在LAM / MPI群集上安裝Rmpi

Question

我正在嘗試在LAM MPI集群機器上安裝Rmpi軟件包。 以前我一直在編譯和測試一些東西（ mpi4py和小型C ++程序），所以我確信MPI本身mpi4py 。 但是，鏈接庫時安裝Rmpi包失敗。

我的主要嫌疑人是在makefile中調用gcc而不是mpicc （我試圖找到配置中的行來改變它，但到目前為止找不到它）。

有人有在LAM上安裝Rmpi的經驗，你是如何管理的？

建築

• LAM MPI（如果存在，也可能是PBS MPI，如何檢查？）。 有一件事是肯定的，我有mpicpp文件，我認為並不是所有的mpi實現都很常見
• R-2.15.2
• Rmpi_0.6-1
• CentOS版本5.5（最終版），x86_64

從R調用的命令

install.packages("Rmpi")

日志

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 âRmpiâ ...
** package âRmpiâ 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 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
** libs
gcc -std=gnu99 -I/g/software/linux/pack/r-2.15.2/centos-5/lib64/R/include -DNDEBUG -DPACKAGE_NAME=\"\" -DPACKAGE_TARNAME=\"\" -DPACKAGE_VERSION=\"\" -DPACKAGE_STRING=\"\" -DPACKAGE_BUGREPORT=\"\" -DPACKAGE_URL=\"\" -I/usr/include  -DMPI2 -DOPENMPI -I/usr/local/include    -fpic  -g -O2  -c RegQuery.c -o RegQuery.o
gcc -std=gnu99 -I/g/software/linux/pack/r-2.15.2/centos-5/lib64/R/include -DNDEBUG -DPACKAGE_NAME=\"\" -DPACKAGE_TARNAME=\"\" -DPACKAGE_VERSION=\"\" -DPACKAGE_STRING=\"\" -DPACKAGE_BUGREPORT=\"\" -DPACKAGE_URL=\"\" -I/usr/include  -DMPI2 -DOPENMPI -I/usr/local/include    -fpic  -g -O2  -c Rmpi.c -o Rmpi.o
gcc -std=gnu99 -I/g/software/linux/pack/r-2.15.2/centos-5/lib64/R/include -DNDEBUG -DPACKAGE_NAME=\"\" -DPACKAGE_TARNAME=\"\" -DPACKAGE_VERSION=\"\" -DPACKAGE_STRING=\"\" -DPACKAGE_BUGREPORT=\"\" -DPACKAGE_URL=\"\" -I/usr/include  -DMPI2 -DOPENMPI -I/usr/local/include    -fpic  -g -O2  -c conversion.c -o conversion.o
gcc -std=gnu99 -I/g/software/linux/pack/r-2.15.2/centos-5/lib64/R/include -DNDEBUG -DPACKAGE_NAME=\"\" -DPACKAGE_TARNAME=\"\" -DPACKAGE_VERSION=\"\" -DPACKAGE_STRING=\"\" -DPACKAGE_BUGREPORT=\"\" -DPACKAGE_URL=\"\" -I/usr/include  -DMPI2 -DOPENMPI -I/usr/local/include    -fpic  -g -O2  -c internal.c -o internal.o
gcc -std=gnu99 -shared -L/usr/local/lib64 -o Rmpi.so RegQuery.o Rmpi.o conversion.o internal.o -L/usr/lib -lmpi -lutil -lpthread -L/g/software/linux/pack/r-2.15.2/centos-5/lib64/R/lib -lR
/usr/bin/ld: skipping incompatible /usr/lib/libutil.so when searching for -lutil
/usr/bin/ld: skipping incompatible /usr/lib/libutil.a when searching for -lutil
/usr/bin/ld: skipping incompatible /usr/lib/libpthread.so when searching for -lpthread
/usr/bin/ld: skipping incompatible /usr/lib/libpthread.a when searching for -lpthread
/usr/bin/ld: skipping incompatible /usr/lib/libc.so when searching for -lc
/usr/bin/ld: skipping incompatible /usr/lib/libc.a when searching for -lc
installing to /home/andrejev/R/x86_64-unknown-linux-gnu-library/2.15/Rmpi/libs
** R
** demo
** inst
** preparing package for lazy loading
** help
*** installing help indices
** building package indices
** testing if installed package can be loaded
Error : .onLoad failed in loadNamespace() for 'Rmpi', details:
  call: dyn.load(file, DLLpath = DLLpath, ...)
  error: unable to load shared object '/home/andrejev/R/x86_64-unknown-linux-gnu-library/2.15/Rmpi/libs/Rmpi.so':
  /usr/lib/libmpi.so.0: undefined symbol: _kio
Error: loading failed
Execution halted
ERROR: loading failed
* removing â/home/andrejev/R/x86_64-unknown-linux-gnu-library/2.15/Rmpiâ

The downloaded source packages are in
        â/tmp/Rtmp5qzeqp/downloaded_packagesâ
Warning message:
In install.packages("Rmpi") :
  installation of package âRmpiâ had non-zero exit status

Answer 1

我們有一個非常相似的設置。 我最近沒有觸及它以便能夠說出你的具體情況，但我可以傳遞一些使我們正常運行的命令......

正如Dirk所說，我們的第一步是讓OpenMPI運行。 完成設置后，您需要確保正確設置了LD_LIBRARY_PATH ：

export LD_LIBRARY_PATH=/opt/openmpiv2/lib/:$LD_LIBRARY_PATH

最終為我們安裝了Rmpi的命令（在運行上述命令並啟動R之后）是：

install.packages("Rmpi", configure.args = c("--with-Rmpi-include=/opt/openmpiv2/include/", "--with-Rmpi-libpath=/opt/openmpiv2/lib/", "--with-Rmpi-type=OPENMPI", "--with-mpi=/opt/openmpiv2/"))

哪個為我們成功安裝了包。 然后我們可以使用以下命令運行作業：

qsub -I -l nodes=2:ppn=12
/opt/openmpiv2/bin/mpirun --hostfile $PBS_NODEFILE -n 1 R --slave -f rmpitest.R

這些命令可能特定於我們的設置，它們可能沒用，但希望這里的東西可以節省你一些時間！

正如@Sergej在下面提出的，更通用的解決方案可能是：

install.packages(
  "Rmpi", 
  configure.args = c(
   "--with-Rmpi-include=/usr/include/", # This is where LAM's mpi.h is located
   "--with-Rmpi-libpath=/usr/lib/",     # This is where liblam.so is located (actually as I type it mine was located in /usr/lib64/liblam.so.0, so maybe this is not needed at all)
   "--with-Rmpi-type=LAM"               # This says that the type is LAM (there is also OPENMPI and MPICH)
))

我們的設置：RHEL 5，Rmpi 0.5-9，OpenMPI 1.4.3。，R 2.15.0

Answer 2

很多年前，我曾經使用過LAM，但之后轉而使用OpenMPI。 Rmpi的作者郝宇建議做同樣的事情。

r-sig-hpc列表對於這些問題來說是個不錯的選擇。

Answer 3

正如傑夫艾倫所說，我去了R並跑了

install.packages(
  "Rmpi", 
  configure.args = c(
   "--with-Rmpi-include=/usr/include/", # This is where LAM's mpi.h is located
   "--with-Rmpi-libpath=/usr/lib64/",     # This is where liblam.so is located
   "--with-Rmpi-type=LAM"               # This says that the type is LAM (there is also OPENMPI and MPICH)
))

我也跑

# Because it was non-standard R location so it din't find shared libraries
export LD_LIBRARY_PATH=/g/software/linux/pack/r-2.15.0/lib64/R/lib

# Rmpi also could not find libutil so I had to add path to it as well
export LD_PRELOAD=/usr/lib64/libutil.so

編輯

即使我在多個節點上運行Rmpi時設法編譯它也會失敗。 所以我想我必須開始推動OpenMPI過渡：/