I eventually want to use valgrind to find what is causing the occasional bizarre output in a C program which refines a model against experimental data using OpenMP parallel programming.
To avoid the use of the nominal gcc (ie clang) compiler, I used brew to install gcc-4.9 on my MacPro running Yosemite (OS x 10.10.5). However, when trying to compile my program with gcc-4.9, with or without -fopenmp, I get numerous error messages of the type:
/var/folders/qc/1j0gr_l48xnfd9001s6tt6f80000gn/T//ccRxnrnU.s:30597:suffix
or operands invalid for `movq'
I have no idea what the problem triggering these error messages is. Can anyone help?
The following summarises what was worked out in the comments section and did lead to the issue being resolved. Not all steps may be necessary, but most are probably good practice.
Step 1 - Clean up
If you have been trying lots of different, potentially incompatible, methods to get OpenMP set up, it is probably a good idea to clean them up first. So, something like:
brew rm --force gcc # or maybe gcc#4.9
Step 2 - Update Xcode and Command Line Tools
If you have upgraded macOS since installing Xcode, it is probably advisable to update Xcode and its "Command Line Tools"
Consider uninstalling and re-installing Xcode - it is available for free from the App Store.
Update/install the "Command Line Tools" after installing/updating with:
xcode-select --install
Step 3 - Install gcc
Now, try installing gcc afresh, ensuring that you use the --without-multilib option:
brew install gcc#4.9 --without-multilib
Hopefully you can now compile OpenMP code with:
/usr/local/bin/gcc -fopenmp program.c -o program
I am unsure exactly why the --without-multilib option is needed and prefer to quote #hristo-iliev:
Multilib usually refers to the coexistence of both 64-bit and 32-bit
versions of each library so that 32-bit software could be run on
64-bit OS. In the GCC case that probably refers to having all GCC
runtime libraries in "fat" Mach-O format, i.e. versions for both i386
and x86_64 in the same shared library file. It could be that libgomp
(the GNU OpenMP runtime library) cannot be built in such a way.
See this question.
Keywords: gcc, g++, GNU Compiler, OpenMP, fopenmp, -fopenmp, Xcode, multilib, Command Line Tools, macOS, OSX, homebrew, brew
Related
I am trying to install a program and when I run make or make -f Makefile (following the installation instructions) I get the following output:
g77 -O5 -Wall -c prep_output.f -o prep_output.o
dyld: Symbol not found: ___keymgr_global
Referenced from: /usr/local/bin/g77
Expected in: /usr/lib/libSystem.B.dylib
make: *** [prep_output.o] Abort trap: 6
I am working on a macOS Mojave 10.14.6. Following some other proposed solutions in similar problems, I've already installed Command line tools in my Xcode but that didn't do the trick. Any suggestions please??
Thanks in advance
EDIT:
I got rid of g77 and installed a compatible version of gcc (gcc8) through MacPorts. After running the make command I get the following:
g77 -O5 -Wall -c prep_output.f -o prep_output.o
make: g77: No such file or directory
make: *** [prep_output.o] Error 1
So I guess the program still needs the g77 setup? This is the program btw ([http://www.cfht.hawaii.edu/~arnouts/LEPHARE/install.html]). Is there a way to rely to the gfortran compiler for building the program?
Thanks in advance
P.S. I noticed that when I install g77 I get the following error:
x usr/local/: Can't set user=0/group=0 for usr/local`
`tar: Error exit delayed from previous errors.
However it still installs g77..
The error message indicates that your installation of g77 is broken. This has nothing to do with make or the particular project you're trying to build, except inasmuch as the project is trying to use g77 in the first place. That is a bit surprising, actually, since g77 has been obsolete for years (gfortran is the current GNU Fortran compiler), but I'm uncertain what exactly to expect from XCode in this area.
On the other hand, since the full path to the binary is /usr/local/bin/g77, I'm further inclined to think that you're not using XCode for this at all. Possibly you've dumped a g77 built on some other system into your /usr/local/bin, and it's not compatible with your Mojave system.
Your best bet is probably to
Get rid of your broken g77 installation.
Install Fink or MacPorts, or a similar project.
Install the Fink / MacPorts / whatever package for gfortran (maybe gcc-gfortran or similar in some of those) to get a working Fortran compiler.
Rely on that compiler to build your project.
I was trying to use the omh.h header file and I realized it was missing. I tried reinstalling gcc on my mac using brew. This is the message I got at the end of the installation.
..
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/4.9.2_1: 1156 files, 203M
It suggests that if I need OpenMP support I need to install brew reinstall gcc --without-multilib. I am not able to understand the meaning of --without-multilib. How is it different from simply installing gcc?
And apparently brew reinstall gcc --without-multilib takes forever to run and brew uninstall gcc && brew install gcc was lightning fast.
Multilib means support for multiple architectures, so you can compile binaries for them.
Given the bug you show, you should be safe as long you build binaries for AMD64 architecture (64-bit) and IA32, as that is what your machine is.
Otherwise, better to choose without multilib.
On Mac OSX 10.9, the default c compiler bundled with Xcode is clang. I installed gcc-4.9 with homebrew. Now I have two different gccs, one is clang, the other is gcc. The default is clang.
I want to use gcc when compiling Haskell files with ghc, and I want also gcc when I launch ghci. How to do this change?
Reproducing my directions I've been sharing with haskellers for the past few months via https://gist.github.com/cartazio/7131371
Type ghc --print-libdir
The output will be a path like /Library/Frameworks/GHC.framework/Versions/7.6.3-x86_64/usr/lib/ghc-7.6.3
Go to that directory and edit the settings file.
There'll be a line indicating the path to the C compiler. It'll probably say /bin/gcc
Change that line to /usr/local/bin/gcc-4.8 (or whichever gcc version you brew installed, such as /usr/local/bin/gcc-4.2)
According to this post (https://github.com/mxcl/homebrew/pull/2953), the flag "--with-mpi" should enable boost_mpi build support for the related homebrew formula, so I am trying to install boost via homebrew like this:
brew install boost --with-mpi
However, the actual boost mpi library is not being build and can not be found.
There is currently some work being done around this, according to: https://github.com/mxcl/homebrew/pull/15689
In summary, I can currently build boost, but it seems the "--with-mpi" flag is being ignored. Could someone please check, if I should be able to build boost (with mpi support) on Mac OS X Mountain Lion (10.8)?
The (verbose) output generates these lines:
MPI auto-detection failed: unknown wrapper compiler mpic++
Please report this error to the Boost mailing list: http://www.boost.org
You will need to manually configure MPI support.
warning: skipping optional Message Passing Interface (MPI) library.
note: to enable MPI support, add "using mpi ;" to user-config.jam.
note: to suppress this message, pass "--without-mpi" to bjam.
note: otherwise, you can safely ignore this message.
Not sure how exactly I can fix this and get the mpi stuff to be build - any ideas?
Just in case this helps anyone else along the line, here's how I fixed this. The main error is MPI auto-detection failed: unknown wrapper compiler mpic++, any typing mpic++ at the command line verified that it was not working properly for me. I used brew to install open-mpi, but the same error was showing in the verbose output for installing boost. A run of brew doctor showed that openmpi was not linked properly, so I fixed those errors and reran brew -v install boost --with-mpi --without-single and it finally built and installed all of the libraries without a problem
To anyone that comes across this, the package migrated to boost-python and boost-mpi separate from boost. Use brew install boost-mpi
Just get it worked on OSX 10.11.5. I've tried brew, but with no luck.
Suppose you already have gcc installed. Here are what I've done:
1. Find and disable (but do not remove) clang
clang alway cause headaches. There would be a lot of warnings when building Boost.
which clang, which should give you /usr/bin/clang
Rename it: sudo mv clang clang_mac_remove, also for clang++: sudo mv clang++ clang++_mac_remove. You can change the names back if you need them in future.
2. Install OpenMPI
If you already installed using brew, uninstall first. Becasue it would have used clang as the compiler wrapper by default. You need to change the wrapper to gcc.
Download the package.
Specify the wrapper compiler to gcc and g++:
./configure CC=gcc CXX=g++ F77=ifort FC=ifort --prefix=/usr/local
Below may take a long time.
make all
sudo make install
Reference: https://wiki.helsinki.fi/display/HUGG/Open+MPI+install+on+Mac+OS+X
3. Install Boost MPI
Download the package.
Run ./bootstrap.sh (can open it first and specify the toolset to gcc, otherwise, the default option is darwin for mac).
Add using mpi ; in project-config.jam file. Then ./b2 —with-mpi will only build the mpi library.
Then, all built libraries can be found in the folder ~/Downloads/boost_1_61_0/stage/lib.
Copy or move them to /usr/local/lib or any other commonly used library path.
Reference: http://www.boost.org/doc/libs/1_61_0/doc/html/mpi/getting_started.html
4. Compile with Boost MPI
LIBRARY DIR = -L/usr/local/lib
INCLUDE = -I/usr/local/include/
LINKER = -lboost_mpi -lboost_serialization
e.g.
mpic++ -std=c++11 -I/usr/local/include/ -c boost_test.cpp -L/usr/local/lib -lboost_mpi -lboost_serialization
Good luck!
I want to start working with C++0x. I see that GCC 4.7 has a fair amount of functionality available. I already have XCode 3.2 installed in /Developer
I downloaded: http://fileboar.com/gcc/snapshots/LATEST-4.7/gcc-4.7-20110528.tar.bz2
Can I somehow compile this in /opt/gcc-4.7? How do I then work with my path so I can compile with GCC 4.7 from the command-line but have OSX use the version it needs?
OSX does not need gcc to run - the Developer tools are optional. So you only need to choose between gcc's when you compile. In Xcode you chose explicitly the gcc andin Makefiles you can set $(CC) or similar to the full path.
Alternatively rename the gcc-4.7 gcc to gcc-4.7 and use that so gcc is always the Apple one.
For ease of using multiple C++ compilers I use macports (or fink or homebrew) which will compile the compilers with the correct patches and also has a port select command to switch between the C++ compilers