ACCarnall
commented
1 year ago I'm afraid this is a multinest installation issue rather than anything to do with Bagpipes. I've copied below my answer to a recent email I got asking a similar question, which hopefully might be helpful.
I also recently had a similar problem with my new M2 mac. I presume you're currently at the stage of trying to compile multinest? Some people have had success installing multinest on M1 macs using conda, with the command "conda install -c conda-forge multinest". However this didn't work for me for some reason.
The process I followed on my new laptop was:
- Download and install anaconda
- Install mpi and mpi4py using conda "conda install -c conda-forge mpi mpi4py"
- Install gfortran using conda "conda install -c conda-forge gfortran" (this is instead of gcc)
- Install cmake from cmake.org
- Download and install multinest from https://github.com/JohannesBuchner/MultiNest using the instructions provided in the readme file of that repository.
Hopefully this might work for you too, if not you might have to speak to someone more involved in MultiNest or your local tech support. I just use MultiNest, I don't have anything to do with developing it, so I can't really offer much specific advice on this.
Hi,
Thank you for creating a public code first! I am having an error downloading MultiNest to use bagpipes.
Even after setting the LD_LIBRARY_PATH, the following error occurred when trying to import bagpipes in jupyter notebook:
import bagpipes -> ERROR: Could not load MultiNest library: libmkl_gf_lp64.so.2 ERROR: You have to build MultiNest, ERROR: and point the LD_LIBRARY_PATH environment variable to it! ERROR: manual: https://johannesbuchner.github.io/PyMultiNest/install.html
problem: libmkl_gf_lp64.so.2: cannot open shared object file: No such file or directory Bagpipes: PyMultiNest import failed, fitting will be unavailable.
"libmkl_gf_lp64.so.2" was not found in my MultiNest/lib directory even after I build MultiNest.
After searching for the file within my team server, I found the file above and added the path to LD_LIBRARY_PATH as well: export LD_LIBRARY_PATH=/opt/intel/oneapi/mkl/2022.1.0/lib/intel64/:$LD_LIBRARY_PATH
Since running this code, I am able to import bagpipes, but as soon as I try to run the "fit.fit" in Example 3 - Fitting photometric data from https://github.com/ACCarnall/bagpipes/blob/master/examples/Example%203%20-%20Fitting%20photometric%20data%20with%20a%20simple%20model.ipynb, the kernel dies.
Do you have any idea what the problem might be and if there is a possible solution?
Thanks in advance Minseong.