Switch between miniforge (Apple Silicon build) and miniconda (Rosetta 2)

[fblamanna]

I am just wondering how to have and switch between:

1. miniforge3 with native support for apple silicon
2. miniconda with all packages installed via Rosetta 2 emulation

on a Mac.
This is because some environment works well under M1, but for other packages (not yet supported) I have to install them via emulation on a classic miniconda environment.

Is there a method to install both (I guess so) and switch between then to update the terminal zsh shell?

[isuruf]

CONDA_SUBDIR=osx-64 conda create -n rosetta python   # create a new environment called rosetta with intel packages.
conda activate rosetta
python -c "import platform;print(platform.machine())"
conda config --env --set subdir osx-64  # make sure that conda commands in this environment use intel packages

[fblamanna]

Thanks it works, just some warnings aobut overwriting the system SUBDIR when switching between arm64 and osx-64 environments.

[hmaarrfk]

Would you like to add a note in the readme explaining and documenting this others?

[avocardio]

Sorry this might be a stupid question, is there any way to just have miniforge and have it run all the non-supported packages through rosetta? I currently use anaconda3, don't really want to be switching between that and miniforge... how would I tell the environments apart?

[hmaarrfk]

Are you trying to run one program (single address space) that is dynamically linked to both x86 code (through Rosetta) and aarch64 code (native)?

[ollie-bell]

CONDA_SUBDIR=osx-64 conda create -n rosetta python   # create a new environment called rosetta with intel packages.
conda activate rosetta
python -c "import platform;print(platform.machine())"
conda env config vars set CONDA_SUBDIR=osx-64  # make sure that conda commands in this environment use intel packages
conda deactivate rosetta
conda activate rosetta
echo "CONDA_SUBDIR: $CONDA_SUBDIR"

So, just to be explicitly clear, is this supposed to work if you only have Miniforge for arm64 installed? i.e. you do not need a separate installation for Miniconda or Anaconda?

[BastianZim]

Yes your current setup (Just Miniforge for Apple Silicon) should be sufficient. You don't need to have a second installation. This will be activated per environment.

[ricardowiseverge]

I've managed to create the environment but when I run jupyter lab the correct kernel doesn't appear.

I've 2 python installed one on M1 and another on rosetta2 because I've specific packages that don't support M1 architecture.

So, I'm running those commands on a Rosetta2 terminal and it can't find the correct kernel version