Re: [R] Rmpi installation trouble
Note: libmpi.dylib: cannot open shared object file: No such file or directory Linux does not have .dylib's: Darwin (part of Mac OS X) does. This is not an R question (it seems to be about your MPI installation) and you need to discuss it with the package maintainer (see the posting quide). OpenMPI was found in /usr/local so it seems not to be a system installation. So you'll need to give full details about that. On 10/10/2012 05:46, Erin Hodgess wrote: Dear R People: I'm trying to install Rmpi (latest version) on Ubuntu. However, I'm getting the following error: erin@erin-Y480:~$ R R version 2.15.1 (2012-06-22) -- Roasted Marshmallows Copyright (C) 2012 The R Foundation for Statistical Computing ISBN 3-900051-07-0 Platform: x86_64-pc-linux-gnu (64-bit) R is free software and comes with ABSOLUTELY NO WARRANTY. You are welcome to redistribute it under certain conditions. Type 'license()' or 'licence()' for distribution details. Natural language support but running in an English locale R is a collaborative project with many contributors. Type 'contributors()' for more information and 'citation()' on how to cite R or R packages in publications. Type 'demo()' for some demos, 'help()' for on-line help, or 'help.start()' for an HTML browser interface to help. Type 'q()' to quit R. install.packages(Rmpi,depen=TRUE) Installing package(s) into ‘/home/erin/R/x86_64-pc-linux-gnu-library/2.15’ (as ‘lib’ is unspecified) --- Please select a CRAN mirror for use in this session --- Loading Tcl/Tk interface ... done trying URL 'http://cran.at.r-project.org/src/contrib/Rmpi_0.6-1.tar.gz' Content type 'application/x-gzip' length 92977 bytes (90 Kb) opened URL == downloaded 90 Kb * installing *source* package ‘Rmpi’ ... ** package ‘Rmpi’ successfully unpacked and MD5 sums checked checking for gcc... gcc -std=gnu99 checking whether the C compiler works... yes checking for C compiler default output file name... a.out checking for suffix of executables... checking whether we are cross compiling... no checking for suffix of object files... o checking whether we are using the GNU C compiler... yes checking whether gcc -std=gnu99 accepts -g... yes checking for gcc -std=gnu99 option to accept ISO C89... none needed I am here /usr/local and it is OpenMPI Trying to find mpi.h ... Found in /usr/local/include Trying to find libmpi.so or libmpich.a ... Found libmpi in /usr/local/lib 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/share/R/include -DNDEBUG -DPACKAGE_NAME=\\ -DPACKAGE_TARNAME=\\ -DPACKAGE_VERSION=\\ -DPACKAGE_STRING=\\ -DPACKAGE_BUGREPORT=\\ -DPACKAGE_URL=\\ -I/usr/local/include -DMPI2 -DOPENMPI -fpic -O3 -pipe -g -c RegQuery.c -o RegQuery.o gcc -std=gnu99 -I/usr/share/R/include -DNDEBUG -DPACKAGE_NAME=\\ -DPACKAGE_TARNAME=\\ -DPACKAGE_VERSION=\\ -DPACKAGE_STRING=\\ -DPACKAGE_BUGREPORT=\\ -DPACKAGE_URL=\\ -I/usr/local/include -DMPI2 -DOPENMPI -fpic -O3 -pipe -g -c Rmpi.c -o Rmpi.o gcc -std=gnu99 -I/usr/share/R/include -DNDEBUG -DPACKAGE_NAME=\\ -DPACKAGE_TARNAME=\\ -DPACKAGE_VERSION=\\ -DPACKAGE_STRING=\\ -DPACKAGE_BUGREPORT=\\ -DPACKAGE_URL=\\ -I/usr/local/include -DMPI2 -DOPENMPI -fpic -O3 -pipe -g -c conversion.c -o conversion.o gcc -std=gnu99 -I/usr/share/R/include -DNDEBUG -DPACKAGE_NAME=\\ -DPACKAGE_TARNAME=\\ -DPACKAGE_VERSION=\\ -DPACKAGE_STRING=\\ -DPACKAGE_BUGREPORT=\\ -DPACKAGE_URL=\\ -I/usr/local/include -DMPI2 -DOPENMPI -fpic -O3 -pipe -g -c internal.c -o internal.o gcc -std=gnu99 -shared -o Rmpi.so RegQuery.o Rmpi.o conversion.o internal.o -L/usr/local/lib -lmpi -lutil -lpthread -L/usr/lib/R/lib -lR installing to /home/erin/R/x86_64-pc-linux-gnu-library/2.15/Rmpi/libs ** R ** demo ** inst ** preparing package for lazy loading ** help ** help *** installing help indices ** building package indices ** testing if installed package can be loaded libmpi.dylib: cannot open shared object file: No such file or directory Error : .onLoad failed in loadNamespace() for 'Rmpi', details: call: fun(libname, pkgname) error: Cannot start MPI_Init(). Exit Error: loading failed Execution halted ERROR: loading failed * removing ‘/home/erin/R/x86_64-pc-linux-gnu-library/2.15/Rmpi’ The downloaded source packages are in ‘/tmp/RtmpWCdvrW/downloaded_packages’ Warning message: In install.packages(Rmpi, depen = TRUE) : installation of package ‘Rmpi’ had non-zero exit status Any suggestions would be much appreciated. Sincerely, Erin -- Brian D. Ripley, rip...@stats.ox.ac.uk Professor of Applied Statistics, http://www.stats.ox.ac.uk/~ripley/ University of Oxford, Tel: +44 1865 272861 (self) 1 South Parks Road, +44 1865 272866 (PA) Oxford OX1 3TG, UKFax: +44 1865 272595
[R] Rmpi installation problems
Hi all, I am trying to install the Rmpi package in R and, while the installation itself works, it breaks down when trying to load the library. I think it has something to do with shared vs static loading of helper libraries, or the order in which shared libraries are loaded, but I am not sure. R version: 2.14.0 Linux version: Gentoo, i686-pc-linux-gnu (32-bit) GCC version: 4.5.3 (Gentoo 4.5.3-r1 p1.0) OpenMPI version: 1.5.4 Output from R CMD INSTALL . in Rmpi source directory: * installing to library ‘/home/jos/R/i686-pc-linux-gnu-library/2.14’ * installing *source* package ‘Rmpi’ ... checking for gcc... i686-pc-linux-gnu-gcc -std=gnu99 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 i686-pc-linux-gnu-gcc -std=gnu99 accepts -g... yes checking for i686-pc-linux-gnu-gcc -std=gnu99 option to accept ISO C89... none needed I am here /usr and it is OpenMPI Trying to find mpi.h ... Found in /usr/include Trying to find libmpi.so or libmpich.a ... Found libmpi in /usr/lib checking for openpty in -lutil... yes checking for main in -lpthread... yes configure: creating ./config.status config.status: creating src/Makevars ** Creating default NAMESPACE file ** libs make: Nothing to be done for `all'. installing to /home/jos/R/i686-pc-linux-gnu-library/2.14/Rmpi/libs ** R ** demo ** inst ** preparing package for lazy loading ** help *** installing help indices ** building package indices ... ** testing if installed package can be loaded /usr/lib/R/bin/exec/R: symbol lookup error: /usr/lib/openmpi/mca_paffinity_hwloc.so: undefined symbol: mca_base_param_reg_int ERROR: loading failed * removing ‘/home/jos/R/i686-pc-linux-gnu-library/2.14/Rmpi’ Any help would be greatly appreciated! Jos __ R-help@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.
Re: [R] Rmpi installation
Another update: I ditched the pre-installed version of openmpi and installed openmpi from source (./configure --without-openib). Compiling and running your test programme gives me the following error: unger@zytosrv01dmi:~/R_projects/OS ./mtest [zytosrv01dmi:6094] *** An error occurred in MPI_Send [zytosrv01dmi:6094] *** on communicator MPI_COMM_WORLD [zytosrv01dmi:6094] *** MPI_ERR_RANK: invalid rank [zytosrv01dmi:6094] *** MPI_ERRORS_ARE_FATAL (your MPI job will now abort) However, Rmpi installs happily and it is also possible to call the package in R. When it comes to running a job under snowfall sfInit( parallel=TRUE,type=MPI, cpus=8) the job is split into eight separate processes which all run on the same CPU. So it seems that the processes are not distributed across the 8 CPUs available. Best wishes Kristian Am 15.06.11 19:25 schrieb Hugo Mildenberger unter hugo.mildenber...@web.de: Dear Kristian, please run exactly mpicc mtest.c -o mtest If you really need it, add -v separately. mpicc is nothing but a compiler wrapper. The -o switch specifies the outfile name, which has to follow immediately after -o, with or without a blank character in between. I'm not sure about what happens with -ov mtest. Probably an output file named v is produced on disk, in addition to the file mtest stemming from a previous with -o mtest Depending on mpicc's, gcc's and ld's version, gcc may instruct the linker to link the previously produced executable mtest with the object newly compiled from mtest.c plus several implicit libraries and startup objects, and output the result to v. ld would then be faced with duplicate defined symbols and output a list of these doublets. My proposal was meant as a test for an else correct openmpi installation, perhaps I should have said this. On successful completion, no message should be printed, but an executable named mtest should have been produced, which could then be run typing ./mtest. If the link process failed, then the openmpi installation was inconsistent. Kind regards Hugo On Wednesday 15 June 2011 17:45:41 Unger, Kristian, Dr. wrote: Dear Hugo I ran the command with the verbose switch and get the following output: mpicc mtest.c -ov mtest mtest: In function `_start': /usr/src/packages/BUILD/glibc-2.11.1/csu/../sysdeps/x86_64/elf/start.S:65 : multiple definition of `_start' /usr/lib64/gcc/x86_64-suse-linux/4.3/../../../../lib64/crt1.o:/usr/src/pa ck ages/BUILD/glibc-2.11.1/csu/../sysdeps/x86_64/elf/start.S:65: first defined here mtest: In function `_fini': (.fini+0x0): multiple definition of `_fini' /usr/lib64/gcc/x86_64-suse-linux/4.3/../../../../lib64/crti.o:initfini.c: (. fini+0x0): first defined here mtest:(.rodata+0x0): multiple definition of `_IO_stdin_used' /usr/lib64/gcc/x86_64-suse-linux/4.3/../../../../lib64/crt1.o:(.rodata.cs t4 +0x0): first defined here mtest: In function `__data_start': (.data+0x0): multiple definition of `__data_start' /usr/lib64/gcc/x86_64-suse-linux/4.3/../../../../lib64/crt1.o:(.data+0x0) : first defined here mtest: In function `main': (.text+0xec): multiple definition of `main' /tmp/ccUyk8e9.o:mtest.c:(.text+0x0): first defined here mtest: In function `_init': (.init+0x0): multiple definition of `_init' /usr/lib64/gcc/x86_64-suse-linux/4.3/../../../../lib64/crti.o:initfini.c: (. init+0x0): first defined here /usr/lib64/gcc/x86_64-suse-linux/4.3/../../../../x86_64-suse-linux/bin/ld : error in mtest(.eh_frame); no .eh_frame_hdr table will be created. collect2: ld returned 1 exit status Not sure if this makes any sense? Best wishes Kristian Dr. Kristian Unger Arbeitsgruppenleiter Integrative Biologie / Head of Integrative Biology Group Abteilung für Strahlenzytogenetik / Research Unit of Radiation Cytogenetics Tel.: +49-89-3187-3515 Mob.: +49-160-90641879 Am 15.06.11 17:25 schrieb Hugo Mildenberger unter hugo.mildenber...@web.de: Kristian, these are the usual problems with binary distributions. Regarding --with-Rmpi-include=/usr/lib64/mpi/gcc/openmpi/include and configure output checking for mpi.h... no ... so does /usr/lib64/mpi/gcc/openmpi/include really exist? At least, that appears to be a very unusual place to look for mpi.h (normally to be found in /usr/include ). And if you try to compile link the attached demo program: does the link phase succeed? Compile link using $ mpicc mtest.c -o mtest Presumably you have already tried to run install.packages(Rmpi). Kind regards Hugo On Wednesday 15 June 2011 16:22:07 Unger, Kristian, Dr. wrote: Thanks Hugo. I am pretty sure openmpi is installed: # zypper se openmpi Loading repository data... Reading installed packages... S | Name | Summary | Type --+---+-+--- i |
Re: [R] Rmpi installation
Kristian, Another update: I ditched the pre-installed version of openmpi and installed openmpi from source (./configure --without-openib). A very reasonable approach. Remembering my own struggles with SuSE years ago, I did not dare to propose that procedure. Compiling and running your test programme gives me the following error: unger@zytosrv01dmi:~/R_projects/OS ./mtest [zytosrv01dmi:6094] *** An error occurred in MPI_Send [zytosrv01dmi:6094] *** on communicator MPI_COMM_WORLD [zytosrv01dmi:6094] *** MPI_ERR_RANK: invalid rank [zytosrv01dmi:6094] *** MPI_ERRORS_ARE_FATAL (your MPI job will now abort) Same here. For the moment I can't figure out why this happens, also because the parameters of the openmpi functions are scarcely documented. Try one of the demo programs in the openmpi examples directory instead, e.g. the attached one However, Rmpi installs happily and it is also possible to call the package in R. When it comes to running a job under snowfall sfInit( parallel=TRUE,type=MPI, cpus=8) the job is split into eight separate processes which all run on the same CPU. So it seems that the processes are not distributed across the 8 CPUs available. sfSetMaxCPUs(number=8) should fix at least the thread distribution. However, on my notebook all cores are kept busy with 100% load after sfInit(), with nothing obvious to compute. I don't know if this has to be considered a bug or a feature. Kind regards Hugo Best wishes Kristian Am 15.06.11 19:25 schrieb Hugo Mildenberger unter hugo.mildenber...@web.de: Dear Kristian, please run exactly mpicc mtest.c -o mtest If you really need it, add -v separately. mpicc is nothing but a compiler wrapper. The -o switch specifies the outfile name, which has to follow immediately after -o, with or without a blank character in between. I'm not sure about what happens with -ov mtest. Probably an output file named v is produced on disk, in addition to the file mtest stemming from a previous with -o mtest Depending on mpicc's, gcc's and ld's version, gcc may instruct the linker to link the previously produced executable mtest with the object newly compiled from mtest.c plus several implicit libraries and startup objects, and output the result to v. ld would then be faced with duplicate defined symbols and output a list of these doublets. My proposal was meant as a test for an else correct openmpi installation, perhaps I should have said this. On successful completion, no message should be printed, but an executable named mtest should have been produced, which could then be run typing ./mtest. If the link process failed, then the openmpi installation was inconsistent. Kind regards Hugo On Wednesday 15 June 2011 17:45:41 Unger, Kristian, Dr. wrote: Dear Hugo I ran the command with the verbose switch and get the following output: mpicc mtest.c -ov mtest mtest: In function `_start': /usr/src/packages/BUILD/glibc-2.11.1/csu/../sysdeps/x86_64/elf/start.S:65 : multiple definition of `_start' /usr/lib64/gcc/x86_64-suse-linux/4.3/../../../../lib64/crt1.o:/usr/src/pa ck ages/BUILD/glibc-2.11.1/csu/../sysdeps/x86_64/elf/start.S:65: first defined here mtest: In function `_fini': (.fini+0x0): multiple definition of `_fini' /usr/lib64/gcc/x86_64-suse-linux/4.3/../../../../lib64/crti.o:initfini.c: (. fini+0x0): first defined here mtest:(.rodata+0x0): multiple definition of `_IO_stdin_used' /usr/lib64/gcc/x86_64-suse-linux/4.3/../../../../lib64/crt1.o:(.rodata.cs t4 +0x0): first defined here mtest: In function `__data_start': (.data+0x0): multiple definition of `__data_start' /usr/lib64/gcc/x86_64-suse-linux/4.3/../../../../lib64/crt1.o:(.data+0x0) : first defined here mtest: In function `main': (.text+0xec): multiple definition of `main' /tmp/ccUyk8e9.o:mtest.c:(.text+0x0): first defined here mtest: In function `_init': (.init+0x0): multiple definition of `_init' /usr/lib64/gcc/x86_64-suse-linux/4.3/../../../../lib64/crti.o:initfini.c: (. init+0x0): first defined here /usr/lib64/gcc/x86_64-suse-linux/4.3/../../../../x86_64-suse-linux/bin/ld : error in mtest(.eh_frame); no .eh_frame_hdr table will be created. collect2: ld returned 1 exit status Not sure if this makes any sense? Best wishes Kristian Dr. Kristian Unger Arbeitsgruppenleiter Integrative Biologie / Head of Integrative Biology Group Abteilung für Strahlenzytogenetik / Research Unit of Radiation Cytogenetics Tel.: +49-89-3187-3515 Mob.: +49-160-90641879 Am 15.06.11 17:25 schrieb Hugo Mildenberger unter hugo.mildenber...@web.de: Kristian, these are the usual problems with binary distributions. Regarding --with-Rmpi-include=/usr/lib64/mpi/gcc/openmpi/include and configure output
[R] Rmpi installation
Hi there I am trying to install Rmpi (version 0.5-4) on our 8-core SUSE Linux Enterprise Server 11 SP1. I read all I could find about Rmpi installation but still cannot get it working. Here the last command that I used: R CMD INSTALL --configure.args=--with-Rmpi-include=/usr/lib64/mpi/gcc/openmpi/include \ --with-Rmpi-libpath=/usr/lib64/mpi/gcc/openmpi/lib64 \ --with-Rmpi-type=OPENMPI Rmpi_0.5-4.tar.gz Resulting in the following: zytosrv01dmi:/home/unger/R_projects/OS # R CMD INSTALL --configure-args=--with-Rmpi-include=/usr/lib64/mpi/gcc/openmpi/include \--with-Rmpi-libpath=/usr/lib64/mpi/gcc/openmpi/lib64 \ --with-Rmpi-type=OPENMPI Rmpi_0.5-4.tar.gz * installing to library Œ/usr/local/lib64/R/library¹ * installing *source* package ŒRmpi¹ ... 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 checking how to run the C preprocessor... gcc -E checking for grep that handles long lines and -e... /usr/bin/grep checking for egrep... /usr/bin/grep -E checking for ANSI C header files... yes checking for sys/types.h... yes checking for sys/stat.h... yes checking for stdlib.h... yes checking for string.h... yes checking for memory.h... yes checking for strings.h... yes checking for inttypes.h... yes checking for stdint.h... yes checking for unistd.h... yes checking mpi.h usability... no checking mpi.h presence... no checking for mpi.h... no Try to find libmpi or libmpich ... checking for main in -lmpi... no libmpi not found. exiting... ERROR: configuration failed for package ŒRmpi¹ * removing Œ/usr/local/lib64/R/library/Rmpi¹ So obviously libmpi is not found. Doing a locate search for libmpi (straight after updatedb) shows: zytosrv01dmi:/home/unger/R_projects/OS # locate libmpi /opt/mpich/ch-p4/lib64/libmpich.a /opt/mpich/ch-p4/lib64/libmpichf90.a /opt/mpich/ch-p4/lib64/libmpichf90nc.a /opt/mpich/ch-p4/lib64/libmpichfarg.a /opt/mpich/ch-p4/lib64/libmpichfsup.a /opt/mpich/ch-p4mpd/lib64/libmpich.a /opt/mpich/ch-p4mpd/lib64/libmpichf90.a /opt/mpich/ch-p4mpd/lib64/libmpichf90nc.a /opt/mpich/ch-p4mpd/lib64/libmpichfarg.a /opt/mpich/ch-p4mpd/lib64/libmpichfsup.a /usr/lib64/mpi/gcc/openmpi/lib64/libmpi_cxx.la /usr/lib64/mpi/gcc/openmpi/lib64/libmpi_cxx.so /usr/lib64/mpi/gcc/openmpi/lib64/libmpi_cxx.so.0 /usr/lib64/mpi/gcc/openmpi/lib64/libmpi_cxx.so.0.0.0 /usr/lib64/mpi/gcc/openmpi/lib64/libmpi_f77.la /usr/lib64/mpi/gcc/openmpi/lib64/libmpi_f77.so /usr/lib64/mpi/gcc/openmpi/lib64/libmpi_f77.so.0 /usr/lib64/mpi/gcc/openmpi/lib64/libmpi_f77.so.0.0.0 /usr/lib64/mpi/gcc/openmpi/lib64/libmpi_f90.la /usr/lib64/mpi/gcc/openmpi/lib64/libmpi_f90.so /usr/lib64/mpi/gcc/openmpi/lib64/libmpi_f90.so.0 /usr/lib64/mpi/gcc/openmpi/lib64/libmpi_f90.so.0.0.0 /usr/lib64/mpi/gcc/openmpi/lib64/libmpi.la /usr/lib64/mpi/gcc/openmpi/lib64/libmpi.so /usr/lib64/mpi/gcc/openmpi/lib64/libmpi.so.0 /usr/lib64/mpi/gcc/openmpi/lib64/libmpi.so.0.0.0 What lets me assume that the path to libmpi is /usr/lib64/mpi/gcc/openmpi/lib64 which I already included in the options... How can I get this working? I would highly appreciate any help on this! Best wishes Kristian Dr. Kristian Unger Arbeitsgruppenleiter Integrative Biologie / Head of Integrative Biology Group Abteilung für Strahlenzytogenetik / Research Unit of Radiation Cytogenetics Helmholtz Zentrum München Deutsches Forschungszentrum für Gesundheit und Umwelt (GmbH) Ingolstädter Landstr. 1 85764 Neuherberg www.helmholtz-muenchen.de Aufsichtsratsvorsitzende: MinDir´in Bärbel Brumme-Bothe Geschäftsführer: Prof. Dr. Günther Wess und Dr. Nikolaus Blum Registergericht: Amtsgericht München HRB 6466 USt-IdNr: DE 129521671 __ R-help@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.
Re: [R] Rmpi installation
Hmm, looks like there was a trailing blank after the backslash and before end of line, resulting in --with-Rmpi-libpath possibly not recognised: \--with-Rmpi-libpath=/usr/lib64/mpi/gcc/openmpi/lib64 \ I also doubt there is real need to escape newlines within a string. But another possible problem source is that according to R CMD INSTALL --help, the parameter is called --configure-args, not configure.args. $ R CMD INSTALL --configure-args=--with-Rmpi-include=/usr/lib64/mpi/gcc/openmpi/include --with-Rmpi-libpath=/usr/lib64/mpi/gcc/openmpi/lib64 -- with-Rmpi-type=OPENMPI Rmpi_0.5-4.tar.gz Best On Wednesday 15 June 2011 14:25:46 Unger, Kristian, Dr. wrote: Hi there I am trying to install Rmpi (version 0.5-4) on our 8-core SUSE Linux Enterprise Server 11 SP1. I read all I could find about Rmpi installation but still cannot get it working. Here the last command that I used: R CMD INSTALL --configure.args=--with-Rmpi-include=/usr/lib64/mpi/gcc/openmpi/include \ --with-Rmpi-libpath=/usr/lib64/mpi/gcc/openmpi/lib64 \ --with-Rmpi-type=OPENMPI Rmpi_0.5-4.tar.gz Resulting in the following: zytosrv01dmi:/home/unger/R_projects/OS # R CMD INSTALL --configure-args=--with-Rmpi-include=/usr/lib64/mpi/gcc/openmpi/include \--with-Rmpi-libpath=/usr/lib64/mpi/gcc/openmpi/lib64 \ --with-Rmpi-type=OPENMPI Rmpi_0.5-4.tar.gz * installing to library Œ/usr/local/lib64/R/library¹ * installing *source* package ŒRmpi¹ ... 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 checking how to run the C preprocessor... gcc -E checking for grep that handles long lines and -e... /usr/bin/grep checking for egrep... /usr/bin/grep -E checking for ANSI C header files... yes checking for sys/types.h... yes checking for sys/stat.h... yes checking for stdlib.h... yes checking for string.h... yes checking for memory.h... yes checking for strings.h... yes checking for inttypes.h... yes checking for stdint.h... yes checking for unistd.h... yes checking mpi.h usability... no checking mpi.h presence... no checking for mpi.h... no Try to find libmpi or libmpich ... checking for main in -lmpi... no libmpi not found. exiting... ERROR: configuration failed for package ŒRmpi¹ * removing Œ/usr/local/lib64/R/library/Rmpi¹ So obviously libmpi is not found. Doing a locate search for libmpi (straight after updatedb) shows: zytosrv01dmi:/home/unger/R_projects/OS # locate libmpi /opt/mpich/ch-p4/lib64/libmpich.a /opt/mpich/ch-p4/lib64/libmpichf90.a /opt/mpich/ch-p4/lib64/libmpichf90nc.a /opt/mpich/ch-p4/lib64/libmpichfarg.a /opt/mpich/ch-p4/lib64/libmpichfsup.a /opt/mpich/ch-p4mpd/lib64/libmpich.a /opt/mpich/ch-p4mpd/lib64/libmpichf90.a /opt/mpich/ch-p4mpd/lib64/libmpichf90nc.a /opt/mpich/ch-p4mpd/lib64/libmpichfarg.a /opt/mpich/ch-p4mpd/lib64/libmpichfsup.a /usr/lib64/mpi/gcc/openmpi/lib64/libmpi_cxx.la /usr/lib64/mpi/gcc/openmpi/lib64/libmpi_cxx.so /usr/lib64/mpi/gcc/openmpi/lib64/libmpi_cxx.so.0 /usr/lib64/mpi/gcc/openmpi/lib64/libmpi_cxx.so.0.0.0 /usr/lib64/mpi/gcc/openmpi/lib64/libmpi_f77.la /usr/lib64/mpi/gcc/openmpi/lib64/libmpi_f77.so /usr/lib64/mpi/gcc/openmpi/lib64/libmpi_f77.so.0 /usr/lib64/mpi/gcc/openmpi/lib64/libmpi_f77.so.0.0.0 /usr/lib64/mpi/gcc/openmpi/lib64/libmpi_f90.la /usr/lib64/mpi/gcc/openmpi/lib64/libmpi_f90.so /usr/lib64/mpi/gcc/openmpi/lib64/libmpi_f90.so.0 /usr/lib64/mpi/gcc/openmpi/lib64/libmpi_f90.so.0.0.0 /usr/lib64/mpi/gcc/openmpi/lib64/libmpi.la /usr/lib64/mpi/gcc/openmpi/lib64/libmpi.so /usr/lib64/mpi/gcc/openmpi/lib64/libmpi.so.0 /usr/lib64/mpi/gcc/openmpi/lib64/libmpi.so.0.0.0 What lets me assume that the path to libmpi is /usr/lib64/mpi/gcc/openmpi/lib64 which I already included in the options... How can I get this working? I would highly appreciate any help on this! Best wishes Kristian Dr. Kristian Unger Arbeitsgruppenleiter Integrative Biologie / Head of Integrative Biology Group Abteilung für Strahlenzytogenetik / Research Unit of Radiation Cytogenetics Helmholtz Zentrum München Deutsches Forschungszentrum für Gesundheit und Umwelt (GmbH) Ingolstädter Landstr. 1 85764 Neuherberg www.helmholtz-muenchen.de Aufsichtsratsvorsitzende: MinDir´in Bärbel Brumme-Bothe Geschäftsführer: Prof. Dr. Günther Wess und Dr. Nikolaus Blum Registergericht: Amtsgericht München HRB 6466 USt-IdNr: DE 129521671 __ R-help@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide
Re: [R] Rmpi installation
Thank you very much Hugo. Using the command as suggested results exactly the same error: # R CMD INSTALL --configure-args=--with-Rmpi-include=/usr/lib64/mpi/gcc/openmpi/include --with-Rmpi-libpath=/usr/lib64/mpi/gcc/openmpi/lib64 --with-Rmpi-type=OPENMPI Rmpi_0.5-4.tar.gz * installing to library ‘/usr/local/lib64/R/library’ * 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 checking how to run the C preprocessor... gcc -E checking for grep that handles long lines and -e... /usr/bin/grep checking for egrep... /usr/bin/grep -E checking for ANSI C header files... yes checking for sys/types.h... yes checking for sys/stat.h... yes checking for stdlib.h... yes checking for string.h... yes checking for memory.h... yes checking for strings.h... yes checking for inttypes.h... yes checking for stdint.h... yes checking for unistd.h... yes checking mpi.h usability... no checking mpi.h presence... no checking for mpi.h... no Try to find libmpi or libmpich ... checking for main in -lmpi... no libmpi not found. exiting... ERROR: configuration failed for package ‘Rmpi’ * removing ‘/usr/local/lib64/R/library/Rmpi’ Best wishes Kristian Dr. Kristian Unger Arbeitsgruppenleiter Integrative Biologie / Head of Integrative Biology Group Abteilung für Strahlenzytogenetik / Research Unit of Radiation Cytogenetics Tel.: +49-89-3187-3515 Mob.: +49-160-90641879 Am 15.06.11 15:12 schrieb Hugo Mildenberger unter hugo.mildenber...@web.de: Hmm, looks like there was a trailing blank after the backslash and before end of line, resulting in --with-Rmpi-libpath possibly not recognised: \--with-Rmpi-libpath=/usr/lib64/mpi/gcc/openmpi/lib64 \ I also doubt there is real need to escape newlines within a string. But another possible problem source is that according to R CMD INSTALL --help, the parameter is called --configure-args, not configure.args. $ R CMD INSTALL --configure-args=--with-Rmpi-include=/usr/lib64/mpi/gcc/openmpi/include --with-Rmpi-libpath=/usr/lib64/mpi/gcc/openmpi/lib64 -- with-Rmpi-type=OPENMPI Rmpi_0.5-4.tar.gz Best On Wednesday 15 June 2011 14:25:46 Unger, Kristian, Dr. wrote: Hi there I am trying to install Rmpi (version 0.5-4) on our 8-core SUSE Linux Enterprise Server 11 SP1. I read all I could find about Rmpi installation but still cannot get it working. Here the last command that I used: R CMD INSTALL --configure.args=--with-Rmpi-include=/usr/lib64/mpi/gcc/openmpi/include \ --with-Rmpi-libpath=/usr/lib64/mpi/gcc/openmpi/lib64 \ --with-Rmpi-type=OPENMPI Rmpi_0.5-4.tar.gz Resulting in the following: zytosrv01dmi:/home/unger/R_projects/OS # R CMD INSTALL --configure-args=--with-Rmpi-include=/usr/lib64/mpi/gcc/openmpi/include \--with-Rmpi-libpath=/usr/lib64/mpi/gcc/openmpi/lib64 \ --with-Rmpi-type=OPENMPI Rmpi_0.5-4.tar.gz * installing to library Œ/usr/local/lib64/R/library¹ * installing *source* package ŒRmpi¹ ... 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 checking how to run the C preprocessor... gcc -E checking for grep that handles long lines and -e... /usr/bin/grep checking for egrep... /usr/bin/grep -E checking for ANSI C header files... yes checking for sys/types.h... yes checking for sys/stat.h... yes checking for stdlib.h... yes checking for string.h... yes checking for memory.h... yes checking for strings.h... yes checking for inttypes.h... yes checking for stdint.h... yes checking for unistd.h... yes checking mpi.h usability... no checking mpi.h presence... no checking for mpi.h... no Try to find libmpi or libmpich ... checking for main in -lmpi... no libmpi not found. exiting... ERROR: configuration failed for package ŒRmpi¹ * removing Œ/usr/local/lib64/R/library/Rmpi¹ So obviously libmpi is not found. Doing a locate search for libmpi (straight after updatedb) shows: zytosrv01dmi:/home/unger/R_projects/OS # locate libmpi /opt/mpich/ch-p4/lib64/libmpich.a /opt/mpich/ch-p4/lib64/libmpichf90.a /opt/mpich/ch-p4/lib64/libmpichf90nc.a /opt/mpich/ch-p4/lib64/libmpichfarg.a /opt/mpich/ch-p4/lib64/libmpichfsup.a /opt/mpich/ch-p4mpd/lib64/libmpich.a /opt/mpich/ch-p4mpd/lib64/libmpichf90.a /opt/mpich/ch-p4mpd/lib64/libmpichf90nc.a
Re: [R] Rmpi installation
Kristian, I just tried that particular command here on a Gentoo system with openmpi-1.5.3 installed, because I wondered why the Rmpi configure script tests for main function in a shared library ... But I got a completly different output from configure. While the linker succeeds here, the package load test does not. However, on your system, may be the openmpi installation really is a kinda private one of gcc? I heard gcc makes use of openmpi internally. So is openmpi really installed? I just recognize that you are trying to use Rmpi_0.5-4.tar.gz while current version on CRAN is Rmpi_0.5-9.tar.gz. Best Hugo R CMD INSTALL --configure-args=--with-Rmpi-include=/usr/include --with-Rmpi-libpath=/usr/lib64/openmpi --with-Rmpi-type=OPENMPI Rmpi_0.5-9.tar.gz * installing to library ‘/home/hm/R/x86_64-pc-linux-gnu-library/2.13’ * installing *source* package ‘Rmpi’ ... checking for openpty in -lutil... no checking for main in -lpthread... no configure: creating ./config.status config.status: creating src/Makevars ** libs ** libs x86_64-pc-linux-gnu-gcc -std=gnu99 -I/usr/lib64/R/include -DPACKAGE_NAME=\\ -DPACKAGE_TARNAME=\\ -DPACKAGE_VERSION=\\ - DPACKAGE_STRING=\\ -DPACKAGE_BUGREPORT=\\ -I/usr/include -DMPI2 -DOPENMPI -I/usr/local/include-fpic -O3 -pipe -march=core2 - mtune=core2 -ggdb -c RegQuery.c -o RegQuery.o x86_64-pc-linux-gnu-gcc -std=gnu99 -I/usr/lib64/R/include -DPACKAGE_NAME=\\ -DPACKAGE_TARNAME=\\ -DPACKAGE_VERSION=\\ - DPACKAGE_STRING=\\ -DPACKAGE_BUGREPORT=\\ -I/usr/include -DMPI2 -DOPENMPI -I/usr/local/include-fpic -O3 -pipe -march=core2 - mtune=core2 -ggdb -c Rmpi.c -o Rmpi.o x86_64-pc-linux-gnu-gcc -std=gnu99 -I/usr/lib64/R/include -DPACKAGE_NAME=\\ -DPACKAGE_TARNAME=\\ -DPACKAGE_VERSION=\\ - DPACKAGE_STRING=\\ -DPACKAGE_BUGREPORT=\\ -I/usr/include -DMPI2 -DOPENMPI -I/usr/local/include-fpic -O3 -pipe -march=core2 - mtune=core2 -ggdb -c conversion.c -o conversion.o x86_64-pc-linux-gnu-gcc -std=gnu99 -I/usr/lib64/R/include -DPACKAGE_NAME=\\ -DPACKAGE_TARNAME=\\ -DPACKAGE_VERSION=\\ - DPACKAGE_STRING=\\ -DPACKAGE_BUGREPORT=\\ -I/usr/include -DMPI2 -DOPENMPI -I/usr/local/include-fpic -O3 -pipe -march=core2 - mtune=core2 -ggdb -c internal.c -o internal.o x86_64-pc-linux-gnu-gcc -std=gnu99 -shared -Wl,-O1 -Wl,--as-needed -o Rmpi.so RegQuery.o Rmpi.o conversion.o internal.o -L/usr/lib64/openmpi -lmpi -L/usr/lib64/R/lib -lR installiert nach /home/hm/R/x86_64-pc-linux-gnu-library/2.13/Rmpi/libs [...] ompi_mpi_init: orte_init failed -- Returned Not found (-13) instead of Success (0) On Wednesday 15 June 2011 15:19:22 Unger, Kristian, Dr. wrote: Thank you very much Hugo. Using the command as suggested results exactly the same error: # R CMD INSTALL --configure-args=--with-Rmpi-include=/usr/lib64/mpi/gcc/openmpi/include --with-Rmpi-libpath=/usr/lib64/mpi/gcc/openmpi/lib64 --with-Rmpi-type=OPENMPI Rmpi_0.5-4.tar.gz * installing to library ‘/usr/local/lib64/R/library’ * 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 checking how to run the C preprocessor... gcc -E checking for grep that handles long lines and -e... /usr/bin/grep checking for egrep... /usr/bin/grep -E checking for ANSI C header files... yes checking for sys/types.h... yes checking for sys/stat.h... yes checking for stdlib.h... yes checking for string.h... yes checking for memory.h... yes checking for strings.h... yes checking for inttypes.h... yes checking for stdint.h... yes checking for unistd.h... yes checking mpi.h usability... no checking mpi.h presence... no checking for mpi.h... no Try to find libmpi or libmpich ... checking for main in -lmpi... no libmpi not found. exiting... ERROR: configuration failed for package ‘Rmpi’ * removing ‘/usr/local/lib64/R/library/Rmpi’ Best wishes Kristian Dr. Kristian Unger Arbeitsgruppenleiter Integrative Biologie / Head of Integrative Biology Group Abteilung für Strahlenzytogenetik / Research Unit of Radiation Cytogenetics Tel.: +49-89-3187-3515 Mob.: +49-160-90641879 Am 15.06.11 15:12 schrieb Hugo Mildenberger unter hugo.mildenber...@web.de: Hmm, looks like there was a trailing blank after the backslash and before end of line, resulting in --with-Rmpi-libpath possibly not recognised: \--with-Rmpi-libpath=/usr/lib64/mpi/gcc/openmpi/lib64 \ I also doubt there is real need to escape newlines within a string. But another possible problem source is that according to R CMD INSTALL --help, the parameter is called
Re: [R] Rmpi installation
Thanks Hugo. I am pretty sure openmpi is installed: # zypper se openmpi Loading repository data... Reading installed packages... S | Name | Summary | Type --+---+-+--- i | openmpi | A powerful implementaion of MPI | package | openmpi | A powerful implementaion of MPI | srcpackage i | openmpi-devel | A powerful implementaion of MPI | package I got the same error message with the latest version available. The reason why I took the somewhat older version is that I wanted to make sure that it is not related to any libraries used by the newest version. Best wishes Kristian Dr. Kristian Unger Arbeitsgruppenleiter Integrative Biologie / Head of Integrative Biology Group Abteilung für Strahlenzytogenetik / Research Unit of Radiation Cytogenetics Tel.: +49-89-3187-3515 Mob.: +49-160-90641879 Am 15.06.11 16:16 schrieb Hugo Mildenberger unter hugo.mildenber...@web.de: Kristian, I just tried that particular command here on a Gentoo system with openmpi-1.5.3 installed, because I wondered why the Rmpi configure script tests for main function in a shared library ... But I got a completly different output from configure. While the linker succeeds here, the package load test does not. However, on your system, may be the openmpi installation really is a kinda private one of gcc? I heard gcc makes use of openmpi internally. So is openmpi really installed? I just recognize that you are trying to use Rmpi_0.5-4.tar.gz while current version on CRAN is Rmpi_0.5-9.tar.gz. Best Hugo R CMD INSTALL --configure-args=--with-Rmpi-include=/usr/include --with-Rmpi-libpath=/usr/lib64/openmpi --with-Rmpi-type=OPENMPI Rmpi_0.5-9.tar.gz * installing to library ‘/home/hm/R/x86_64-pc-linux-gnu-library/2.13’ * installing *source* package ‘Rmpi’ ... checking for openpty in -lutil... no checking for main in -lpthread... no configure: creating ./config.status config.status: creating src/Makevars ** libs ** libs x86_64-pc-linux-gnu-gcc -std=gnu99 -I/usr/lib64/R/include -DPACKAGE_NAME=\\ -DPACKAGE_TARNAME=\\ -DPACKAGE_VERSION=\\ - DPACKAGE_STRING=\\ -DPACKAGE_BUGREPORT=\\ -I/usr/include -DMPI2 -DOPENMPI -I/usr/local/include-fpic -O3 -pipe -march=core2 - mtune=core2 -ggdb -c RegQuery.c -o RegQuery.o x86_64-pc-linux-gnu-gcc -std=gnu99 -I/usr/lib64/R/include -DPACKAGE_NAME=\\ -DPACKAGE_TARNAME=\\ -DPACKAGE_VERSION=\\ - DPACKAGE_STRING=\\ -DPACKAGE_BUGREPORT=\\ -I/usr/include -DMPI2 -DOPENMPI -I/usr/local/include-fpic -O3 -pipe -march=core2 - mtune=core2 -ggdb -c Rmpi.c -o Rmpi.o x86_64-pc-linux-gnu-gcc -std=gnu99 -I/usr/lib64/R/include -DPACKAGE_NAME=\\ -DPACKAGE_TARNAME=\\ -DPACKAGE_VERSION=\\ - DPACKAGE_STRING=\\ -DPACKAGE_BUGREPORT=\\ -I/usr/include -DMPI2 -DOPENMPI -I/usr/local/include-fpic -O3 -pipe -march=core2 - mtune=core2 -ggdb -c conversion.c -o conversion.o x86_64-pc-linux-gnu-gcc -std=gnu99 -I/usr/lib64/R/include -DPACKAGE_NAME=\\ -DPACKAGE_TARNAME=\\ -DPACKAGE_VERSION=\\ - DPACKAGE_STRING=\\ -DPACKAGE_BUGREPORT=\\ -I/usr/include -DMPI2 -DOPENMPI -I/usr/local/include-fpic -O3 -pipe -march=core2 - mtune=core2 -ggdb -c internal.c -o internal.o x86_64-pc-linux-gnu-gcc -std=gnu99 -shared -Wl,-O1 -Wl,--as-needed -o Rmpi.so RegQuery.o Rmpi.o conversion.o internal.o -L/usr/lib64/openmpi -lmpi -L/usr/lib64/R/lib -lR installiert nach /home/hm/R/x86_64-pc-linux-gnu-library/2.13/Rmpi/libs [...] ompi_mpi_init: orte_init failed -- Returned Not found (-13) instead of Success (0) On Wednesday 15 June 2011 15:19:22 Unger, Kristian, Dr. wrote: Thank you very much Hugo. Using the command as suggested results exactly the same error: # R CMD INSTALL --configure-args=--with-Rmpi-include=/usr/lib64/mpi/gcc/openmpi/include --with-Rmpi-libpath=/usr/lib64/mpi/gcc/openmpi/lib64 --with-Rmpi-type=OPENMPI Rmpi_0.5-4.tar.gz * installing to library ‘/usr/local/lib64/R/library’ * 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 checking how to run the C preprocessor... gcc -E checking for grep that handles long lines and -e... /usr/bin/grep checking for egrep... /usr/bin/grep -E checking for ANSI C header files... yes checking for sys/types.h... yes checking for sys/stat.h... yes checking for stdlib.h... yes checking for string.h... yes checking for memory.h... yes checking for strings.h... yes checking for inttypes.h... yes checking for stdint.h... yes checking for unistd.h... yes checking mpi.h usability... no checking mpi.h presence... no checking for
Re: [R] Rmpi installation
Kristian, these are the usual problems with binary distributions. Regarding --with-Rmpi-include=/usr/lib64/mpi/gcc/openmpi/include and configure output checking for mpi.h... no ... so does /usr/lib64/mpi/gcc/openmpi/include really exist? At least, that appears to be a very unusual place to look for mpi.h (normally to be found in /usr/include ). And if you try to compile link the attached demo program: does the link phase succeed? Compile link using $ mpicc mtest.c -o mtest Presumably you have already tried to run install.packages(Rmpi). Kind regards Hugo On Wednesday 15 June 2011 16:22:07 Unger, Kristian, Dr. wrote: Thanks Hugo. I am pretty sure openmpi is installed: # zypper se openmpi Loading repository data... Reading installed packages... S | Name | Summary | Type --+---+-+--- i | openmpi | A powerful implementaion of MPI | package | openmpi | A powerful implementaion of MPI | srcpackage i | openmpi-devel | A powerful implementaion of MPI | package I got the same error message with the latest version available. The reason why I took the somewhat older version is that I wanted to make sure that it is not related to any libraries used by the newest version. Best wishes Kristian Dr. Kristian Unger Arbeitsgruppenleiter Integrative Biologie / Head of Integrative Biology Group Abteilung für Strahlenzytogenetik / Research Unit of Radiation Cytogenetics Tel.: +49-89-3187-3515 Mob.: +49-160-90641879 Am 15.06.11 16:16 schrieb Hugo Mildenberger unter hugo.mildenber...@web.de: Kristian, I just tried that particular command here on a Gentoo system with openmpi-1.5.3 installed, because I wondered why the Rmpi configure script tests for main function in a shared library ... But I got a completly different output from configure. While the linker succeeds here, the package load test does not. However, on your system, may be the openmpi installation really is a kinda private one of gcc? I heard gcc makes use of openmpi internally. So is openmpi really installed? I just recognize that you are trying to use Rmpi_0.5-4.tar.gz while current version on CRAN is Rmpi_0.5-9.tar.gz. Best Hugo R CMD INSTALL --configure-args=--with-Rmpi-include=/usr/include --with-Rmpi-libpath=/usr/lib64/openmpi --with-Rmpi-type=OPENMPI Rmpi_0.5-9.tar.gz * installing to library ‘/home/hm/R/x86_64-pc-linux-gnu-library/2.13’ * installing *source* package ‘Rmpi’ ... checking for openpty in -lutil... no checking for main in -lpthread... no configure: creating ./config.status config.status: creating src/Makevars ** libs ** libs x86_64-pc-linux-gnu-gcc -std=gnu99 -I/usr/lib64/R/include -DPACKAGE_NAME=\\ -DPACKAGE_TARNAME=\\ -DPACKAGE_VERSION=\\ - DPACKAGE_STRING=\\ -DPACKAGE_BUGREPORT=\\ -I/usr/include -DMPI2 -DOPENMPI -I/usr/local/include-fpic -O3 -pipe -march=core2 - mtune=core2 -ggdb -c RegQuery.c -o RegQuery.o x86_64-pc-linux-gnu-gcc -std=gnu99 -I/usr/lib64/R/include -DPACKAGE_NAME=\\ -DPACKAGE_TARNAME=\\ -DPACKAGE_VERSION=\\ - DPACKAGE_STRING=\\ -DPACKAGE_BUGREPORT=\\ -I/usr/include -DMPI2 -DOPENMPI -I/usr/local/include-fpic -O3 -pipe -march=core2 - mtune=core2 -ggdb -c Rmpi.c -o Rmpi.o x86_64-pc-linux-gnu-gcc -std=gnu99 -I/usr/lib64/R/include -DPACKAGE_NAME=\\ -DPACKAGE_TARNAME=\\ -DPACKAGE_VERSION=\\ - DPACKAGE_STRING=\\ -DPACKAGE_BUGREPORT=\\ -I/usr/include -DMPI2 -DOPENMPI -I/usr/local/include-fpic -O3 -pipe -march=core2 - mtune=core2 -ggdb -c conversion.c -o conversion.o x86_64-pc-linux-gnu-gcc -std=gnu99 -I/usr/lib64/R/include -DPACKAGE_NAME=\\ -DPACKAGE_TARNAME=\\ -DPACKAGE_VERSION=\\ - DPACKAGE_STRING=\\ -DPACKAGE_BUGREPORT=\\ -I/usr/include -DMPI2 -DOPENMPI -I/usr/local/include-fpic -O3 -pipe -march=core2 - mtune=core2 -ggdb -c internal.c -o internal.o x86_64-pc-linux-gnu-gcc -std=gnu99 -shared -Wl,-O1 -Wl,--as-needed -o Rmpi.so RegQuery.o Rmpi.o conversion.o internal.o -L/usr/lib64/openmpi -lmpi -L/usr/lib64/R/lib -lR installiert nach /home/hm/R/x86_64-pc-linux-gnu-library/2.13/Rmpi/libs [...] ompi_mpi_init: orte_init failed -- Returned Not found (-13) instead of Success (0) On Wednesday 15 June 2011 15:19:22 Unger, Kristian, Dr. wrote: Thank you very much Hugo. Using the command as suggested results exactly the same error: # R CMD INSTALL --configure-args=--with-Rmpi-include=/usr/lib64/mpi/gcc/openmpi/include --with-Rmpi-libpath=/usr/lib64/mpi/gcc/openmpi/lib64 --with-Rmpi-type=OPENMPI Rmpi_0.5-4.tar.gz * installing to library ‘/usr/local/lib64/R/library’ * 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
Re: [R] Rmpi installation
Dear Hugo I ran the command with the verbose switch and get the following output: mpicc mtest.c -ov mtest mtest: In function `_start': /usr/src/packages/BUILD/glibc-2.11.1/csu/../sysdeps/x86_64/elf/start.S:65: multiple definition of `_start' /usr/lib64/gcc/x86_64-suse-linux/4.3/../../../../lib64/crt1.o:/usr/src/pack ages/BUILD/glibc-2.11.1/csu/../sysdeps/x86_64/elf/start.S:65: first defined here mtest: In function `_fini': (.fini+0x0): multiple definition of `_fini' /usr/lib64/gcc/x86_64-suse-linux/4.3/../../../../lib64/crti.o:initfini.c:(. fini+0x0): first defined here mtest:(.rodata+0x0): multiple definition of `_IO_stdin_used' /usr/lib64/gcc/x86_64-suse-linux/4.3/../../../../lib64/crt1.o:(.rodata.cst4 +0x0): first defined here mtest: In function `__data_start': (.data+0x0): multiple definition of `__data_start' /usr/lib64/gcc/x86_64-suse-linux/4.3/../../../../lib64/crt1.o:(.data+0x0): first defined here mtest: In function `main': (.text+0xec): multiple definition of `main' /tmp/ccUyk8e9.o:mtest.c:(.text+0x0): first defined here mtest: In function `_init': (.init+0x0): multiple definition of `_init' /usr/lib64/gcc/x86_64-suse-linux/4.3/../../../../lib64/crti.o:initfini.c:(. init+0x0): first defined here /usr/lib64/gcc/x86_64-suse-linux/4.3/../../../../x86_64-suse-linux/bin/ld: error in mtest(.eh_frame); no .eh_frame_hdr table will be created. collect2: ld returned 1 exit status Not sure if this makes any sense? Best wishes Kristian Dr. Kristian Unger Arbeitsgruppenleiter Integrative Biologie / Head of Integrative Biology Group Abteilung für Strahlenzytogenetik / Research Unit of Radiation Cytogenetics Tel.: +49-89-3187-3515 Mob.: +49-160-90641879 Am 15.06.11 17:25 schrieb Hugo Mildenberger unter hugo.mildenber...@web.de: Kristian, these are the usual problems with binary distributions. Regarding --with-Rmpi-include=/usr/lib64/mpi/gcc/openmpi/include and configure output checking for mpi.h... no ... so does /usr/lib64/mpi/gcc/openmpi/include really exist? At least, that appears to be a very unusual place to look for mpi.h (normally to be found in /usr/include ). And if you try to compile link the attached demo program: does the link phase succeed? Compile link using $ mpicc mtest.c -o mtest Presumably you have already tried to run install.packages(Rmpi). Kind regards Hugo On Wednesday 15 June 2011 16:22:07 Unger, Kristian, Dr. wrote: Thanks Hugo. I am pretty sure openmpi is installed: # zypper se openmpi Loading repository data... Reading installed packages... S | Name | Summary | Type --+---+-+--- i | openmpi | A powerful implementaion of MPI | package | openmpi | A powerful implementaion of MPI | srcpackage i | openmpi-devel | A powerful implementaion of MPI | package I got the same error message with the latest version available. The reason why I took the somewhat older version is that I wanted to make sure that it is not related to any libraries used by the newest version. Best wishes Kristian Dr. Kristian Unger Arbeitsgruppenleiter Integrative Biologie / Head of Integrative Biology Group Abteilung für Strahlenzytogenetik / Research Unit of Radiation Cytogenetics Tel.: +49-89-3187-3515 Mob.: +49-160-90641879 Am 15.06.11 16:16 schrieb Hugo Mildenberger unter hugo.mildenber...@web.de: Kristian, I just tried that particular command here on a Gentoo system with openmpi-1.5.3 installed, because I wondered why the Rmpi configure script tests for main function in a shared library ... But I got a completly different output from configure. While the linker succeeds here, the package load test does not. However, on your system, may be the openmpi installation really is a kinda private one of gcc? I heard gcc makes use of openmpi internally. So is openmpi really installed? I just recognize that you are trying to use Rmpi_0.5-4.tar.gz while current version on CRAN is Rmpi_0.5-9.tar.gz. Best Hugo R CMD INSTALL --configure-args=--with-Rmpi-include=/usr/include --with-Rmpi-libpath=/usr/lib64/openmpi --with-Rmpi-type=OPENMPI Rmpi_0.5-9.tar.gz * installing to library ‘/home/hm/R/x86_64-pc-linux-gnu-library/2.13’ * installing *source* package ‘Rmpi’ ... checking for openpty in -lutil... no checking for main in -lpthread... no configure: creating ./config.status config.status: creating src/Makevars ** libs ** libs x86_64-pc-linux-gnu-gcc -std=gnu99 -I/usr/lib64/R/include -DPACKAGE_NAME=\\ -DPACKAGE_TARNAME=\\ -DPACKAGE_VERSION=\\ - DPACKAGE_STRING=\\ -DPACKAGE_BUGREPORT=\\ -I/usr/include -DMPI2 -DOPENMPI -I/usr/local/include-fpic -O3 -pipe -march=core2 - mtune=core2 -ggdb -c RegQuery.c -o RegQuery.o x86_64-pc-linux-gnu-gcc -std=gnu99 -I/usr/lib64/R/include -DPACKAGE_NAME=\\
Re: [R] Rmpi installation
Dear Kristian, please run exactly mpicc mtest.c -o mtest If you really need it, add -v separately. mpicc is nothing but a compiler wrapper. The -o switch specifies the outfile name, which has to follow immediately after -o, with or without a blank character in between. I'm not sure about what happens with -ov mtest. Probably an output file named v is produced on disk, in addition to the file mtest stemming from a previous with -o mtest Depending on mpicc's, gcc's and ld's version, gcc may instruct the linker to link the previously produced executable mtest with the object newly compiled from mtest.c plus several implicit libraries and startup objects, and output the result to v. ld would then be faced with duplicate defined symbols and output a list of these doublets. My proposal was meant as a test for an else correct openmpi installation, perhaps I should have said this. On successful completion, no message should be printed, but an executable named mtest should have been produced, which could then be run typing ./mtest. If the link process failed, then the openmpi installation was inconsistent. Kind regards Hugo On Wednesday 15 June 2011 17:45:41 Unger, Kristian, Dr. wrote: Dear Hugo I ran the command with the verbose switch and get the following output: mpicc mtest.c -ov mtest mtest: In function `_start': /usr/src/packages/BUILD/glibc-2.11.1/csu/../sysdeps/x86_64/elf/start.S:65: multiple definition of `_start' /usr/lib64/gcc/x86_64-suse-linux/4.3/../../../../lib64/crt1.o:/usr/src/pack ages/BUILD/glibc-2.11.1/csu/../sysdeps/x86_64/elf/start.S:65: first defined here mtest: In function `_fini': (.fini+0x0): multiple definition of `_fini' /usr/lib64/gcc/x86_64-suse-linux/4.3/../../../../lib64/crti.o:initfini.c:(. fini+0x0): first defined here mtest:(.rodata+0x0): multiple definition of `_IO_stdin_used' /usr/lib64/gcc/x86_64-suse-linux/4.3/../../../../lib64/crt1.o:(.rodata.cst4 +0x0): first defined here mtest: In function `__data_start': (.data+0x0): multiple definition of `__data_start' /usr/lib64/gcc/x86_64-suse-linux/4.3/../../../../lib64/crt1.o:(.data+0x0): first defined here mtest: In function `main': (.text+0xec): multiple definition of `main' /tmp/ccUyk8e9.o:mtest.c:(.text+0x0): first defined here mtest: In function `_init': (.init+0x0): multiple definition of `_init' /usr/lib64/gcc/x86_64-suse-linux/4.3/../../../../lib64/crti.o:initfini.c:(. init+0x0): first defined here /usr/lib64/gcc/x86_64-suse-linux/4.3/../../../../x86_64-suse-linux/bin/ld: error in mtest(.eh_frame); no .eh_frame_hdr table will be created. collect2: ld returned 1 exit status Not sure if this makes any sense? Best wishes Kristian Dr. Kristian Unger Arbeitsgruppenleiter Integrative Biologie / Head of Integrative Biology Group Abteilung für Strahlenzytogenetik / Research Unit of Radiation Cytogenetics Tel.: +49-89-3187-3515 Mob.: +49-160-90641879 Am 15.06.11 17:25 schrieb Hugo Mildenberger unter hugo.mildenber...@web.de: Kristian, these are the usual problems with binary distributions. Regarding --with-Rmpi-include=/usr/lib64/mpi/gcc/openmpi/include and configure output checking for mpi.h... no ... so does /usr/lib64/mpi/gcc/openmpi/include really exist? At least, that appears to be a very unusual place to look for mpi.h (normally to be found in /usr/include ). And if you try to compile link the attached demo program: does the link phase succeed? Compile link using $ mpicc mtest.c -o mtest Presumably you have already tried to run install.packages(Rmpi). Kind regards Hugo On Wednesday 15 June 2011 16:22:07 Unger, Kristian, Dr. wrote: Thanks Hugo. I am pretty sure openmpi is installed: # zypper se openmpi Loading repository data... Reading installed packages... S | Name | Summary | Type --+---+-+--- i | openmpi | A powerful implementaion of MPI | package | openmpi | A powerful implementaion of MPI | srcpackage i | openmpi-devel | A powerful implementaion of MPI | package I got the same error message with the latest version available. The reason why I took the somewhat older version is that I wanted to make sure that it is not related to any libraries used by the newest version. Best wishes Kristian Dr. Kristian Unger Arbeitsgruppenleiter Integrative Biologie / Head of Integrative Biology Group Abteilung für Strahlenzytogenetik / Research Unit of Radiation Cytogenetics Tel.: +49-89-3187-3515 Mob.: +49-160-90641879 Am 15.06.11 16:16 schrieb Hugo Mildenberger unter hugo.mildenber...@web.de: Kristian, I just tried that particular command here on
Re: [R] Rmpi installation
Hi Hugo One step closer! I added the path to libmpi to ld.so.conf and ran ldconfig (see http://www.linuxforums.org/forum/programming-scripting/80405-linking-shared -libraries.html). Compiling the code you sent me works but running the file gets stuck with the following message: ./mtest librdmacm: couldn't read ABI version. librdmacm: assuming: 4 libibverbs: Fatal: couldn't read uverbs ABI version. CMA: unable to open /dev/infiniband/rdma_cm -- WARNING: Failed to open OpenIB-cma [DAT_INTERNAL_ERROR:]. This may be a real error or it may be an invalid entry in the uDAPL Registry which is contained in the dat.conf file. Contact your local System Administrator to confirm the availability of the interfaces in the dat.conf file. -- Any suggestions? Best wishes Kristian Dr. Kristian Unger Arbeitsgruppenleiter Integrative Biologie / Head of Integrative Biology Group Abteilung für Strahlenzytogenetik / Research Unit of Radiation Cytogenetics Tel.: +49-89-3187-3515 Mob.: +49-160-90641879 Am 15.06.11 19:25 schrieb Hugo Mildenberger unter hugo.mildenber...@web.de: Dear Kristian, please run exactly mpicc mtest.c -o mtest If you really need it, add -v separately. mpicc is nothing but a compiler wrapper. The -o switch specifies the outfile name, which has to follow immediately after -o, with or without a blank character in between. I'm not sure about what happens with -ov mtest. Probably an output file named v is produced on disk, in addition to the file mtest stemming from a previous with -o mtest Depending on mpicc's, gcc's and ld's version, gcc may instruct the linker to link the previously produced executable mtest with the object newly compiled from mtest.c plus several implicit libraries and startup objects, and output the result to v. ld would then be faced with duplicate defined symbols and output a list of these doublets. My proposal was meant as a test for an else correct openmpi installation, perhaps I should have said this. On successful completion, no message should be printed, but an executable named mtest should have been produced, which could then be run typing ./mtest. If the link process failed, then the openmpi installation was inconsistent. Kind regards Hugo On Wednesday 15 June 2011 17:45:41 Unger, Kristian, Dr. wrote: Dear Hugo I ran the command with the verbose switch and get the following output: mpicc mtest.c -ov mtest mtest: In function `_start': /usr/src/packages/BUILD/glibc-2.11.1/csu/../sysdeps/x86_64/elf/start.S:65 : multiple definition of `_start' /usr/lib64/gcc/x86_64-suse-linux/4.3/../../../../lib64/crt1.o:/usr/src/pa ck ages/BUILD/glibc-2.11.1/csu/../sysdeps/x86_64/elf/start.S:65: first defined here mtest: In function `_fini': (.fini+0x0): multiple definition of `_fini' /usr/lib64/gcc/x86_64-suse-linux/4.3/../../../../lib64/crti.o:initfini.c: (. fini+0x0): first defined here mtest:(.rodata+0x0): multiple definition of `_IO_stdin_used' /usr/lib64/gcc/x86_64-suse-linux/4.3/../../../../lib64/crt1.o:(.rodata.cs t4 +0x0): first defined here mtest: In function `__data_start': (.data+0x0): multiple definition of `__data_start' /usr/lib64/gcc/x86_64-suse-linux/4.3/../../../../lib64/crt1.o:(.data+0x0) : first defined here mtest: In function `main': (.text+0xec): multiple definition of `main' /tmp/ccUyk8e9.o:mtest.c:(.text+0x0): first defined here mtest: In function `_init': (.init+0x0): multiple definition of `_init' /usr/lib64/gcc/x86_64-suse-linux/4.3/../../../../lib64/crti.o:initfini.c: (. init+0x0): first defined here /usr/lib64/gcc/x86_64-suse-linux/4.3/../../../../x86_64-suse-linux/bin/ld : error in mtest(.eh_frame); no .eh_frame_hdr table will be created. collect2: ld returned 1 exit status Not sure if this makes any sense? Best wishes Kristian Dr. Kristian Unger Arbeitsgruppenleiter Integrative Biologie / Head of Integrative Biology Group Abteilung für Strahlenzytogenetik / Research Unit of Radiation Cytogenetics Tel.: +49-89-3187-3515 Mob.: +49-160-90641879 Am 15.06.11 17:25 schrieb Hugo Mildenberger unter hugo.mildenber...@web.de: Kristian, these are the usual problems with binary distributions. Regarding --with-Rmpi-include=/usr/lib64/mpi/gcc/openmpi/include and configure output checking for mpi.h... no ... so does /usr/lib64/mpi/gcc/openmpi/include really exist? At least, that appears to be a very unusual place to look for mpi.h (normally to be found in /usr/include ). And if you try to compile link the attached demo program: does the link phase succeed? Compile link using $ mpicc mtest.c -o mtest Presumably you have already tried to run install.packages(Rmpi). Kind regards
[R] Rmpi installation problems on slave nodes
Hi guys, I'm trying to install Rmpi on the slave nodes of my cluster. The installation seems to work with no errors or warnings using this command: R CMD INSTALL --configure-args=--with-Rmpi-include=/usr/lib64/openmpi/1.3.2-gcc/include --with-Rmpi-libpath=/usr/lib64/openmpi/1.3.2-gcc/lib --with-Rmpi-type=OPENMPI Rmpi_0.5-8.tar.gz But when I attempt to load Rmpi I get the following error: library(Rmpi) Error in dyn.load(file, DLLpath = DLLpath, ...) : unable to load shared library '/usr/lib64/R/library/Rmpi/libs/Rmpi.so': libmpi.so.0: cannot open shared object file: No such file or directory Error in library(Rmpi) : .First.lib failed for 'Rmpi' Error in dyn.unload(file.path(libpath, libs, paste(Rmpi, .Platform$dynlib.ext, : dynamic/shared library '/usr/lib64/R/library/Rmpi/libs/Rmpi.so' was not loaded I wonder does anybody have an ideas as to what might be wrong here? It seemed to install just fine on my master node by the way. Thanks, Paul. -- Paul Geeleher School of Mathematics, Statistics and Applied Mathematics National University of Ireland Galway Ireland -- www.bioinformaticstutorials.com __ R-help@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.