Open lorenzsp opened 2 months ago
I am currently facing the same problem (M1 chip trying to install FEW 1.5.5 master branch). However, now when I try installing following this path (or any of the others), I get the following error:
src/inspiralwrap.cpp:5181:20: error: no member named 'dealloc' in 'ODECarrier' __pyx_v_self->g->dealloc();
Running the command twice doesn't fix it.
I found a temporary solution to install FEW on M1-2 chips.
git clone https://github.com/BlackHolePerturbationToolkit/FastEMRIWaveforms.git
cd FastEMRIWaveforms
git checkout e4038da
conda create -n few_env -c conda-forge wget gsl hdf5 numpy Cython scipy tqdm jupyter ipython h5py requests matplotlib python=3.9 openblas lapack liblapacke\nconda activate few_env
conda activate few_env
python setup.py install --ccbin /usr/bin/
python -m unittest discover
This should solve the installation temporarily and solve issues #96 #93 . Let me know if it works!
I found a temporary solution to install FEW on M1-2 chips.
git clone https://github.com/BlackHolePerturbationToolkit/FastEMRIWaveforms.git cd FastEMRIWaveforms git checkout e4038da conda create -n few_env -c conda-forge wget gsl hdf5 numpy Cython scipy tqdm jupyter ipython h5py requests matplotlib python=3.9 openblas lapack liblapacke\nconda activate few_env conda activate few_env python setup.py install --ccbin /usr/bin/ python -m unittest discover
This should solve the installation temporarily and solve issues #96 #93 . Let me know if it works!
For me, I changed python setup.py install --ccbin /usr/bin/
to python -m pip install .
then it worked.
otherwise, it has this error when testing:
ImportError: dlopen(/Path/to/conda/envs/emri_env/lib/python3.9/site-packages/few-1.5.4-py3.9-macosx-11.0-arm64.egg/pyUtility.cpython-39-darwin.so, 0x0002): tried: '/Path/to/conda/envs/emri_env/lib/python3.9/site-packages/few-1.5.4-py3.9-macosx-11.0-arm64.egg/pyUtility.cpython-39-darwin.so' (mach-o file, but is an incompatible architecture (have 'x86_64', need 'arm64')), '/System/Volumes/Preboot/Cryptexes/OS/Path/to/conda/envs/emri_env/lib/python3.9/site-packages/few-1.5.4-py3.9-macosx-11.0-arm64.egg/pyUtility.cpython-39-darwin.so' (no such file), '/Path/to/conda/miniconda3/envs/emri_env/lib/python3.9/site-packages/few-1.5.4-py3.9-macosx-11.0-arm64.egg/pyUtility.cpython-39-darwin.so' (mach-o file, but is an incompatible architecture (have 'x86_64', need 'arm64'))
(I am on an M1 Mac)
I am running into problems when installing on an M1. The old installation method described below works only when running the final command
python setup.py install --ccbin /usr/bin/
two times. The first time I run the command, I get a cython error. If I run the command again, it works.Old installation:
0) Install Anaconda if you do not have it. 1) Create a virtual environment. Note: There is no available
conda
compiler for Windows. If you want to install for Windows, you will probably need to add libraries and include paths to thesetup.py
file. If on linux:If on MACOSX, substitute
gcc_linux-64
andgxx_linus-64
withclang_osx-64
andclangxx_osx-64
as follows:If on M1 chip use the following command:
2) Clone the repository.
3) If on MACOSX or linux run install:
If on M1 chip use the following command: