About the UConn Health HPC

The UConn Health HPC facility provides access to over 12,000 processor cores, 77 TB of memory and over 14 PB of storage housed in a secure, state of the art data center supported 24/7 by a dedicated staff of professionals. These resources are available to any University of Connecticut of UConn Health researcher for use in support of their research activities.

HPC Clusters

Most users access the HPC resources through one of the HPC clusters. The clusters are designed to support a wide range of reserach applications, including computational biology, bioinformatics, machine learning, and data analysis.

Mantis Cluster

Mantis is the newest cluster, intended to replace the Xanadu cluster. Mantis is recommended over Xanadu for nearly all users.

Xanadu Cluster

Xanadu is a legacy cluster that has been replaced by Mantis. A limited number of resources will remain in Xanadu for a period of time to allow users to transition to Mantis. This cluster will eventually be removed from service so users are encouraged to transition to Mantis as soon as possible.

Storage

The UConn HPC facility provides access to over 14 PB of storage through a variety of storage systems which are accessible from HPC clusters as well as virtual machines.

Virtual Machines

Users can request access to self-supported virtual Linux or Windows instances for running applications that are not compatible with the HPC clusters. These virtual machines are hosted on a high performance virtualization infrastructure and can be accessed from anywhere with an internet connection.

Hardware

Compute Clusters:

4 Dell R620 nodes with two, 10-core Intel(R) Xeon(R) CPU E5-2660 v2 processors, 256 GB RAM, and 10 GB interfaces
17 Dell R660 nodes with two, 64-core Intel(R) Xeon(R) Platinum 8462Y+ processors, 512 GB RAM, and 10 GB interfaces
21 Dell R730 nodes with two, 18-core Intel Xeon E5-2697v4 processors, 256 GB RAM, and 10 GB interfaces
27 Dell R740 nodes with two, 40-core Intel Xeon Gold 6138 processors, 192/384 GB RAM, and 10 GB interfaces
3 Dell R815 nodes with four, 12-core AMD Opteron 6172 processors, 396 GB RAM, and 10 GB interfaces
1 Dell R905 node with four, 6-core AMD Opteron 8435 processors, 256 GB RAM, and 10 GB interfaces
9 Dell C6145 nodes with four, 12-core AMD Opteron 6172 processors, 256 GB RAM, and 10 GB interfaces
8 Dell R6525 nodes with two, 64-core AMD EPYC 7662 64-Core processors, 256 GB RAM, and 10 GB interfaces
21 Dell R7525 nodes with two, 64-core AMD EPYC 7453 processors, 256 GB RAM, and 10 GB interfaces
24 Penguin MH61-HD3-ZB nodes with Intel Xeon Gold 6230 processors, 192 GB RAM, and 10 GB interfaces
15 Supermicro AS 2024US-TRT nodes with two, 64-Core AMD EPYC 7713 processors, 512 GB RAM, and 10 GB interfaces
1 Penguin MZ92-FS0-00 node with eight, 6-Core AMD EPYC 7352 processors, 2048 GB RAM, and 10 GB interfaces
2 Dell XE8545 AI nodes with two, 64-Core AMD EPYC 7763 processors, 2048 GB RAM, and 10 GB interfaces
1 Supermicro 6049GP-TRT node with four, 28-Core Intel Xeon 8176 processors, 1500 GB RAM, and 10 GB interfaces
2 Quanta Cloud S76 nodes with NVIDIA Grace Hopper Superchip, Grace CPU, Hopper GPU processors, 480 GB RAM, and 10 GB interfaces
15 NVIDIA A100 80gb PCIe GPUs, 6 NVIDIA A10 GPUs, and 2 NVIDIA M10 GPUs

Virtualization Infrastructure

16 Dell PowerEdge R640 hosts with 1280 CPU cores and 10.8 TB RAM running VMWare 8 hosting 300+ Windows and Linux virtual machines with SSD high OPS performance cache tier

Datacenter Infrastructure

UPS generator backed power with redundant cooling
3×40 Gbe dark fiber connection to off-site DR location

Network (100+ Gbe)

Full non-oversubscribed 10/40/100/400 GbE datacenter network core layer
BioScienceCT Research Network – 100 GbE to CEN, Internet2, Storrs
HPC Science DMZ – low latency, 80 Gb-capable firewall

Storage

Over 22 PB of storage including 775 TB Quantum, 3.4 PB EMC Isilon, 10.5 PB Qumulo scale-out clusters, 2 PiB Spectra tape backup and archive along with 2.3 PB Amplidata, and 3.1 PB Scality Geo-Spread cloud storage