PAC Meeting Summary, April 30, 2021
Review survey results and come to consensus around top three priorities. These were the results:
PAC ranking of concerns (highest to lowest)
- Expiration of credits (5 yrs.)
- Support for unfunded computational explorations (e.g., graduate student research and thesis work)
- The new model will increase costs on grants, which is a competitive disadvantage
- Negative sponsor reaction to the new model
- Wait times in the new model as compared with dedicated resources
- Storage costs
- Concern that F&A will be charged on PACE after initial 2-year waiver
- Need for a way to pool credits (library model) across larger groups
- Uncertainty around handling of cost share
- E-routing check box confusion
PAC ranking of speculative benefits (highest to lowest)
- Indirect cost waiver on PACE services
- Flexibility to leverage new hardware releases instead of being restricted to hardware purchased at a specific point in time
- Projected lower overall cost to PIs versus the equipment model
- Indirect cost waiver on commercial cloud services
- Backfill queue
- Insulation from failure of compute nodes
- Free basic tier for all PIs (10000 CPU hours and 1 Tb project storage)
- Improved reporting on usage per project, user, job, etc. for data-informed resource management
- Capacity and usage planning allows for pre-purchase of resources, drastically reducing procurement wait times for new PIs
- Higher utilization of existing hardware improves environmental impact
Extensive and robust discussion of top concern – expiration of credits from Rich Building Refresh. Various options considered. Settled on the following proposal:
- Credits from Rich Building Refresh will last 7.5 years (instead of 5 years)
- PI may request 1 year no-cost extension
- Neil announced his departure to New Zealand.
Action Item:
Review CoS memo.
