You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Copy file name to clipboardexpand all lines: README.md
+3-1
Original file line number
Diff line number
Diff line change
@@ -4,7 +4,9 @@ This project provides provisioned HPC cluster models using underlying virtualiza
4
4
5
5
The purpose of this project is to provide a common baseline for repeatable HPC experiments. This has been used for education, distributed collaboration, tool development colaboration, failure signature discovery, local HPC debugging and cluster configuration comparisons, enabled by construction and use of short-lived and common baseline hpc cluster models. In short, extend the "systems as cattle not pets" <AHREF="http://www.pass.org/eventdownload.aspx?suid=1902">[1]</A> <AHREF="http://cloudscaling.com/blog/cloud-computing/the-history-of-pets-vs-cattle/">[2]</A> analogy into the realm of "clusters as cattle, not pets."
6
6
7
-
The initial release requires local enablers: gmake, vagrant and virtualbox. Lighterweight and multi-node mechanisms, such as containers, jails and pods, are planned.
7
+
The initial release requires local enablers: gmake, vagrant and virtualbox and/or libvirt. Lighterweight and multi-node mechanisms, such as containers, jails and pods, are planned. Virtualbox is slower than libvirt provisioning by 50%, although
8
+
<AHREF="https://github.com/hpc/hpc-collab/issues/158">more consistent</A> and
Two representative HPC cluster recipes are provided. Cluster recipes are in the <EM>clusters</EM> directory. Presently, recipes generate clusters local to the installation host.
0 commit comments