Package Manager List

This is a list of package managers available within Ramble, for managing experiment software stacks. It is automatically generated based on the package managers in this Ramble version.

Ramble currently has 7 mainline package_managers:

eessi spack user-managed
environment-modules spack-lightweight
pip spack-pip

eessi

Ramble package_managers:
Description:
Package manager definition for EESSI This package manager uses the European Environment for Scientific Software Installations to stream binaries from an EESSI mirror. https://www.eessi.io/ To control the target architecture used, add EESSI_SOFTWARE_SUBDIR_OVERRIDE to the workspace's environment variable definitions.

environment-modules

Ramble package_managers:
Description:
Definition for using environment-modules as a package manager This definition allows experiments to use environment-modules to manage the software used in an experiment. It assumes the `module` command will be in the path of the experiment at execution time.

pip

Ramble package_managers:
Description:
Pip package manager class definition

spack

Ramble package_managers:
Description:
Full Spack package manager class definition This inherits from the spack-lightweight package manager, and extends it by adding the software_install and define_package_paths phases.

spack-lightweight

Ramble package_managers:
Description:
Lightweight version of Spack package manager class definition This implements most of the logic for the spack package manager. The primary difference is the software_install and define_package_paths are not executed during setup, as they should be deferred to installation time. The installation portion is not defined in this class, as applications that require this should add this on separately.

spack-pip

Ramble package_managers:
Description:
Chained spack + pip package manager This package manager uses spack to install a spack environment, then using the python that is installed within this environment (or whichever one is available in your path) will install pip packages into a python virtual environment. Experiments will load both environments, in the order of spack then pip.

user-managed

Ramble package_managers:
Description:
Package manager representing a user managed environment. This package manager is used when the software required for the experiments is manually installed outside of Ramble. Generally, a user will need to convey the paths to these installed packages to Ramble through specific variable definitions.