Research Computing Resources at URI are hosted in the Tyler Hall data center located on-campus in Kingston, RI. This data center is operated by the University IT Services under the leadership of URI’s chief information officer, Gabriele Fariello. The data center is configured with emergency power supplies and high-speed internet access through the URI central power and network connections. Central cooling is provided to maintain a climate-controlled environment for the operation of HPC systems. Dedicated IT staff provide operation, management and monitoring support to ensure the smooth operation of the data center facility.
The URI Center for Computational Research (https://ccr.uri.edu) includes an IT Research Computing Services team that consists of several specialists: Dr. Kevin Bryan manages HPC systems, Drs. Michael Puerrer and Cecile Cres are Computational Scientists, Christian Vye handles the support for statistics and other software applications, licensing and 4 graduate students offer consultation and training. In addition, the Center also administers the URI AI Lab that is a faculty-run resource that focuses on AI related research collaboration, education and outreach. The Center is led by Dr. Gaurav Khanna who serves as Director and is also a Professor in the Physics Department. The Center is advised by a committee of URI faculty representing widely different areas of research. The Center also has a supporting external advisory group consisting of regional experts drawn from the Massachusetts High-Performance Computing Consortium (MGHPCC) in the area of research computing.
Existing high-performance computing (HPC) resources
Andromeda high-performance computing cluster
The Andromeda cluster is a large scale HPC cluster that largely serves groups of contributing users with some support for the broader URI user community. It is established with contributions from individual researchers and currently has 47 nodes, providing 1704 cores with nodes having 64GB (8), 128GB (29), 256GB (3), 512GB (4), or 768GB (1) memory. These are connected via an Omni-Path 100Gbps network, with shared storage to a 1.1 PB hard disk.
URI campus central data center
URI HPC clusters are located in a Data Center located in the Tyler Hall Building on the Kingston campus. The 2,300 sq. ft. Data Center provides a climate-controlled environment with 90 tons of cooling for the operation of HPC systems and is equipped with 160 kVA of UPS battery backup as well as a 450 kW Generator that provides emergency power supplies. The Tyler Hall Data Center is operated by the URI Information Technology Services (ITS) and is monitored and maintained by ITS staff.
Collaborative high-performance computing (HPC) resources
The Massachusetts Green High-Performance Computing Center (MGHPCC) is a collaboration of the five major research universities in Massachusetts including Boston University, Harvard, MIT, Northeastern and UMass. The collaboration built a dedicated data center in 2012 located in Holyoke, MA that hosts the research computing infrastructure of these universities. The URI Center for Computational Research has multiple collaborative programs with the MGHPCC, allowing for URI researchers to get access to the following MGHPCC resources:
UMass – URI collaboration: UNITY cluster
URI and UMass have built and operate a new shared HPC/AI environment at the MGHPCC. Researchers have access to this “UNITY” cluster via the InCommon Federation. Details of the cluster may be found at the UNITY cluster portal (https://unity.uri.edu). The cluster currently offers ~500 nodes with 25,000 compute-cores and 1,500 AI graphics processing units (GPUs). The nodes include both Intel and AMD multi-core processors and Nvidia GPGPUs for HPC and AI/ML computations.
MIT – URI collaboration: SuperCloud
URI researchers have access to the MIT Lincoln Labs “SuperCloud” resource via the InCommon Federation. Details of the cluster may be found at the SuperCloud cluster portal (https://supercloud.mghpcc.org). The cluster currently offers 200+ nodes, each with 40 Intel Xeon cores, 2 Nvidia V100 GPGPUs and 384GB memory and 400+ nodes with 48 Xeon cores and 192GB memory.
Harvard/BU — URI collaboration: NESE
The North-East Storage Exchange (NESE) is a storage collaboration led by Harvard and Boston University (https://nese.mghpcc.org) funded by NSF hosted at the MGHPCC. URI researchers have access to this facility for their data backup and tape archival needs via UNITY.
Harvard/BU — URI collaboration: NERC
The New England Research Cloud (NERC) is an academic Cloud computing resource led by Harvard and Boston University (https://nerc.mghpcc.org) hosted at the MGHPCC. URI researchers have access to this facility for their data Cloud computing needs via the InCommon Federation.
Active research data storage: OSN
The OpenStorageNetwork (OSN) is a national distributed storage collaboration (https://openstoragenetwork.org) funded by NSF, with a node hosted at the MGHPCC. URI researchers have access to this facility for their active research data storage and collaboration needs. Access to URI researchers is enabled through the InCommon Federation.
High-speed access (10 Gbps) between the MGHPCC and URI is enabled via a collaboration between OSHEAN and UMassNet.
Artificial Intelligence Lab
The Artificial Intelligence Lab contains:
- Nvidia DGX-1, a high-performance GPU (graphics processing unit) server
- 6 Lambda Tensorbooks and one DELL PC, run Ubuntu OS and preinstalled with AI development frameworks, TensorFlow, PyTorch, Caffe, Caffe 2, Keras, CUDA, and cuDNN
- Processor: Intel™ Core® i7-7700HQ; GPUs: 1x NVIDIA GeForce GTX 1070; Memory: 32 GB DDR4 RAM; Storage: 1 TB NVME SSD
- 6 Spiderbot: Hiwonder Hexapod Programming Robot for Arduino
- 6 Programmable Dancing Robots
- Samsung Flip 55” Interactive Digital Flipchart