Re: [firedrake] FD on mac
On Mon, 30 Nov 2015 at 18:26 Mitchell, Lawrence < lawrence.mitchell@imperial.ac.uk> wrote:
On 30 Nov 2015, at 18:01, David Ham <David.Ham@imperial.ac.uk> wrote:
This is quite possible. Unfortunately we have no access to Mavericks machines which makes it difficult to work fix this problem.
The script does currently work for me on El Capitan!
The proximate cause of Onno's problem appears to be the lack of fortran compiler. In the past, this came via the brew package "gfortran" is this still the case? Those of us for whom everything works have had it all in place for a while, no?
gfortran is currently provided by the plain gcc package. ls -l $(which gfortran) lrwxr-xr-x 1 dham admin 32 23 Jul 12:44 /usr/local/bin/gfortran -> ../Cellar/gcc/5.2.0/bin/gfortran
Lawrence _______________________________________________ firedrake mailing list firedrake@imperial.ac.uk https://mailman.ic.ac.uk/mailman/listinfo/firedrake
I get this: ls -l $(which gfortran) lrwxr-xr-x 1 root admin 32 24 Mar 2014 /usr/local/bin/gfortran -> /usr/local/gfortran/bin/gfortran gfortran works fine; hello.f90 works fine; it is the openmpi bit when I add "use mpi" use mpi 1 Fatal Error: Can't open module file 'mpi.mod' for reading at (1): No such file or directory mpif90 hello.f90 -o hello command does not work. ________________________________ From: firedrake-bounces@imperial.ac.uk <firedrake-bounces@imperial.ac.uk> on behalf of David Ham <David.Ham@imperial.ac.uk> Sent: Monday, November 30, 2015 6:43 PM To: firedrake Subject: Re: [firedrake] FD on mac On Mon, 30 Nov 2015 at 18:26 Mitchell, Lawrence <lawrence.mitchell@imperial.ac.uk<mailto:lawrence.mitchell@imperial.ac.uk>> wrote:
On 30 Nov 2015, at 18:01, David Ham <David.Ham@imperial.ac.uk<mailto:David.Ham@imperial.ac.uk>> wrote:
This is quite possible. Unfortunately we have no access to Mavericks machines which makes it difficult to work fix this problem.
The script does currently work for me on El Capitan!
The proximate cause of Onno's problem appears to be the lack of fortran compiler. In the past, this came via the brew package "gfortran" is this still the case? Those of us for whom everything works have had it all in place for a while, no? gfortran is currently provided by the plain gcc package. ls -l $(which gfortran) lrwxr-xr-x 1 dham admin 32 23 Jul 12:44 /usr/local/bin/gfortran -> ../Cellar/gcc/5.2.0/bin/gfortran Lawrence _______________________________________________ firedrake mailing list firedrake@imperial.ac.uk<mailto:firedrake@imperial.ac.uk> https://mailman.ic.ac.uk/mailman/listinfo/firedrake
Onno, that's strange. Are you sure you followed my instructions correctly? Because simply reinstalling Homebrew is not going to fix the fortran issue. You need to manually reconfigure openmpi to use the latest gcc compilers. First ensure you have the latest gcc via Homebrew (I think it should be gcc-5). Then you "install" openmpi initially (via brew install openmpi). Finally you need to rebuild openmpi via brew reinstall openmpi --build-from-source Because by default, the openmpi links to the apple compilers, and for some reason mpif90 is not hooked up to any gfortran. Hence you need those environment variables I stated. They do something similar here http://stackoverflow.com/questions/26812780/linking-homebrew-compiled-openmp... This should fix that fortran problem within the PETSc build process. Spatialindex needs to be rebuilt the same way. Thanks, Justin On Monday, November 30, 2015, Onno Bokhove <O.Bokhove@leeds.ac.uk> wrote:
I get this:
ls -l $(which gfortran)
lrwxr-xr-x 1 root admin 32 24 Mar 2014 /usr/local/bin/gfortran -> /usr/local/gfortran/bin/gfortran
gfortran works fine; hello.f90 works fine; it is the openmpi bit when I add "use mpi"
use mpi
1
Fatal Error: Can't open module file 'mpi.mod' for reading at (1): No such file or directory
mpif90 hello.f90 -o hello
command does not work.
------------------------------ *From:* firedrake-bounces@imperial.ac.uk <javascript:_e(%7B%7D,'cvml','firedrake-bounces@imperial.ac.uk');> < firedrake-bounces@imperial.ac.uk <javascript:_e(%7B%7D,'cvml','firedrake-bounces@imperial.ac.uk');>> on behalf of David Ham <David.Ham@imperial.ac.uk <javascript:_e(%7B%7D,'cvml','David.Ham@imperial.ac.uk');>> *Sent:* Monday, November 30, 2015 6:43 PM *To:* firedrake *Subject:* Re: [firedrake] FD on mac
On Mon, 30 Nov 2015 at 18:26 Mitchell, Lawrence < lawrence.mitchell@imperial.ac.uk <javascript:_e(%7B%7D,'cvml','lawrence.mitchell@imperial.ac.uk');>> wrote:
On 30 Nov 2015, at 18:01, David Ham <David.Ham@imperial.ac.uk <javascript:_e(%7B%7D,'cvml','David.Ham@imperial.ac.uk');>> wrote:
This is quite possible. Unfortunately we have no access to Mavericks machines which makes it difficult to work fix this problem.
The script does currently work for me on El Capitan!
The proximate cause of Onno's problem appears to be the lack of fortran compiler. In the past, this came via the brew package "gfortran" is this still the case? Those of us for whom everything works have had it all in place for a while, no?
gfortran is currently provided by the plain gcc package.
ls -l $(which gfortran)
lrwxr-xr-x 1 dham admin 32 23 Jul 12:44 /usr/local/bin/gfortran -> ../Cellar/gcc/5.2.0/bin/gfortran
Lawrence _______________________________________________ firedrake mailing list firedrake@imperial.ac.uk <javascript:_e(%7B%7D,'cvml','firedrake@imperial.ac.uk');> https://mailman.ic.ac.uk/mailman/listinfo/firedrake
Hi Justin, On 30/11/15 23:59, Justin Chang wrote:
Onno, that's strange. Are you sure you followed my instructions correctly? Because simply reinstalling Homebrew is not going to fix the fortran issue. You need to manually reconfigure openmpi to use the latest gcc compilers. First ensure you have the latest gcc via Homebrew (I think it should be gcc-5). Then you "install" openmpi initially (via brew install openmpi). Finally you need to rebuild openmpi via
brew reinstall openmpi --build-from-source
Because by default, the openmpi links to the apple compilers, and for some reason mpif90 is not hooked up to any gfortran. Hence you need those environment variables I stated.
They do something similar here http://stackoverflow.com/questions/26812780/linking-homebrew-compiled-openmp...
This should fix that fortran problem within the PETSc build process.
Spatialindex needs to be rebuilt the same way.
It shouldn't be necessary to rebuild openmpi yourself unless you want to use the GNU compilers everywhere. Clang (+ gfortran) works fine as far as I can tell. If you do that, you then won't need to rebuild spatialindex either. Cheers, Lawrence
The issue was that homebrew's openmpi 1.10.1 does not link to any gfortran compiler. I rebuilt openmpi primarily to include the gcc fortran compiler. Though I think the cleaner way would be to modify the mpif90-wrapper-data.txt to ensure that the compiler points to a working gfortran. On Tuesday, December 1, 2015, Lawrence Mitchell < lawrence.mitchell@imperial.ac.uk> wrote:
Hi Justin,
On 30/11/15 23:59, Justin Chang wrote:
Onno, that's strange. Are you sure you followed my instructions correctly? Because simply reinstalling Homebrew is not going to fix the fortran issue. You need to manually reconfigure openmpi to use the latest gcc compilers. First ensure you have the latest gcc via Homebrew (I think it should be gcc-5). Then you "install" openmpi initially (via brew install openmpi). Finally you need to rebuild openmpi via
brew reinstall openmpi --build-from-source
Because by default, the openmpi links to the apple compilers, and for some reason mpif90 is not hooked up to any gfortran. Hence you need those environment variables I stated.
They do something similar here http://stackoverflow.com/questions/26812780/linking-homebrew-compiled-openmp...
This should fix that fortran problem within the PETSc build process.
Spatialindex needs to be rebuilt the same way.
It shouldn't be necessary to rebuild openmpi yourself unless you want to use the GNU compilers everywhere. Clang (+ gfortran) works fine as far as I can tell. If you do that, you then won't need to rebuild spatialindex either.
Cheers,
Lawrence
On 01/12/15 11:48, Justin Chang wrote:
The issue was that homebrew's openmpi 1.10.1 does not link to any gfortran compiler. I rebuilt openmpi primarily to include the gcc fortran compiler. Though I think the cleaner way would be to modify the mpif90-wrapper-data.txt to ensure that the compiler points to a working gfortran.
Oh, weird. So I'm on El Cap, and so brew installs gcc and openmpi from bottles. I did: brew uninstall gcc brew uninstall openmpi brew install openmpi brew install gcc And now mpif90 -show gives: /usr/local/bin/gfortran ... So it's all set up correctly AFAICT. Cheers, Lawrence
Yeah, that's not the case for me. I did: brew uninstall gcc brew uninstall openmpi brew install openmpi brew install gcc and here's what I have: mpicc --show clang -I/usr/local/Cellar/open-mpi/1.10.1/include -L/usr/local/opt/libevent/lib -L/usr/local/Cellar/open-mpi/1.10.1/lib -lmpi mpicxx --show clang++ -I/usr/local/Cellar/open-mpi/1.10.1/include -L/usr/local/opt/libevent/lib -L/usr/local/Cellar/open-mpi/1.10.1/lib -lmpi_cxx -lmpi mpif90 --show -------------------------------------------------------------------------- No underlying compiler was specified in the wrapper compiler data file (e.g., mpicc-wrapper-data.txt) -------------------------------------------------------------------------- On Tue, Dec 1, 2015 at 4:51 AM, Lawrence Mitchell < lawrence.mitchell@imperial.ac.uk> wrote:
On 01/12/15 11:48, Justin Chang wrote:
The issue was that homebrew's openmpi 1.10.1 does not link to any gfortran compiler. I rebuilt openmpi primarily to include the gcc fortran compiler. Though I think the cleaner way would be to modify the mpif90-wrapper-data.txt to ensure that the compiler points to a working gfortran.
Oh, weird. So I'm on El Cap, and so brew installs gcc and openmpi from bottles. I did:
brew uninstall gcc brew uninstall openmpi
brew install openmpi brew install gcc
And now mpif90 -show gives:
/usr/local/bin/gfortran ...
So it's all set up correctly AFAICT.
Cheers,
Lawrence
_______________________________________________ firedrake mailing list firedrake@imperial.ac.uk https://mailman.ic.ac.uk/mailman/listinfo/firedrake
Okay, so mpif90-wrapper-data.txt is located in /usr/local/Cellar/open-mpi/1.10.1/share/openmpi/mpif90-wrapper-data.txt and in the row that says "compiler=", the RHS is blank. Simply adding gfortran-5 now gives: mpif90 --show gfortran-5 -I/usr/local/Cellar/open-mpi/1.10.1/include -L/usr/local/opt/libevent/lib -L/usr/local/Cellar/open-mpi/1.10.1/lib -lmpi_mpifh -lmpi so this is definitely the much better alternative On Tue, Dec 1, 2015 at 4:58 AM, Justin Chang <jychang48@gmail.com> wrote:
Yeah, that's not the case for me. I did:
brew uninstall gcc brew uninstall openmpi
brew install openmpi brew install gcc
and here's what I have:
mpicc --show
clang -I/usr/local/Cellar/open-mpi/1.10.1/include -L/usr/local/opt/libevent/lib -L/usr/local/Cellar/open-mpi/1.10.1/lib -lmpi
mpicxx --show
clang++ -I/usr/local/Cellar/open-mpi/1.10.1/include -L/usr/local/opt/libevent/lib -L/usr/local/Cellar/open-mpi/1.10.1/lib -lmpi_cxx -lmpi
mpif90 --show
--------------------------------------------------------------------------
No underlying compiler was specified in the wrapper compiler data file
(e.g., mpicc-wrapper-data.txt)
--------------------------------------------------------------------------
On Tue, Dec 1, 2015 at 4:51 AM, Lawrence Mitchell < lawrence.mitchell@imperial.ac.uk> wrote:
On 01/12/15 11:48, Justin Chang wrote:
The issue was that homebrew's openmpi 1.10.1 does not link to any gfortran compiler. I rebuilt openmpi primarily to include the gcc fortran compiler. Though I think the cleaner way would be to modify the mpif90-wrapper-data.txt to ensure that the compiler points to a working gfortran.
Oh, weird. So I'm on El Cap, and so brew installs gcc and openmpi from bottles. I did:
brew uninstall gcc brew uninstall openmpi
brew install openmpi brew install gcc
And now mpif90 -show gives:
/usr/local/bin/gfortran ...
So it's all set up correctly AFAICT.
Cheers,
Lawrence
_______________________________________________ firedrake mailing list firedrake@imperial.ac.uk https://mailman.ic.ac.uk/mailman/listinfo/firedrake
On 1 Dec 2015, at 12:05, Justin Chang <jychang48@gmail.com> wrote:
Okay, so mpif90-wrapper-data.txt is located in
/usr/local/Cellar/open-mpi/1.10.1/share/openmpi/mpif90-wrapper-data.txt
and in the row that says "compiler=", the RHS is blank. Simply adding gfortran-5 now gives:
mpif90 --show gfortran-5 -I/usr/local/Cellar/open-mpi/1.10.1/include -L/usr/local/opt/libevent/lib -L/usr/local/Cellar/open-mpi/1.10.1/lib -lmpi_mpifh -lmpi
so this is definitely the much better alternative
A better way to do this is to set the environment variable OMPI_FC to the name of the fortran compiler you want mpif90 to use. Lawrence
So without any fortran environments: *➜ * *~ * brew uninstall gcc Uninstalling /usr/local/Cellar/gcc/5.2.0... (1353 files, 248M) *➜ **~ * brew uninstall openmpi Uninstalling /usr/local/Cellar/open-mpi/1.10.1... (765 files, 21M) *➜ **~ * brew install openmpi *==>** Downloading https://homebrew.bintray.com/bottles/open-mpi-1.10.1.el_capitan.bottle.tar <https://homebrew.bintray.com/bottles/open-mpi-1.10.1.el_capitan.bottle.tar>* Already downloaded: /Library/Caches/Homebrew/open-mpi-1.10.1.el_capitan.bottle.tar.gz *==>** Pouring open-mpi-1.10.1.el_capitan.bottle.tar.gz* 🍺 /usr/local/Cellar/open-mpi/1.10.1: 765 files, 21M *➜ **~ * mpif90 -show -------------------------------------------------------------------------- No underlying compiler was specified in the wrapper compiler data file (e.g., mpicc-wrapper-data.txt) -------------------------------------------------------------------------- *➜ * *~ * mpif90 -------------------------------------------------------------------------- No underlying compiler was specified in the wrapper compiler data file (e.g., mpicc-wrapper-data.txt) -------------------------------------------------------------------------- *➜ * *~ * brew install gcc *==>** Downloading https://homebrew.bintray.com/bottles/gcc-5.2.0.el_capitan.bottle.3.tar.gz <https://homebrew.bintray.com/bottles/gcc-5.2.0.el_capitan.bottle.3.tar.gz>* Already downloaded: /Library/Caches/Homebrew/gcc-5.2.0.el_capitan.bottle.3.tar.gz *==>** Pouring gcc-5.2.0.el_capitan.bottle.3.tar.gz* *==>** Caveats* GCC has been built with multilib support. Notably, OpenMP may not work: https://gcc.gnu.org/bugzilla/show_bug.cgi?id=60670 If you need OpenMP support you may want to brew reinstall gcc --without-multilib *==>** Summary* 🍺 /usr/local/Cellar/gcc/5.2.0: 1353 files, 248M *➜ **~ * mpif90 -show -------------------------------------------------------------------------- No underlying compiler was specified in the wrapper compiler data file (e.g., mpicc-wrapper-data.txt) -------------------------------------------------------------------------- *➜ * *~ * mpif90 -------------------------------------------------------------------------- No underlying compiler was specified in the wrapper compiler data file (e.g., mpicc-wrapper-data.txt) -------------------------------------------------------------------------- Now with export OMPI_FC=gfortran in ~/.profile: *➜ * *~ * brew uninstall gcc Uninstalling /usr/local/Cellar/gcc/5.2.0... (1353 files, 248M) *➜ **~ * brew uninstall openmpi Uninstalling /usr/local/Cellar/open-mpi/1.10.1... (765 files, 21M) *➜ **~ * brew install openmpi *==>** Downloading https://homebrew.bintray.com/bottles/open-mpi-1.10.1.el_capitan.bottle.tar <https://homebrew.bintray.com/bottles/open-mpi-1.10.1.el_capitan.bottle.tar>* Already downloaded: /Library/Caches/Homebrew/open-mpi-1.10.1.el_capitan.bottle.tar.gz *==>** Pouring open-mpi-1.10.1.el_capitan.bottle.tar.gz* 🍺 /usr/local/Cellar/open-mpi/1.10.1: 765 files, 21M *➜ **~ * mpif90 -show -------------------------------------------------------------------------- No underlying compiler was specified in the wrapper compiler data file (e.g., mpicc-wrapper-data.txt) -------------------------------------------------------------------------- *➜ * *~ * mpif90 -------------------------------------------------------------------------- No underlying compiler was specified in the wrapper compiler data file (e.g., mpicc-wrapper-data.txt) -------------------------------------------------------------------------- *➜ * *~ * brew install gcc *==>** Downloading https://homebrew.bintray.com/bottles/gcc-5.2.0.el_capitan.bottle.3.tar.gz <https://homebrew.bintray.com/bottles/gcc-5.2.0.el_capitan.bottle.3.tar.gz>* Already downloaded: /Library/Caches/Homebrew/gcc-5.2.0.el_capitan.bottle.3.tar.gz *==>** Pouring gcc-5.2.0.el_capitan.bottle.3.tar.gz* *==>** Caveats* GCC has been built with multilib support. Notably, OpenMP may not work: https://gcc.gnu.org/bugzilla/show_bug.cgi?id=60670 If you need OpenMP support you may want to brew reinstall gcc --without-multilib *==>** Summary* 🍺 /usr/local/Cellar/gcc/5.2.0: 1353 files, 248M *➜ **~ * mpif90 -show -------------------------------------------------------------------------- No underlying compiler was specified in the wrapper compiler data file (e.g., mpicc-wrapper-data.txt) -------------------------------------------------------------------------- *➜ * *~ * mpif90 -------------------------------------------------------------------------- No underlying compiler was specified in the wrapper compiler data file (e.g., mpicc-wrapper-data.txt) -------------------------------------------------------------------------- *➜ * *~ * vi ~/.profile *➜ **~ * . ~/.profile *➜ **~ * brew uninstall gcc Uninstalling /usr/local/Cellar/gcc/5.2.0... (1353 files, 248M) *➜ **~ * brew uninstall openmpi Uninstalling /usr/local/Cellar/open-mpi/1.10.1... (765 files, 21M) *➜ **~ * brew install openmpi *==>** Downloading https://homebrew.bintray.com/bottles/open-mpi-1.10.1.el_capitan.bottle.tar <https://homebrew.bintray.com/bottles/open-mpi-1.10.1.el_capitan.bottle.tar>* Already downloaded: /Library/Caches/Homebrew/open-mpi-1.10.1.el_capitan.bottle.tar.gz *==>** Pouring open-mpi-1.10.1.el_capitan.bottle.tar.gz* 🍺 /usr/local/Cellar/open-mpi/1.10.1: 765 files, 21M *➜ **~ * mpif90 -show gfortran -I/usr/local/Cellar/open-mpi/1.10.1/include -L/usr/local/opt/libevent/lib -L/usr/local/Cellar/open-mpi/1.10.1/lib -lmpi_mpifh -lmpi *➜ **~ * which gfortran gfortran not found *➜ * *~ * mpif90 --version -------------------------------------------------------------------------- The Open MPI wrapper compiler was unable to find the specified compiler gfortran in your PATH. Note that this compiler was either specified at configure time or in one of several possible environment variables. -------------------------------------------------------------------------- *➜ * *~ * brew install openmpi Warning: open-mpi-1.10.1 already installed *➜ **~ * brew install gcc *==>** Downloading https://homebrew.bintray.com/bottles/gcc-5.2.0.el_capitan.bottle.3.tar.gz <https://homebrew.bintray.com/bottles/gcc-5.2.0.el_capitan.bottle.3.tar.gz>* Already downloaded: /Library/Caches/Homebrew/gcc-5.2.0.el_capitan.bottle.3.tar.gz *==>** Pouring gcc-5.2.0.el_capitan.bottle.3.tar.gz* *==>** Caveats* GCC has been built with multilib support. Notably, OpenMP may not work: https://gcc.gnu.org/bugzilla/show_bug.cgi?id=60670 If you need OpenMP support you may want to brew reinstall gcc --without-multilib *==>** Summary* 🍺 /usr/local/Cellar/gcc/5.2.0: 1353 files, 248M *➜ **~ * mpif90 --version GNU Fortran (Homebrew gcc 5.2.0) 5.2.0 Copyright (C) 2015 Free Software Foundation, Inc. GNU Fortran comes with NO WARRANTY, to the extent permitted by law. You may redistribute copies of GNU Fortran under the terms of the GNU General Public License. For more information about these matters, see the file named COPYING but my firedrake-install script gives the same fortran error during the petsc configure. Now when I tried your hello world example: mpif90 hello.f90 -o hello *hello.f90:2:4:* use mpi * 1* *Fatal Error:* Can't open module file '*mpi.mod*' for reading at (1): No such file or directory compilation terminated. Clearly something is wrong, and setting the environment flags (or modifying the mpif90-wrapper-data.txt) does not resolve the issue. Do you have a gfortran environment set anywhere? Thanks, Justin On Tue, Dec 1, 2015 at 5:08 AM, Lawrence Mitchell < lawrence.mitchell@imperial.ac.uk> wrote:
On 1 Dec 2015, at 12:05, Justin Chang <jychang48@gmail.com> wrote:
Okay, so mpif90-wrapper-data.txt is located in
/usr/local/Cellar/open-mpi/1.10.1/share/openmpi/mpif90-wrapper-data.txt
and in the row that says "compiler=", the RHS is blank. Simply adding gfortran-5 now gives:
mpif90 --show gfortran-5 -I/usr/local/Cellar/open-mpi/1.10.1/include -L/usr/local/opt/libevent/lib -L/usr/local/Cellar/open-mpi/1.10.1/lib -lmpi_mpifh -lmpi
so this is definitely the much better alternative
A better way to do this is to set the environment variable OMPI_FC to the name of the fortran compiler you want mpif90 to use.
Lawrence
_______________________________________________ firedrake mailing list firedrake@imperial.ac.uk https://mailman.ic.ac.uk/mailman/listinfo/firedrake
Ah, we've found the issue. On 01/12/15 12:45, Justin Chang wrote:
https://homebrew.bintray.com/bottles/open-mpi-1.10.1.el_capitan.bottle.tar* ^^^^^^
I had 1.10.0 It turns out the latest openmpi homebrew "bottle" was built without Fortran support. https://github.com/Homebrew/homebrew/pull/46461 Hopefully they will fix this soon! Lawrence
I worked on the matter the entire day, in part with our IT man, and Anna's suggestions, seemed to have passed all hoops at some point, passing the firedrake make, but then a test code did not work. We failed. It seems indeed to be fortran mli compiler. One thing I noticed with Anna's suggestions, that some of the instructions are not required, and this was for me not clear from reading them. Our IT man is going to try an install FD on a free mac laptop (with Yosemite alas) trying to use openmpi installed from another source, as well as use anaconda. More people in our dept. want to try firedrake, on their mac laptop, so resolving this issue soon may be a good idea! TBC. -- Onno ________________________________________ From: firedrake-bounces@imperial.ac.uk <firedrake-bounces@imperial.ac.uk> on behalf of Lawrence Mitchell <lawrence.mitchell@imperial.ac.uk> Sent: Tuesday, December 1, 2015 1:04 PM To: firedrake@imperial.ac.uk Subject: Re: [firedrake] FD on mac Ah, we've found the issue. On 01/12/15 12:45, Justin Chang wrote:
https://homebrew.bintray.com/bottles/open-mpi-1.10.1.el_capitan.bottle.tar* ^^^^^^
I had 1.10.0 It turns out the latest openmpi homebrew "bottle" was built without Fortran support. https://github.com/Homebrew/homebrew/pull/46461 Hopefully they will fix this soon! Lawrence
Hi Onno,
On 1 Dec 2015, at 20:14, Onno Bokhove <O.Bokhove@leeds.ac.uk> wrote:
I worked on the matter the entire day, in part with our IT man, and Anna's suggestions, seemed to have passed all hoops at some point, passing the firedrake make, but then a test code did not work.
Please send the error message from this failure.
We failed. It seems indeed to be fortran mli compiler.
Yes, we believe that the current version of openmpi built by homebrew does not come with fortran support. This will hopefully be fixed in the next few days.
One thing I noticed with Anna's suggestions, that some of the instructions are not required, and this was for me not clear from reading them.
Suggestions for things to change, and or remove are very welcome!
Our IT man is going to try an install FD on a free mac laptop (with Yosemite alas) trying to use openmpi installed from another source, as well as use anaconda.
If using anaconda, beware that they ship with a number of Python packages that require MPI but are not typically linked against whatever system MPI is already available which can lead to more problems.
More people in our dept. want to try firedrake, on their mac laptop, so resolving this issue soon may be a good idea!
We hope very much to have Mac build testing working soon. It appears additionally that this week was a particularly bad week to try installs on a Mac due to breakages out with our control. Lawrence
Hi Lawrence, We are exactly stuck on the mpif90 issues indicated below, after I mysteriously was able to have a successful "make" of FD. But the codes did not work and Tim-IT spotted the problem: two version of python, one used by anaconda and one old version were interfering with another. So we removed the anaconda from the root, installed one in the user directory; set the path, and at that point the first petsc installing command failed again on the point remarked below. Onno ________________________________________ From: firedrake-bounces@imperial.ac.uk <firedrake-bounces@imperial.ac.uk> on behalf of Lawrence Mitchell <lawrence.mitchell@imperial.ac.uk> Sent: Tuesday, December 1, 2015 8:49 PM To: firedrake@imperial.ac.uk Subject: Re: [firedrake] FD on mac Hi Onno,
On 1 Dec 2015, at 20:14, Onno Bokhove <O.Bokhove@leeds.ac.uk> wrote:
I worked on the matter the entire day, in part with our IT man, and Anna's suggestions, seemed to have passed all hoops at some point, passing the firedrake make, but then a test code did not work.
Please send the error message from this failure.
We failed. It seems indeed to be fortran mli compiler.
Yes, we believe that the current version of openmpi built by homebrew does not come with fortran support. This will hopefully be fixed in the next few days.
One thing I noticed with Anna's suggestions, that some of the instructions are not required, and this was for me not clear from reading them.
Suggestions for things to change, and or remove are very welcome!
Our IT man is going to try an install FD on a free mac laptop (with Yosemite alas) trying to use openmpi installed from another source, as well as use anaconda.
If using anaconda, beware that they ship with a number of Python packages that require MPI but are not typically linked against whatever system MPI is already available which can lead to more problems.
More people in our dept. want to try firedrake, on their mac laptop, so resolving this issue soon may be a good idea!
We hope very much to have Mac build testing working soon. It appears additionally that this week was a particularly bad week to try installs on a Mac due to breakages out with our control. Lawrence _______________________________________________ firedrake mailing list firedrake@imperial.ac.uk https://mailman.ic.ac.uk/mailman/listinfo/firedrake
Hi Onno, (and others) On 01/12/15 21:51, Onno Bokhove wrote:
Hi Lawrence,
We are exactly stuck on the mpif90 issues indicated below, after I mysteriously was able to have a successful "make" of FD. But the codes did not work and Tim-IT spotted the problem: two version of python, one used by anaconda and one old version were interfering with another.
So we removed the anaconda from the root, installed one in the user directory; set the path, and at that point the first petsc installing command failed again on the point remarked below.
OK, so the issue is absolutely that the homebrew provided openmpi binary package has not been built correctly. A workaround until it is fixed upstream is the following: # Remove any existing openmpi package brew uninstall openmpi # Ensure the gcc package (including gfortran) is available brew install gcc # Build openmpi from source (this picks up the fortran compiler # and provides the correct set of wrappers) brew install --with-fortran --build-from-source openmpi On my machine, I immediately see something like the following: ==> Using Homebrew-provided fortran compiler This may be changed by setting the FC environment variable Once openmpi has been installed, the firedrake install script /should/ work: it works locally for me after this sequence of steps. Please try it and let us know. Cheers, Lawrence
Dear all, In my case, the installation seemed to work, but when I try to run "from firedrake import *" I get the following error: Traceback (most recent call last): File "<stdin>", line 1, in <module> File "/Users/mmfg/anaconda/lib/python2.7/site-packages/firedrake/__init__.py", line 31, in <module> from firedrake.assemble import * File "/Users/mmfg/anaconda/lib/python2.7/site-packages/firedrake/assemble.py", line 10, in <module> from firedrake import assemble_expressions File "/Users/mmfg/anaconda/lib/python2.7/site-packages/firedrake/assemble_expressions.py", line 17, in <module> from firedrake import function File "/Users/mmfg/anaconda/lib/python2.7/site-packages/firedrake/function.py", line 14, in <module> from firedrake import functionspace File "/Users/mmfg/anaconda/lib/python2.7/site-packages/firedrake/functionspace.py", line 17, in <module> import firedrake.mesh as mesh_t File "/Users/mmfg/anaconda/lib/python2.7/site-packages/firedrake/mesh.py", line 16, in <module> import firedrake.spatialindex as spatialindex ImportError: dlopen(/Users/mmfg/anaconda/lib/python2.7/site-packages/firedrake/spatialindex.so, 2): Symbol not found: __ZN5Tools11PropertySet11setPropertyESsRKNS_7VariantE Referenced from: /Users/mmfg/anaconda/lib/python2.7/site-packages/firedrake/spatialindex.so Expected in: dynamic lookup So I tried to run "import firedrake.spatialindex as spatial index" and I get this error: Traceback (most recent call last): File "<stdin>", line 1, in <module> File "/Users/mmfg/anaconda/lib/python2.7/site-packages/firedrake/__init__.py", line 3, in <module> import firedrake.petsc as petsc AttributeError: 'module' object has no attribute 'petsc' Do you know where it could come from? Thank you, Floriane _______________________________________ De : firedrake-bounces@imperial.ac.uk <firedrake-bounces@imperial.ac.uk> de la part de Lawrence Mitchell <lawrence.mitchell@imperial.ac.uk> Envoyé : mercredi 2 décembre 2015 09:27 À : firedrake@imperial.ac.uk Objet : Re: [firedrake] FD on mac Hi Onno, (and others) On 01/12/15 21:51, Onno Bokhove wrote:
Hi Lawrence,
We are exactly stuck on the mpif90 issues indicated below, after I mysteriously was able to have a successful "make" of FD. But the codes did not work and Tim-IT spotted the problem: two version of python, one used by anaconda and one old version were interfering with another.
So we removed the anaconda from the root, installed one in the user directory; set the path, and at that point the first petsc installing command failed again on the point remarked below.
OK, so the issue is absolutely that the homebrew provided openmpi binary package has not been built correctly. A workaround until it is fixed upstream is the following: # Remove any existing openmpi package brew uninstall openmpi # Ensure the gcc package (including gfortran) is available brew install gcc # Build openmpi from source (this picks up the fortran compiler # and provides the correct set of wrappers) brew install --with-fortran --build-from-source openmpi On my machine, I immediately see something like the following: ==> Using Homebrew-provided fortran compiler This may be changed by setting the FC environment variable Once openmpi has been installed, the firedrake install script /should/ work: it works locally for me after this sequence of steps. Please try it and let us know. Cheers, Lawrence
On 02/12/15 12:09, Floriane Gidel [RPG] wrote:
Dear all,
In my case, the installation seemed to work, but when I try to run "from firedrake import *" I get the following error:
Traceback (most recent call last): File "<stdin>", line 1, in <module> File "/Users/mmfg/anaconda/lib/python2.7/site-packages/firedrake/__init__.py", line 31, in <module> from firedrake.assemble import * File "/Users/mmfg/anaconda/lib/python2.7/site-packages/firedrake/assemble.py", line 10, in <module> from firedrake import assemble_expressions File "/Users/mmfg/anaconda/lib/python2.7/site-packages/firedrake/assemble_expressions.py", line 17, in <module> from firedrake import function File "/Users/mmfg/anaconda/lib/python2.7/site-packages/firedrake/function.py", line 14, in <module> from firedrake import functionspace File "/Users/mmfg/anaconda/lib/python2.7/site-packages/firedrake/functionspace.py", line 17, in <module> import firedrake.mesh as mesh_t File "/Users/mmfg/anaconda/lib/python2.7/site-packages/firedrake/mesh.py", line 16, in <module> import firedrake.spatialindex as spatialindex ImportError: dlopen(/Users/mmfg/anaconda/lib/python2.7/site-packages/firedrake/spatialindex.so, 2): Symbol not found: __ZN5Tools11PropertySet11setPropertyESsRKNS_7VariantE Referenced from: /Users/mmfg/anaconda/lib/python2.7/site-packages/firedrake/spatialindex.so Expected in: dynamic lookup
This appears because the C++ compiler used to build the firedrake spatialindex extension was not the same as used to build the installed libspatialindex library I think. I presume you installed the latter using homebrew. What is the output of: brew info spatialindex I don't know where you got MPI from, but what do the following produce: mpicc -show mpicc --version mpicxx -show mpicxx --version
So I tried to run "import firedrake.spatialindex as spatial index" and I get this error:
Traceback (most recent call last): File "<stdin>", line 1, in <module> File "/Users/mmfg/anaconda/lib/python2.7/site-packages/firedrake/__init__.py", line 3, in <module> import firedrake.petsc as petsc AttributeError: 'module' object has no attribute 'petsc'
This broke because the previous import failed. Thanks, Lawrence
Dear Lawrence, Here is what I get when trying the commands you suggested: campus-10-41-49-255:~ mmfg$ mpicc -show clang -I/usr/local/Cellar/open-mpi/1.8.4/include -L/usr/local/opt/libevent/lib -L/usr/local/Cellar/open-mpi/1.8.4/lib -lmpi campus-10-41-49-255:~ mmfg$ mpicc --version Apple LLVM version 5.1 (clang-503.0.40) (based on LLVM 3.4svn) Target: x86_64-apple-darwin13.4.0 Thread model: posix campus-10-41-49-255:~ mmfg$ mpicxx -show clang++ -I/usr/local/Cellar/open-mpi/1.8.4/include -L/usr/local/opt/libevent/lib -L/usr/local/Cellar/open-mpi/1.8.4/lib -lmpi_cxx -lmpi campus-10-41-49-255:~ mmfg$ mpicxx --version Apple LLVM version 5.1 (clang-503.0.40) (based on LLVM 3.4svn) Target: x86_64-apple-darwin13.4.0 Thread model: posix Thanks, Floriane ________________________________________ De : firedrake-bounces@imperial.ac.uk <firedrake-bounces@imperial.ac.uk> de la part de Lawrence Mitchell <lawrence.mitchell@imperial.ac.uk> Envoyé : mercredi 2 décembre 2015 13:26 À : firedrake@imperial.ac.uk Objet : Re: [firedrake] FD on mac On 02/12/15 12:09, Floriane Gidel [RPG] wrote:
Dear all,
In my case, the installation seemed to work, but when I try to run "from firedrake import *" I get the following error:
Traceback (most recent call last): File "<stdin>", line 1, in <module> File "/Users/mmfg/anaconda/lib/python2.7/site-packages/firedrake/__init__.py", line 31, in <module> from firedrake.assemble import * File "/Users/mmfg/anaconda/lib/python2.7/site-packages/firedrake/assemble.py", line 10, in <module> from firedrake import assemble_expressions File "/Users/mmfg/anaconda/lib/python2.7/site-packages/firedrake/assemble_expressions.py", line 17, in <module> from firedrake import function File "/Users/mmfg/anaconda/lib/python2.7/site-packages/firedrake/function.py", line 14, in <module> from firedrake import functionspace File "/Users/mmfg/anaconda/lib/python2.7/site-packages/firedrake/functionspace.py", line 17, in <module> import firedrake.mesh as mesh_t File "/Users/mmfg/anaconda/lib/python2.7/site-packages/firedrake/mesh.py", line 16, in <module> import firedrake.spatialindex as spatialindex ImportError: dlopen(/Users/mmfg/anaconda/lib/python2.7/site-packages/firedrake/spatialindex.so, 2): Symbol not found: __ZN5Tools11PropertySet11setPropertyESsRKNS_7VariantE Referenced from: /Users/mmfg/anaconda/lib/python2.7/site-packages/firedrake/spatialindex.so Expected in: dynamic lookup
This appears because the C++ compiler used to build the firedrake spatialindex extension was not the same as used to build the installed libspatialindex library I think. I presume you installed the latter using homebrew. What is the output of: brew info spatialindex I don't know where you got MPI from, but what do the following produce: mpicc -show mpicc --version mpicxx -show mpicxx --version
So I tried to run "import firedrake.spatialindex as spatial index" and I get this error:
Traceback (most recent call last): File "<stdin>", line 1, in <module> File "/Users/mmfg/anaconda/lib/python2.7/site-packages/firedrake/__init__.py", line 3, in <module> import firedrake.petsc as petsc AttributeError: 'module' object has no attribute 'petsc'
This broke because the previous import failed. Thanks, Lawrence
Dear Floriane,
On 2 Dec 2015, at 15:35, Floriane Gidel [RPG] <mmfg@leeds.ac.uk> wrote:
Dear Lawrence,
Here is what I get when trying the commands you suggested:
campus-10-41-49-255:~ mmfg$ mpicc -show clang -I/usr/local/Cellar/open-mpi/1.8.4/include -L/usr/local/opt/libevent/lib -L/usr/local/Cellar/open-mpi/1.8.4/lib -lmpi
campus-10-41-49-255:~ mmfg$ mpicc --version Apple LLVM version 5.1 (clang-503.0.40) (based on LLVM 3.4svn) Target: x86_64-apple-darwin13.4.0 Thread model: posix
campus-10-41-49-255:~ mmfg$ mpicxx -show clang++ -I/usr/local/Cellar/open-mpi/1.8.4/include -L/usr/local/opt/libevent/lib -L/usr/local/Cellar/open-mpi/1.8.4/lib -lmpi_cxx -lmpi
campus-10-41-49-255:~ mmfg$ mpicxx --version Apple LLVM version 5.1 (clang-503.0.40) (based on LLVM 3.4svn) Target: x86_64-apple-darwin13.4.0 Thread model: posix
OK, so my hypothesis for what was going wrong doesn't appear to be the case. Let's try and fix this tomorrow when we're in the same room. Cheers, Lawrence
Ok, thanks a lot Lawrence. Floriane ________________________________________ De : firedrake-bounces@imperial.ac.uk <firedrake-bounces@imperial.ac.uk> de la part de Lawrence Mitchell <lawrence.mitchell@imperial.ac.uk> Envoyé : jeudi 3 décembre 2015 10:11 À : firedrake@imperial.ac.uk Objet : Re: [firedrake] FD on mac Dear Floriane,
On 2 Dec 2015, at 15:35, Floriane Gidel [RPG] <mmfg@leeds.ac.uk> wrote:
Dear Lawrence,
Here is what I get when trying the commands you suggested:
campus-10-41-49-255:~ mmfg$ mpicc -show clang -I/usr/local/Cellar/open-mpi/1.8.4/include -L/usr/local/opt/libevent/lib -L/usr/local/Cellar/open-mpi/1.8.4/lib -lmpi
campus-10-41-49-255:~ mmfg$ mpicc --version Apple LLVM version 5.1 (clang-503.0.40) (based on LLVM 3.4svn) Target: x86_64-apple-darwin13.4.0 Thread model: posix
campus-10-41-49-255:~ mmfg$ mpicxx -show clang++ -I/usr/local/Cellar/open-mpi/1.8.4/include -L/usr/local/opt/libevent/lib -L/usr/local/Cellar/open-mpi/1.8.4/lib -lmpi_cxx -lmpi
campus-10-41-49-255:~ mmfg$ mpicxx --version Apple LLVM version 5.1 (clang-503.0.40) (based on LLVM 3.4svn) Target: x86_64-apple-darwin13.4.0 Thread model: posix
OK, so my hypothesis for what was going wrong doesn't appear to be the case. Let's try and fix this tomorrow when we're in the same room. Cheers, Lawrence
Have el-capitan installed now so t'row at IC we'll see whether the script works! Off on the train. ________________________________________ From: firedrake-bounces@imperial.ac.uk <firedrake-bounces@imperial.ac.uk> on behalf of Floriane Gidel [RPG] <mmfg@leeds.ac.uk> Sent: Thursday, December 3, 2015 2:17 PM To: firedrake@imperial.ac.uk Subject: Re: [firedrake] FD on mac Ok, thanks a lot Lawrence. Floriane ________________________________________ De : firedrake-bounces@imperial.ac.uk <firedrake-bounces@imperial.ac.uk> de la part de Lawrence Mitchell <lawrence.mitchell@imperial.ac.uk> Envoyé : jeudi 3 décembre 2015 10:11 À : firedrake@imperial.ac.uk Objet : Re: [firedrake] FD on mac Dear Floriane,
On 2 Dec 2015, at 15:35, Floriane Gidel [RPG] <mmfg@leeds.ac.uk> wrote:
Dear Lawrence,
Here is what I get when trying the commands you suggested:
campus-10-41-49-255:~ mmfg$ mpicc -show clang -I/usr/local/Cellar/open-mpi/1.8.4/include -L/usr/local/opt/libevent/lib -L/usr/local/Cellar/open-mpi/1.8.4/lib -lmpi
campus-10-41-49-255:~ mmfg$ mpicc --version Apple LLVM version 5.1 (clang-503.0.40) (based on LLVM 3.4svn) Target: x86_64-apple-darwin13.4.0 Thread model: posix
campus-10-41-49-255:~ mmfg$ mpicxx -show clang++ -I/usr/local/Cellar/open-mpi/1.8.4/include -L/usr/local/opt/libevent/lib -L/usr/local/Cellar/open-mpi/1.8.4/lib -lmpi_cxx -lmpi
campus-10-41-49-255:~ mmfg$ mpicxx --version Apple LLVM version 5.1 (clang-503.0.40) (based on LLVM 3.4svn) Target: x86_64-apple-darwin13.4.0 Thread model: posix
OK, so my hypothesis for what was going wrong doesn't appear to be the case. Let's try and fix this tomorrow when we're in the same room. Cheers, Lawrence _______________________________________________ firedrake mailing list firedrake@imperial.ac.uk https://mailman.ic.ac.uk/mailman/listinfo/firedrake
Dec 4: Firedrake works, on 3 mac laptops! Tested many demos on the train fare back with Will. Plus, ++: great progress on the firedrake computer science of 2 conservative FSI challenges, a Hamiltonian 3D linear internal GW problem, and a variational 3D potential flow water-wave model. Thx to Lawrence, Miklos, Colin & David at our 5-(wo)man visit! Thx guys for having us! ________________________________________ From: firedrake-bounces@imperial.ac.uk <firedrake-bounces@imperial.ac.uk> on behalf of Onno Bokhove <O.Bokhove@leeds.ac.uk> Sent: Thursday, December 3, 2015 4:32 PM To: firedrake@imperial.ac.uk Subject: Re: [firedrake] FD on mac Have el-capitan installed now so t'row at IC we'll see whether the script works! Off on the train. ________________________________________ From: firedrake-bounces@imperial.ac.uk <firedrake-bounces@imperial.ac.uk> on behalf of Floriane Gidel [RPG] <mmfg@leeds.ac.uk> Sent: Thursday, December 3, 2015 2:17 PM To: firedrake@imperial.ac.uk Subject: Re: [firedrake] FD on mac Ok, thanks a lot Lawrence. Floriane ________________________________________ De : firedrake-bounces@imperial.ac.uk <firedrake-bounces@imperial.ac.uk> de la part de Lawrence Mitchell <lawrence.mitchell@imperial.ac.uk> Envoyé : jeudi 3 décembre 2015 10:11 À : firedrake@imperial.ac.uk Objet : Re: [firedrake] FD on mac Dear Floriane,
On 2 Dec 2015, at 15:35, Floriane Gidel [RPG] <mmfg@leeds.ac.uk> wrote:
Dear Lawrence,
Here is what I get when trying the commands you suggested:
campus-10-41-49-255:~ mmfg$ mpicc -show clang -I/usr/local/Cellar/open-mpi/1.8.4/include -L/usr/local/opt/libevent/lib -L/usr/local/Cellar/open-mpi/1.8.4/lib -lmpi
campus-10-41-49-255:~ mmfg$ mpicc --version Apple LLVM version 5.1 (clang-503.0.40) (based on LLVM 3.4svn) Target: x86_64-apple-darwin13.4.0 Thread model: posix
campus-10-41-49-255:~ mmfg$ mpicxx -show clang++ -I/usr/local/Cellar/open-mpi/1.8.4/include -L/usr/local/opt/libevent/lib -L/usr/local/Cellar/open-mpi/1.8.4/lib -lmpi_cxx -lmpi
campus-10-41-49-255:~ mmfg$ mpicxx --version Apple LLVM version 5.1 (clang-503.0.40) (based on LLVM 3.4svn) Target: x86_64-apple-darwin13.4.0 Thread model: posix
OK, so my hypothesis for what was going wrong doesn't appear to be the case. Let's try and fix this tomorrow when we're in the same room. Cheers, Lawrence _______________________________________________ firedrake mailing list firedrake@imperial.ac.uk https://mailman.ic.ac.uk/mailman/listinfo/firedrake _______________________________________________ firedrake mailing list firedrake@imperial.ac.uk https://mailman.ic.ac.uk/mailman/listinfo/firedrake
Humph, I get the following:
On 1 Dec 2015, at 11:58, Justin Chang <jychang48@gmail.com> wrote:
Yeah, that's not the case for me. I did:
brew uninstall gcc brew uninstall openmpi
brew install openmpi brew install gcc
and here's what I have:
mpicc --show clang -I/usr/local/Cellar/open-mpi/1.10.1/include -L/usr/local/opt/libevent/lib -L/usr/local/Cellar/open-mpi/1.10.1/lib -lmpi
mpicxx --show clang++ -I/usr/local/Cellar/open-mpi/1.10.1/include -L/usr/local/opt/libevent/lib -L/usr/local/Cellar/open-mpi/1.10.1/lib -lmpi_cxx -lmpi
mpif90 --show -------------------------------------------------------------------------- No underlying compiler was specified in the wrapper compiler data file (e.g., mpicc-wrapper-data.txt) --------------------------------------------------------------------------
$ brew uninstall gcc Uninstalling /usr/local/Cellar/gcc/5.2.0... (1353 files, 248M) $ brew uninstall openmpi Uninstalling /usr/local/Cellar/open-mpi/1.10.0... (785 files, 22M) $ brew install openmpi ==> Downloading https://homebrew.bintray.com/bottles/open-mpi-1.10.0.el_capitan.bottle.tar.g... Already downloaded: /Library/Caches/Homebrew/open-mpi-1.10.0.el_capitan.bottle.tar.gz ==> Pouring open-mpi-1.10.0.el_capitan.bottle.tar.gz /usr/local/Cellar/open-mpi/1.10.0: 785 files, 22M $ mpif90 -show /usr/local/bin/gfortran -I/usr/local/Cellar/open-mpi/1.10.0/include -Wl,-flat_namespace -Wl,-commons,use_dylibs -I/usr/local/Cellar/open-mpi/1.10.0/lib -L/usr/local/opt/libevent/lib -L/usr/local/Cellar/open-mpi/1.10.0/lib -lmpi_usempif08 -lmpi_usempi_ignore_tkr -lmpi_mpifh -lmpi $ mpif90 -------------------------------------------------------------------------- The Open MPI wrapper compiler was unable to find the specified compiler /usr/local/bin/gfortran in your PATH. Note that this compiler was either specified at configure time or in one of several possible environment variables. -------------------------------------------------------------------------- $ brew install gcc ==> Downloading https://homebrew.bintray.com/bottles/gcc-5.2.0.el_capitan.bottle.2.tar.gz Already downloaded: /Library/Caches/Homebrew/gcc-5.2.0.el_capitan.bottle.2.tar.gz ==> Pouring gcc-5.2.0.el_capitan.bottle.2.tar.gz ==> Caveats GCC has been built with multilib support. Notably, OpenMP may not work: https://gcc.gnu.org/bugzilla/show_bug.cgi?id=60670 If you need OpenMP support you may want to brew reinstall gcc --without-multilib ==> Summary /usr/local/Cellar/gcc/5.2.0: 1353 files, 248M $ mpif90 -show /usr/local/bin/gfortran -I/usr/local/Cellar/open-mpi/1.10.0/include -Wl,-flat_namespace -Wl,-commons,use_dylibs -I/usr/local/Cellar/open-mpi/1.10.0/lib -L/usr/local/opt/libevent/lib -L/usr/local/Cellar/open-mpi/1.10.0/lib -lmpi_usempif08 -lmpi_usempi_ignore_tkr -lmpi_mpifh -lmpi lmitche1@dyn1224-221:~/Documents/work/mapdes/petsc g firedrake d st $ mpif90 gfortran: fatal error: no input files compilation terminated. So the ElCap bottle of openmpi definitely has /usr/local/bin/gfortran baked in. I don't recall if you're on el cap or mavericks, or something else. Cheers, Lawrence
Dear Onno, On 30/11/15 22:14, Onno Bokhove wrote:
I get this:
ls -l $(which gfortran)
lrwxr-xr-x 1 root admin 32 24 Mar 2014 /usr/local/bin/gfortran -> /usr/local/gfortran/bin/gfortran
gfortran works fine; hello.f90 works fine; it is the openmpi bit when I add "use mpi"
use mpi
1
Fatal Error: Can't open module file 'mpi.mod' for reading at (1): No such file or directory
mpif90 hello.f90 -o hello
command does not work.
I /think/ what has happened is that homebrew has got in a mess somehow and mpif90 is pointing at the wrong place. What is the output of: mpicc -show mpif90 -show Cheers, Lawrence
participants (5)
-
David Ham
-
Floriane Gidel [RPG]
-
Justin Chang
-
Lawrence Mitchell
-
Onno Bokhove