> liblam not found. Probably not LAM-MPI
> checking for openpty in -lutil... yes
> checking for main in -lpthread... yes
> configure: creating ./config.status
> config.status: creating src/Makevars
> ** libs
> gcc -std=gnu99 -I/usr/lib64/R/include -DPACKAGE_NAME=\"\"
> -DPACKAGE_TARNAME=\"\" -DPACKAGE_VERSION=\"\" -DPACKAGE_STRING=\"\"
> -DPACKAGE_BUGREPORT=\"\" -I/usr/include -I/usr/include/64 -DMPI2 -DLAM
> -fPIC -I/usr/local/include -fpic -fmessage-length=0 -O2 -Wall
> -D_FORTIFY_SOURCE=2 -fstack-protector -funwind-tables
> -fasynchronous-unwind-tables -c conversion.c -o conversion.o
> gcc -std=gnu99 -I/usr/lib64/R/include -DPACKAGE_NAME=\"\"
> -DPACKAGE_TARNAME=\"\" -DPACKAGE_VERSION=\"\" -DPACKAGE_STRING=\"\"
> -DPACKAGE_BUGREPORT=\"\" -I/usr/include -I/usr/include/64 -DMPI2 -DLAM
> -fPIC -I/usr/local/include -fpic -fmessage-length=0 -O2 -Wall
> -D_FORTIFY_SOURCE=2 -fstack-protector -funwind-tables
> -fasynchronous-unwind-tables -c internal.c -o internal.o
> gcc -std=gnu99 -I/usr/lib64/R/include -DPACKAGE_NAME=\"\"
> -DPACKAGE_TARNAME=\"\" -DPACKAGE_VERSION=\"\" -DPACKAGE_STRING=\"\"
> -DPACKAGE_BUGREPORT=\"\" -I/usr/include -I/usr/include/64 -DMPI2 -DLAM
> -fPIC -I/usr/local/include -fpic -fmessage-length=0 -O2 -Wall
> -D_FORTIFY_SOURCE=2 -fstack-protector -funwind-tables
> -fasynchronous-unwind-tables -c RegQuery.c -o RegQuery.o
> gcc -std=gnu99 -I/usr/lib64/R/include -DPACKAGE_NAME=\"\"
> -DPACKAGE_TARNAME=\"\" -DPACKAGE_VERSION=\"\" -DPACKAGE_STRING=\"\"
> -DPACKAGE_BUGREPORT=\"\" -I/usr/include -I/usr/include/64 -DMPI2 -DLAM
> -fPIC -I/usr/local/include -fpic -fmessage-length=0 -O2 -Wall
> -D_FORTIFY_SOURCE=2 -fstack-protector -funwind-tables
> -fasynchronous-unwind-tables -c Rmpi.c -o Rmpi.o
> gcc -std=gnu99 -shared -L/usr/local/lib64 -o Rmpi.so conversion.o
> internal.o RegQuery.o Rmpi.o -L/usr/lib64 -lmpi -lutil -lpthread -fPIC
> -L/usr/lib64/R/lib -lR
> ** R
> ** demo
> ** inst
> ** preparing package for lazy loading
> ** help
> >>> Building/Updating help pages for package 'Rmpi'
> [...]
> ** building package indices ...
> * DONE (Rmpi)
> ####end snippet
>
>
>
>
>
>
> 2009/6/26 Hao Yu <hyu at stats.uwo.ca>:
>> Hi Melwin,
>>
>> Did you try
>>
>> R CMD INSTALL Rmpi_0.5-7.tar.gz --configure-args=--with-mpi=/path_to_lam
>>
>> Your system may have several mpi and Rmpi did not find a right mpi.
>>
>> BTW, when you are installing lam, make sure to turn dynamic lib on. The
>> default one is static one and Rmpi needs libmpi.so to work.
>>
>> Hao
>>
>> Melwin wrote:
>>> Thanks for the reply.
>>>
>>> 2009/6/24 Rowe, Brian Lee Yung (Portfolio Analytics) <B_Rowe at ml.com>:
>>>> Did you confirm that your MPI installation is working? That is the
>>>> probably the first step.
>>> LAM seems to work ok:
>>> johnson at linuxmodel:~> lamboot -v
>>>
>>> LAM 7.1.4/MPI 2 C++/ROMIO - Indiana University
>>>
>>> n-1<4165> ssi:boot:base:linear: booting n0 (localhost)
>>> n-1<4165> ssi:boot:base:linear: finished
>>> johnson at linuxmodel:~> lamexec C hostname
>>> linuxmodel
>>>
>>>
>>>> Regarding LD_LIBRARY_PATH, since I don't know what your configuration
>>>> is, make sure you export it to your environment and also don't
>>>> overwrite
>>>> any existing values.
>>>> i.e. This is better:
>>>> $ export LD_LIBRARY_PATH=/usr/lib64:$LD_LIBRARY_PATH
>>>> since it looks like you're on a 64 bit platform.
>>>
>>> Tried that. Same results as before:
>>> johnson at linuxmodel:~> R CMD INSTALL Rmpi_0.5-7.tar.gz
>>> * Installing to library
>>> '/home/johnson/R/x86_64-unknown-linux-gnu-library/2.8'
>>> * Installing *source* package 'Rmpi' ...
>>> checking for gcc... gcc
>>> checking for C compiler default output file name... a.out
>>> checking whether the C compiler works... yes
>>> checking whether we are cross compiling... no
>>> checking for suffix of executables...
>>> checking for suffix of object files... o
>>> checking whether we are using the GNU C compiler... yes
>>> checking whether gcc accepts -g... yes
>>> checking for gcc option to accept ISO C89... none needed
>>> I am here /usr
>>> Try to find mpi.h ...
>>> Found in /usr/include
>>> Try to find libmpi.so or libmpich.a
>>> checking for main in -lmpi... no
>>> libmpi not found. exiting...
>>> ERROR: configuration failed for package 'Rmpi'
>>> ** Removing
>>> /home/johnson/R/x86_64-unknown-linux-gnu-library/2.8/Rmpi'
>>>
>>>> I use OpenMPI and it include an mpi-based compiler which I found
>>>> to work
>>>> better than gcc directly. You might try that. Just make sure you export
>>>> your MPI_ROOT.
>>> I would love to use OpenMPI, but I gather the slave spawning from
>>> within R
>>> is not yet possible - which is what I need in interactive mode
>>> (https://stat.ethz.ch/pipermail/r-sig-hpc/2009-February/000104.html).
>>> I'd
>>> love to hear that this has changed.
>>>
>>>> It would also probably help to include a bit more output
>>>> from the compilation process.
>>> The dump above contains the entire output, I hope that helps.
>>>
>>> thanks again,
>>> Melwin
>>>
>>> _______________________________________________
>>> R-sig-hpc mailing list
>>> R-sig-hpc at r-project.org
>>> https://stat.ethz.ch/mailman/listinfo/r-sig-hpc
>>>
>>
>>
>> --
>> Department of Statistics & Actuarial Sciences
>> Fax Phone#:(519)-661-3813
>> The University of Western Ontario
>> Office Phone#:(519)-661-3622
>> London, Ontario N6A 5B7
>> http://www.stats.uwo.ca/faculty/yu
>
> _______________________________________________
> R-sig-hpc mailing list
> R-sig-hpc at r-project.org
> https://stat.ethz.ch/mailman/listinfo/r-sig-hpc
--
Martin Morgan
Computational Biology / Fred Hutchinson Cancer Research Center
1100 Fairview Ave. N.
PO Box 19024 Seattle, WA 98109
Location: Arnold Building M1 B861
Phone: (206) 667-2793
More information about the R-sig-hpc
mailing list