Production supercomputing
![]() |
|||
|
Estimated sustained computing capacity (in teraflops) available to the NCAR community over the past 10 years. On average CISL provides substantial capacity increases every 16 months to meet the scientific research demands of the community. |
|||
![]() |
|||
|
Average utilization (defined as the percent of compute nodes assigned to users at any one time) of the production computing platforms provided by CISL. During the time period shown, the major IBM systems (blackforest, bluesky, bluevista, and blueice) were delivered on time and reached nearly 90% utilization within a few months. Keeping scientific production at a high level meets NCAR's goal of providing reliable and robust computational services to the community. |
|||
To fulfill its mission, CISL provides robust, reliable, effective, and efficient production supercomputing, coupled with state-of-the-art storage, data analysis, and data visualization services and tools for its community of users.
CISL's goals are to provide equitable access to reliable computing resources with minimal user wait times, while maximizing resource utilization. These goals are achieved by balancing the allocation and scheduling of resources with a well-tuned queue structure and prescribing single-job resource limits.
CISL works to increase the computational capacity available to the community on a regular basis. Deploying the IBM POWER5+ system blueice during FY2007 is another example of such an increase. Peripheral resources, which complement and supplement the high-end computing environment, are upgraded and enhanced as appropriate to match the growth in compute capacity.
In FY2008, CISL will install an IBM POWER6 system as the second phase of the ICESS procurement. The power and cooling demands of this new system will require the removal of both the POWER5 system bluevista and the POWER5+ system blueice. Despite the removal of these two large systems, both computational capacity and capability available to users will increase significantly.
This work fulfills NCAR's strategic priority of "Enhancing capability and capacity of NCAR supercomputing" and is supported by NSF Core funds, including CSL funding.

