This module is a singularity container wrapper for ghcr.io/autamus/intel-mkl. Intel oneAPI Math Kernel Library, formerly just Intel Math Kernel Library, is a library of optimized math routines for science, engineering, and financial applications. After installing shpc you will want to install this container module:

$ shpc install ghcr.io/autamus/intel-mkl

Or a specific version:

$ shpc install ghcr.io/autamus/intel-mkl:2020.4.304

And then you can tell lmod about your modules folder:

$ module use ./modules

And load the module, and ask for help, or similar.

$ module load ghcr.io/autamus/intel-mkl/2020.4.304
$ module help ghcr.io/autamus/intel-mkl/2020.4.304

You can use tab for auto-completion of module names or commands that are provided.

Commands

When you install this module, you will be able to load it to make the following commands accessible. Examples for both Singularity, Podman, and Docker (container technologies supported) are included.

intel-mkl-run:

$ singularity run <container>
$ podman run --rm  -v ${PWD} -w ${PWD} <container>
$ docker run --rm  -v ${PWD} -w ${PWD} <container>

intel-mkl-shell:

$ singularity shell -s /bin/sh <container>
$ podman run --it --rm --entrypoint /bin/sh  -v ${PWD} -w ${PWD} <container>
$ docker run --it --rm --entrypoint /bin/sh  -v ${PWD} -w ${PWD} <container>

intel-mkl-exec:

$ singularity exec <container> "$@"
$ podman run --it --rm --entrypoint ""  -v ${PWD} -w ${PWD} <container> "$@"
$ docker run --it --rm --entrypoint ""  -v ${PWD} -w ${PWD} <container> "$@"

intel-mkl-inspect:

Podman and Docker only have one inspect type.

$ podman inspect <container>
$ docker inspect <container>

intel-mkl-inspect-runscript:

$ singularity inspect -r <container>

intel-mkl-inspect-deffile:

$ singularity inspect -d <container>

intel-mkl

$ singularity run <container>
$ podman run --rm  -v ${PWD} -w ${PWD} <container>
$ docker run --rm  -v ${PWD} -w ${PWD} <container>

In the above, the <container> directive will reference an actual container provided by the module, for the version you have chosen to load. An environment file in the module folder will also be bound. Note that although a container might provide custom commands, every container exposes unique exec, shell, run, and inspect aliases. For anycommands above, you can export:

• SINGULARITY_OPTS: to define custom options for singularity (e.g., –debug)
• SINGULARITY_COMMAND_OPTS: to define custom options for the command (e.g., -b)
• PODMAN_OPTS: to define custom options for podman or docker
• PODMAN_COMMAND_OPTS: to define custom options for the command

Install

You can install shpc locally (for yourself or your user base) as follows:

$ git clone https://github.com/singularityhub/singularity-hpc
$ cd singularity-hpc
$ pip install -e .

Have any questions, or want to request a new module or version? ask for help!