fatiando\gravmag\_polyprism.c:349:17: fatal error: omp.h: No such file or directory

[mtb-za]

Currently trying to reinstall Fatiando in a new conda environment, after following this guide: http://conda.pydata.org/docs/examples/create.html. After switching to the newly created environment (with python2.7 and numpy 1.9), I ran 'pip install fatiando'. Unfortunately, I get the above error.

Relevant extract from log:

building 'fatiando.gravmag._polyprism' extension

C:\Users\foo\AppData\Local\Continuum\Anaconda\Scripts\gcc.bat -DMS_WIN64 -mdll -O -Wall -IC:\Users\foo\AppData\Local\Continuum\Anaconda\lib\site-packages\numpy\core\include -IC:\Users\foo\AppData\Local\Continuum\Anaconda\include -IC:\Users\foo\AppData\Local\Continuum\Anaconda\PC -c fatiando\gravmag_polyprism.c -o build\temp.win-amd64-2.7\Release\fatiando\gravmag_polyprism.o -fopenmp

In file included from C:\Users\foo\AppData\Local\Continuum\Anaconda\lib\site-packages\numpy\core\include/numpy/ndarraytypes.h:1761:0,

from C:\Users\foo\AppData\Local\Continuum\Anaconda\lib\site-packages\numpy\core\include/numpy/ndarrayobject.h:17,

from C:\Users\foo\AppData\Local\Continuum\Anaconda\lib\site-packages\numpy\core\include/numpy/arrayobject.h:4,

from fatiando\gravmag_polyprism.c:347:

C:\Users\foo\AppData\Local\Continuum\Anaconda\lib\site-packages\numpy\core\include/numpy/npy_1_7_deprecated_api.h:12:9: note: #pragma message: C:\Users\foo\AppData\Local\Continuum\Anaconda\lib\site-packages\numpy\core\include/numpy/npy_1_7_deprecated_api.h(12) : Warning Msg: Using deprecated NumPy API, disable it by #defining NPY_NO_DEPRECATED_API NPY_1_7_API_VERSION

fatiando\gravmag_polyprism.c:349:17: fatal error: omp.h: No such file or directory

compilation terminated.

error: command 'C:\Users\foo\AppData\Local\Continuum\Anaconda\Scripts\gcc.bat' failed with exit status 1

[eusoubrasileiro]

Hi @mtb-za I am glad in your interest in fatiando.

From docs Extra dependencies for Windows users you have to install TDM-GCC to be able to compile OpenMP code.

I bet it will solve your problem but if you have any other just share with us.

[leouieda]

Hi @mtb-za, thanks for reporting the issue.

My first thought was what @eusoubrasileiro said, just try (re)installing TDM. We get this error a lot under windows. But it might also be caused by the conda environment. TDM installs on top of Anaconda gcc. So it might not be carried along when you switch conda environments. This is mere speculation, by the way.

Is the install working if you don't use conda environments?

PS: These errors are making me seriously question the utility of the badly implemented parallelism that we have. Might be better to ditch Cython openmp and use a higher level multiprocessing in the Python code. (Update: discussing this in #165)

[mtb-za]

This definitely solved the issue, although since I tested it by installing it outside, I am not sure if it has just used that one.

Feel free to close this issue as solved.