Priority Tier is an opt-in YCRC compute service for computational jobs. Computations submitted to a Priority Tier partition will precede all pending jobs in the corresponding standard tier partitions in the scheduling queue to provide a “fast-lane”. To ensure researchers receive the best and consistent value for their funds, Priority Tier partitions are composed of the YCRC’s newest generation of nodes and GPUs.
All computation on the “standard” tier of partitions (e.g. day, week, mpi, gpu) as well as private nodes and scavenge partitions do not incur any charges. Researchers can escalate specific important computations to Priority Tier and run less urgent workloads in the Standard Tier.
The Service Unit rate structure for the Priority Tier is derived to closely match the prorated cost of a similar dedicated node over a 5 year expected lifetime. As such, Priority Tier is an alternative to purchasing dedicated nodes. For the same cost, researchers can realize 100% of the value of their funds (compared to inevitable idle time on dedicated nodes) and always be computing on the YCRC’s latest resources (vs dedicated nodes which get more out of date every year and are decommissioned after 5 years).
More detailed information on Priority Tier, including how to request access, rates, and how to use the Priority Tier partitions can be found in our User Guide.
The Priority Tier service model is effective December 1st, 2024. This compute charging model was developed in close collaboration with faculty, YCRC staff and university administrators to ensure the YCRC service charging models support the researchers who rely on our systems and the needs of the University.
Estimate Costs
To assist with cost estimates and budgeting, we provide a Cost Calculator.