The HPC/HTC 'Buy In' program is designed to be a “win‐win” proposition for PIs and campus researchers. Of most importance are the advantages to a PI's research group using a shared, centrally administered computing system vs. a dedicated, locally housed and administered system.
Key Advantages Include:
- access to professionally‐administered computing resources, including system administration and user support, for the cost of the required compute hardware alone
- access to common ancillary infrastructure such as high performance interconnect fabrics, high performance data storage systems, and high bandwidth research networks
- guaranteed use of the number of cores/nodes contributed while having the ability to access idle cores across the entire cluster, including the opportunity to run jobs at a scale larger than the purchased nodes
- short lead times for high priority job startup
- professional system administration of contributed cores, avoiding both the operational costs as well as the professional distractions of having researchers or graduate students perform system administration and support duties
- higher system uptimes due to professional administration
- improved grant competitiveness due to lower costs that must be included in the grant (i.e. hardware only, no operations costs); evidence of campus cost‐sharing; and more positive funding agency review of a cost‐effective centrally‐administered facilities
- provide access to a system to bridge the gap between technical workstations/small clusters and national‐scale systems
Advantages to Campus:
- reduced energy costs due to operational efficiencies of a shared versus standalone system and use of more energy‐efficient data center facilities
- housing of the cluster in a purpose‐built data center facility, avoiding costly modifications to space, HVAC, and power systems in existing office/laboratory space
- achieving economies of scale by avoiding duplication of secondary infrastructure such as large‐scale storage systems and interconnection to external research networks
- improved grant competitiveness, resulting in increased probability of award
Advantages to HPC/HTC 'Buy In' Provide:
- a computational base for multi‐UA campus grants and research as a UA system resource
- additional capacity for specific research needs or projects
- additional capability that is not supported in campus clusters
- an alternative to building additional campus capacity
- facilities that may be a cost‐effective alternative to building out new energy‐efficient data center space for other existing or planned shared cluster systems
Please note that for researchers that already have shared cluster facilities and local support staff, HPC/HTC is designed such that the local technical staff can remain as primary research group support.
The HPC/HTC 'Buy In'program is not designed to replace or compete with the very large‐scale resources at national NSF and DOE facilities. National resources are available at no financial cost to many researchers based on competitive proposal processes. The HPC/HTC 'Buy In' program is designed to meet the needs of many researchers with lower or medium‐scale HPC/HTC requirements, who want guaranteed access to compute resources and control of the scheduling priorities for their resources.
Site map: http://rc.arizona.edu/sitemap