Research Computing Facilities

Posted by Research Computing Facilities il Monday, January 1, 0001

TOC

Research Computing Facilities (RCF) is a management structure for High Performance Computing, GPU Computing and Cloud Computing computational resources transversal to the Computational Intelligence and Smart Systems (CI&SS, http://cisslab.uniparthenope.it), Computer Vision and Pattern Recognition (CVPR, http://cvprlab.uniparthenope.it) and High Performance Scientific Computing-SmartLab (HPSC, http://hpsclab.uniparthenope.it), as well as in the Neptun-IA multidisciplinary laboratory dedicated to the applications of artificial intelligence at the sea (http://neptunia.uniparthenope.it), of the Department of Science and Technology of the University of Naples “Parthenope”.

  • Enabling research is at the heart of RCF’s mission. The calculation tools made available by RCF have contributed to the advancement of critical investigation in many fields, from mathematical and physical to social and humanistic sciences. In addition to cutting-edge hardware and advanced software managed by RCF, the computation experts (machine learning, big data, computer vision, signal processing, intelligent systems, high performance computing, GPU computing, cloud computing, workflows), the developers and the administrators of the HPC systems assist the researchers involved in other projects that require computation such as, for example, those relating to the Center for Marine and Atmospheric Monitoring and Modeling. Support for research projects is carried out through consultancy and training with the aim of helping to effectively implement hardware, software and visualization resources.

  • purpleJeans, a high performance computing cluster dedicated to the production of “research products” forms the fourth generation core of RCF’s advanced computational infrastructure. The clusters for high performance computing of previous generations (blueJeans, 2007; greenJeans, 2009), were definitively deactivated between 2018 and 2019. The redJeans cluster (2012) is dedicated to teaching and is undergoing restructuring. blackJeans is a high-performance computing cluster equipped with a GPU managed by RCF on behalf of the Center for Marine and Atmospheric Monitoring and Modeling (meteo @ uniparthenope, http://meteo.uniparthenope.it) and is dedicated to the production of meteo-oceanographic simulations and forecasts and the management of environmental data collected by the network of dedicated sensors (eg Weather Radar in X band). purpleJeans includes a large pool of servers, software, and storage that researchers can use to increase the efficiency and scale of their computational science. The RCF provides resources for distributed and shared memory computing, as well as emerging technologies including accelerators and big data management systems. RCF resources are free for researchers from the Department of Science and Technology of the University of Naples “Parthenope”, as long as they comply with the rules of use established by the RCF user policies. For more information, see the Introduction page.

  • RCF has availability of high-end 3D graphics processing tools and visualization hardware, virtual reality, support for interactive supercomputing and customized remote viewing tools for the data stored on the purpleJeans compute cluster.

  • RCF organizes seminars on a variety of research-relevant topics. To date, these educational / informational meetings have included introductory, intermediate and advanced seminars on programming languages, overview of data management tools and best practices, and sessions focused on using purpleJeans and other RCF resources.

  • RCF supports the computational aspect of external projects. If a funded research project (i.e. that has funding available) whose scientific manager (PI, principal investigator) is not affiliated with the CI&SS, CVPR, HPSC laboratories, needs support for high performance computing activities, the use of technologies related to machine learning, artificial intelligence, computer vision, big data processing, data visualization and, in general, other enabling technologies belonging to the skills of the laboratories associated with RCF, can purchase its own hardware (CPU, GPU, Storage) to be included in the resources managed by RCF. Hardware purchased from funds from funded projects is in the exclusive availability of the project for the duration of the project (unless otherwise agreed). At the end of the project, the computational resources will remain available to RCF and its users.

Requesting an account

All teachers and researchers belonging to the Department of Science and Technology of the University of Naples “Parthenope” who belong to one of the CI&SS, CVPR, HPSC laboratories are entitled to obtain an RCF account by completing a request for a PI account.

A PI account allows a user to log in to RCF systems; request allocations of RCF resources and grant access to other users; and delegate responsibilities to the collaborators of the University of Naples “Parthenope” and other institutions.

Users who are not eligible for an IP account can have access to RCF resources through a PI. These users must complete a General User Account Request and identify the sponsor PI, who will be asked to approve the request before creating the account. General Users cannot apply directly for resource allocations or grant access to other users.

Teachers and researchers belonging to the DiST who are PIs of funded projects who contribute through the funding received by the project to the maintenance and / or increase of the computational resources offered by RCF can request a PI account even if they do not belong to the CI&SS, CVPR, HPSC through an assessment of the specific case.

The use of RCF systems is governed by the RCF user policy.

Requesting an allocation unit

An allocation is an amount of compute time and storage resources granted to PI. To apply for a PI account, please refer to “Request Account”.

The basic unit of computational resources is the service unit (SU), which represents the use of one core for one hour on the purpleJeans cluster. The grants are intended to support researchers who promote specific research or teaching objectives and are of limited duration.