converged-computing / bare-vm-container-study

Testing the impact of containers vs. "bare meta" VM on the cloud

Geek Repo:Geek Repo

Github PK Tool:Github PK Tool

Bare Metal (VM) vs. Container

We want to test the difference between:

  • building an application on bare metal

  • building with a container on the same instruction set (the instance)

  • building with a container on a different instruction set (my host)

  • for each container build:

    • running on bare metal
    • running with singularity
    • running with docker
  • docker: includes Docker builds for:

  • host machine "a different machines" instances for 14 applications

  • a similar build script c2d-build.sh to build the "same" containers on c2d instances (I don't expect these to be different)

  • bare-vm: includes a monster script for that combines each previous docker build to build on bare metal, along with installing Singularity.

  • experiment: runs the experiment, once all the software is ready!

I was going to add Podman, but I think I hate it too much and I'm worried it has a conflicting dependency with docker. I'm going to be consistent and build flux / mpi into the containers. I don't know if I'll wind up using flux or MPI to run on one node, but it's worth having it.

License

HPCIC DevTools is distributed under the terms of the MIT license. All new contributions must be made under this license.

See LICENSE, COPYRIGHT, and NOTICE for details.

SPDX-License-Identifier: (MIT)

LLNL-CODE- 842614

About

Testing the impact of containers vs. "bare meta" VM on the cloud

License:MIT License


Languages

Language:Python 42.2%Language:Dockerfile 28.7%Language:Shell 8.6%Language:C 7.2%Language:Makefile 6.9%Language:Fortran 5.2%Language:C++ 0.9%Language:CMake 0.3%