Using EESSI with Apple Silicon on macOS 26¶

Requirements

Mac with Apple Silicon (M-series chip)

macOS 26 (Tahoe) or later

internet connection (EESSI software is streamed on demand from CVMFS Stratum 1 servers)/Other advanced solutions are also possible

1. Install the `container` tool¶

Download the latest signed installer package (.pkg) from the apple/container GitHub releases page.

Double-click the downloaded .pkg file and follow the on-screen instructions. When prompted, enter your administrator password to allow the installer to place files under /usr/local

Alternatively, using Homebrew:

brew install container

Verify the installation:

container --version

You should see output similar to:

container CLI version 0.10.0 (build: release, commit: unspeci)

2. Start the `container` system service¶

The container tool requires a background API server. Start it with:

container system start

To confirm the service is running:

container system status

3. Pull the EESSI client container image¶

Pull the EESSI ubuntu22.04-macOS-26 image from the GitHub Container Registry:

container image pull ghcr.io/eessi/client:ubuntu22.04-macOS-26

This image is an arm64 (Apple Silicon) Ubuntu 22.04 image that includes the CernVM-FS client pre-installed and configured for EESSI.

4. Start an interactive EESSI session¶

Run the container interactively:

container run -it --rm --name eessi ghcr.io/eessi/client:ubuntu22.04-macOS-26

Note: -it provides an interactive session inside the container and --rm keeps your system clean by automatically removing the stopped container after you exit.

5. Once inside the container shell¶

Verify that the EESSI repository is accessible:

ls /cvmfs/software.eessi.io

You should see output similar to:

README.eessi  defaults  host_injections  init  versions

Initialize the EESSI environment:

source /cvmfs/software.eessi.io/versions/2023.06/init/lmod/bash

Expected output:

Note: EESSI currently supports NVIDIA GPUs, so the message EESSI did not identify an accelerator on the system is expected.

Module for EESSI/2023.06 loaded successfully
EESSI has selected aarch64/neoverse_n1 as the compatible CPU target for EESSI/2023.06
EESSI did not identify an accelerator on the system
(for debug information when loading the EESSI module, set the environment variable EESSI_MODULE_DEBUG_INIT)

EESSI is now ready to use inside the container.

7. Use EESSI software¶

List available modules:

module avail

Search for a specific package:

module avail TensorFlow

Load and use a module:

module load TensorFlow/2.13.0-foss-2023a
python3 -c "import tensorflow as tf; print(tf.__version__)"

Cleanup¶

Remove the pulled image:

container image rm ghcr.io/eessi/client:ubuntu22.04-macOS-26

Stop the container system service when it is no longer needed:

container system stop