I suggest you ask the maintainer (as does the posting guide).
Red Hat headers for mpi are in a different layout from the sources, and
recent versions of mpi from earlier ones. But the path to mpi.h (which I
guess is /usr/local/lam-7.1.4/include in your setup) is missing. You
might try adding it to the packages' PKG_CPPFLAGS or to C_INCLUDE_PATH.
I suspect you will then have problems with the libs. One I have is that
the lam libs built from sources need -lutil, and Rmpi and taskPR do not
allow for that.
There are a lot of other problems with this package: one is that it
includes a long-outdated (>2.5 years) copy of the private R header Defh.h
that is incompatible with the current Rinternals.h
As a result the package is basically skipped (a fake install is used) in
the CRAN checks and I am pretty sure it is not going to work with recent
versions of R without very extensive changes.
On Wed, 5 Dec 2007, Baker D.J. wrote:
> Hello,
>
> I'd like to take a look at how the taskPR package performs on our
computational cluster, and this morning I've spend some time trying to do a
test install of this package on one of our linux machines. Unfortunately
I've been unsuccessful in completing the install, and I wondered if someone
could please advise me.
>
> As I've noted I working on a linux machine, and it's running RHEL
4.0. I've build the lam package (we're very familiar with all the mpi
libraries and so that is not an issue), and then I attempted to install
taskPR...
>
> A). I added the lam binaries to my command search PATH
> B). Defined CC=mpicc
> C). Started R, and tried the install using
> install.packages(c("taskPR_0.2.7.tar.gz"), repos=NULL,
configure.args= c(taskPR="--with-mpi='/usr/local/lam-7.1.4'"))
>
> The configure/install fails, and I've appended the output below. I can
understand why it's failing, however I don't understand how to fix the
problem. Could anyone please shead any light on this issue.
>
> Thanks & Regards,
>
> David Baker.
>
>> install.packages(c("taskPR_0.2.7.tar.gz"), repos=NULL,
configure.args= c(taskPR="--with-mpi='/usr/local/lam-7.1.4'"))
> * Installing *source* package 'taskPR' ...
> creating cache ./config.cache
> checking for gcc... mpicc
> checking whether the C compiler (mpicc ) works... yes
> checking whether the C compiler (mpicc ) is a cross-compiler... no
> checking whether we are using GNU C... yes
> checking whether mpicc accepts -g... yes
> checking how to run the C preprocessor... mpicc -E
> checking for mpi.h... yes
> checking for pthread_detach in -lpthread... yes
> checking for MPI_Comm_spawn... yes
> checking whether mpicc -Wall is understood... yes
> updating cache ./config.cache
> creating ./config.status
> creating src/Makevars
> ** libs
> ** arch -
> gcc -std=gnu99 -I/usr/local/R-2.5.0/lib64/R/include
-I/usr/local/R-2.5.0/lib64/R/include -I/usr/local/include -g -O2 -Wall
-DPARALLEL_R_PACKAGE -fpic -g -O2 -c spawnhelper.c -o spawnhelper.o
> spawnhelper.c:30:17: mpi.h: No such file or directory
> spawnhelper.c: In function `main':
> spawnhelper.c:41: warning: implicit declaration of function `MPI_Init'
> spawnhelper.c:41: error: `MPI_SUCCESS' undeclared (first use in this
function)
> spawnhelper.c:41: error: (Each undeclared identifier is reported only once
> spawnhelper.c:41: error: for each function it appears in.)
> spawnhelper.c:46: warning: implicit declaration of function
`MPI_Comm_rank'
> spawnhelper.c:46: error: `MPI_COMM_WORLD' undeclared (first use in this
function)
> spawnhelper.c:47: warning: implicit declaration of function
`MPI_Comm_size'
> spawnhelper.c:49: warning: implicit declaration of function `MPI_Bcast'
> spawnhelper.c:49: error: `MPI_INT' undeclared (first use in this
function)
> spawnhelper.c:75: error: `MPI_CHAR' undeclared (first use in this
function)
> make: *** [spawnhelper.o] Error 1
> ERROR: compilation failed for package 'taskPR'
> ** Removing '/usr/local/R-2.5.0/lib64/R/library/taskPR'
>
> ______________________________________________
> R-help at 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.
>
--
Brian D. Ripley, ripley at 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, UK Fax: +44 1865 272595