Researchers may, upon approval, purchase nodes to add to CHPC Clusters where they may have exclusive or priority access. CHPC will provide the environmental infrastructure (nests, networking, power and cooling) and system administration. Researchers may purchase nodes leveraging this infrastructure provided by CHPC. Overhead on hardware added to the clusters may be waived by the VP of Research.

Below are the policies related to participating in this program:

• Please work closely with CHPC to determine specifications. Due to the additional administrative workload associated with establishing and maintaining a group of nodes with the same characteristcs, CHPC requires a minimum of four identical nodes (8 proc minimum) initial purchase. You may in general select the speed and memory for these purchases, but there may be minimum requirements on memory and disk space, as negotiated with CHPC, and the networking cards may be specifically required (i.e. inifiniband or myrinet). Please contact CHPC to obtain latest quotes for what you're interested in. Your purchase will need to include network card(s).
• Once you have decided on a configuration, send a memo to the Director of CHPC, Julio Facelli at Julio.Facelli@utah.edu, requesting the addition of xx nodes to with the following yy properties, which cost zz dollars.
• Request that your research program get priority for the nodes according to CHPC policies (see below).
• Upon approval the Director of CHPC will ask you for an account number to which to charge the purchase. CHPC will do all the paper work so we can keep track of delivery and inventory.
• Once the equipment is received by CHPC, please allow two weeks for configuration and burn in before your nodes are available.
• If your purchase is for more that 16 nodes, additional time will be required and we will give you an estimated availability date.
• Priority Policies: Priority and policies will be determined on a case by case basis, but usually read as follows:
  • Users in your group will be given highest priority for running on your nodes, as long as your group still has available allocation.
  • Users in your group will compete against each other for priority and CHPC will use fairshare and other techniques to determine scheduling. Parallelism will be rewarded.
  • Each calendar quarter, you will be awarded an allocation on the cluster equal to the total number of Service Units available on your nodes during the entire quarter.
• When contributing nodes to a CHPC cluster, it is expected that CHPC will maintain and house these nodes in the cluster for the useful life of the nodes (~3 years.)
• When a node's lifetime exceeds the warranty period, CHPC does not have an obligation to maintain that hardware. After the warranty period (typically 3 years) CHPC will do basic maintenance on the nodes, but if there is a hardware failure of some kind rendering the node offline, CHPC may choose not to fix it. Minor hardware repairs will be performed by CHPC on a best effort basis.
• CHPC reserves the right to retire nodes from the clusters in the event that the warranty has expired and CHPC determines a best effort has been made to repair the node.

Please contact CHPC with any questions regarding these procedures.