General information regarding OpenHPC

  1. What is the release cycle of OpenHPC?
    The release cycle is expected to evolve over time. An initial release is timed for November 2015 and subsequent incremental releases will be relative to this initial 1.0 release. To start, we expect the incremental updates to occur at roughly quarterly intervals.
  2. What is the overall support plan for OpenHPC?
    OpenHPC offers community driven support via several email lists for end users and developers.
  3. What tool sets are included in OpenHPC? When will they be available?
    Please see the GitHub site for more details on which components are included in each OpenHPC release. These can be tracked via the Milestone definitions. In addition, the documentation for each release includes a manifest of pre-packaged components (and their version info).
  4. Is it required that all components from the stack be used or can you pick and choose?
    No, it is not required to install all components. In particular, development tools and libraries can be chosen based on local application requirements. Note, however, that package dependencies are includes as part of the OpenHPC integration effort, so sub dependencies may be required in order to install particular components.
  5. How can one view the dependencies of packages and the resulting growth in code size as packages are added?
    Package dependencies can be ascertained via package manager tools during installation from an OpenHPC repository (or mirror). Similarly, the repository size can also be queried via standard package manager tools (e.g. “yum repoinfo” on CentOS).
  6. Is the base OS included in with the stack?
    No. A supported base operating system should be installed first.
  7. What is meant by build and test environment?
    The build environment refers to the Open Build Service (OBS) that serves as the underlying build infrastructure. It initiates package builds based on commit triggers, tracks inter-package dependencies, and publishes resulting repositories. The test environment corresponds to the use of continuous integration (CI) infrastructure and the companion OpenHPC test suite to carry out a variety of integration oriented tests as the project evolves.
  8. How is the OpenHPC stack validated?
    Included within OpenHPC is an integration test suite that is executed on a variety of cluster platforms prior to release. The intent of this suite is to ensure basic functionality of each tested component when installed on a cluster beginning with a bare metal installation.
  9. How were the initial packages for the stack selected?
    Package selection was prioritized initially with an initial goal of being able to provide basic end-to-end functionality for common services typical of HPC deployments. The initial functionality includes example provisioning and resource management services along with a variety of administrative utilities and development tools/libraries.
  10. How are future packages for the stack selected?
    Package selection will be prioritized based on a number of factors including (but not limited to) its relevance and usage by the HPC community, development stability, required integration effort, coexistence capability with other included packages, functionality coverage, and willingness of community members to help integrate and maintain the package.
  11. I see that OpenHPC includes 3rd-party libraries compiled with Intel® Parallel Studio. Can I use these for free?
    While Intel® Parallel Studio compilers and runtimes are not a part of OpenHPC, we provide compatibility packages to allow existing installations of the Intel® toolchains to integrate with available development/scientific libraries available via the OpenHPC repositories.  Note that licenses for Intel® Parallel Studio development tools are freely available to academic researchers and open-source contributors.  For more information on this topic, please see: https://software.intel.com/en-us/qualify-for-free-software