HPC TeraStat2

Description

TeraStat2 (TS2) is the Department of Statistical Sciences’ general-purpose high-performance computing cluster, designed for solving mathematical and statistical models on Big Data. It consists of 12 computing nodes running Linux, each equipped with 2 AMD Epyc 7452 processors and 256GB of RAM, and 12 computing nodes equipped with two Intel Xeon E5-2630 8-Core processors (16 virtual cores) and 64GB of RAM. One of the latter is equipped with 512GB of RAM. Overall, the system provides 1,920 cores. The cluster also includes 76TB of shared storage and an InfiniBand QDR optical fiber network supporting a bandwidth of 40Gb/s between nodes and between nodes and storage. 

Usage

As a high-performance computing system, access to the infrastructure is provided in batch mode through job execution. Users wishing to run a specific experiment prepare a description of the experiment in a designated format, together with all materials required for its execution (e.g., input data). Once prepared, this description is submitted to the queue management system operating on TS2, which tracks the computing resources consumed by each user. As soon as sufficient resources become available, the user’s requested experiment is automatically executed. Upon completion, the user will receive an email notification and may reconnect to TS2 to retrieve the results.

Access Criteria

Access to TS2 is provided free of charge to all Sapienza staff members (professors, researchers, PhD candidates, and research fellows) who require it for projects involving a high-performance scientific computing component. For particularly intensive use of computing resources, an annual subscription with a non-commercial fee is available to contribute to the maintenance and upgrade costs of TS2. This subscription grants exclusive access to a predefined allocation of computing resources and is valid for one year. Different subscription plans are available, varying in the amount of allocated resources and the number of associated accounts.

Click here to request a free account

Software Availability

The TS2 staff is responsible for the installation, maintenance, and updating of a set of software packages and libraries of common interest to the general user community.
Users requiring software and libraries not initially available on TS2 may install them locally within their user environment or request installation by the staff, who will proceed where the request is deemed to be of general interest.
In all cases, only open-source software or software covered by a valid license may be installed.

Funding

The primary source of funding for TS2 was the University Medium-Scale Equipment Project entitled “TeraStat2: High-density supercomputing infrastructure for solving highly complex data-centric problems” (project code no. MA31916B8891453A), under the scientific responsibility of Prof. Umberto Ferraro Petrillo, funded in 2019 with an amount of EUR 55,000. Additional co-funding contributions were provided by the Department of Statistical Sciences, including funding from the Second-Level Master’s Programme in Big Data. Statistical Methods for the Knowledge Society and the Second-Level Master’s Programme in Data Intelligence and Decision Strategies. 

History

TeraStat2 is the evolution of the TeraStat cluster, originally developed by the Department of Statistical Sciences in 2011. Additional information about TeraStat is available at this link.