ARC High Performance Computing

Background
The Advanced Research Computing (ARC) program provides research services and support for computational and storage technology. The goal of this contract is to refresh the general-purpose High-Performance Computing (HPC) cluster to support a broad range of research workloads at the University.

The current HPC cluster consists of approximately 21,000 CPU cores and 504 GPUs, many of which are faculty-owned. It has an existing storage infrastructure based on GPFS with a capacity of roughly 3 PB.

Work Details
The contractor will be responsible for refreshing the HPC cluster to scale the storage capacity to at least 8 PB while continuing to use Slurm as the primary job scheduler.

All submissions must include a quote with a Bill of Materials (BoM) breakdown for each configuration proposed. The following attachments must be referred to for detailed requirements:
1) Supplier Questionnaire Excel (Attachment E)
2) Pricing Excel (Attachment F)
3) Requirements document (Attachment G).

Period of Performance
The contract period will commence upon award and is expected to conclude by the end of the project timeline as specified in the proposal.

Place of Performance
The work will be performed at the University’s facilities where the HPC cluster is located.

Bidder Requirements
Bidders are required to submit an 'intent to bid' before question submissions can be accepted by the system. This decision can be withdrawn at any time up until the bid closing.