Home of the Singularity container project.

Containers for HPC, analytics, machine learning, reproducible and trusted computing

  • Thank you for your work on the Singularity project, which my lab and others at JHU use extensively.

    Johns Hopkins University
  • Singularity is the best option among the big three considerations for HPC

    DoD Army Research Laboratory
  • Singularity allows us to containerize previous Docker workflows without exposing us to simple root escalation attacks.

    R Systems Inc.
  • Singularity is a fabulous tool for providing forward and backward software compatibility on clusters and for reproducibility

    University of Michigan
  • Users have been asking for containers for years, but I've always resisted. Singularity addressed the majority of my complaints and couldn't have been easier to install.

    Washington University in St. Louis
  • This is a brilliant containerization technology which will gain traction very quickly.

    University of Cape Town
  • We used Singularity to get around a GLIBC version requirement for a binary distribution of the NCI GDC download tool on CentOS 6

    Rutgers University
  • Singularity allowed us to use software that was otherwise impossible to install under SL6, such as TensorFlow

    University of Chicago
  • Our focus is mainly on containers converted from Docker to singularity

    University of Oslo
  • We are using CentOS6 and not able to use TensorFlow without Singularity.

    Hong Kong University of Science & Technology

There are containers for microservices, and then there is Singularity

Simplicity

Singularity uses a single image format for runtime. You can copy them, move them, set permissions on them... whatever!  If you can use Linux, you can use Singularity! 

Security

Singularity blocks privilege escalation inside of containers with a single file container format that can be signed and verified

Community

Singularity runs on some of the biggest super computers in the world.  Come join our friendly and knowledgeable communities, via GitHub, Google Groups, or in the Slack community channel. 

Open Source

People all over the world contribute to the Singularity project, and you can too! Just fork the Singularity project on GitHub and send us a pull request.


The features you need to succeed

Download and run containers directly from Docker Hub

Read and write data directly to your host system with ease

Seamlessly integrate with resource managers and MPI

Leverage NVIDIA GPUs on the host with zero configuration

Run Singularity instances in the background for services support

Use persistent overlays to bring your data anywhere