New HPC Usage Model
In 2019, an HPC cost-recovery model was introduced to coincide with our newest HPC cluster, Wendian. Upon review of the cost-recovery model, it was found that it did not fully satisfy certain federal funding requirements while mixing shared-use and paid-use in the same model. In 2021-22, the Mines Research Council appointed a HPC sub-committee to explore options for a viable cost-recovery model for current and future HPC systems. The goal of the new model will ensure that HPC systems are consistent with federal funding requirements (2 CFR 200’s), the HPC business plan is sustainable, and consistent with forthcoming cloud options.
Starting in April 2023, HPC usage will be billed as a pay-for-use model at a rate of $0.02 per core-hour instead of passing capital equipment costs to the end-user. IT will continue to fully subsidize class usage for HPC. To prepare for the April launch of the new model, several efforts are underway. The HPC support team will be creating additional tools such as a self-service portal, billing and usage tools, and budgeting guidance.
Please contact us at hpcinfo@mines.edu if you have specific questions related to the change.
Upcoming Changes
- Easier project management and self-service using ColdFront
- Expansion of HPC resources with university partners
- Additional cloud resources in AWS
Supporting Documentation
- HPC Funding and Usage Proposal (pdf)
- Fall 2021 HPC Business Plan Discussion – Meeting One (pptx)
- Fall 2021 HPC Business Plan Discussion – Meeting Two (pptx)
- HPC Business Plan Strategy – February 2023 (pptx)
FAQ
Q: What if I purchased priority access under the old model?
A: Those that purchased priority access under the old model should reach out to the HPC team to discuss options (hpcinfo@mines.edu)
Q: What HPC clusters does this apply to?
A: This change is only applicable to Wendian. Those that use Mio are unaffected by this change.
Q: Do I have to pay for data storage?
A: We are only addressing cpu core-hour usage at this phase.