The University of Naples “Parthenope” participates in the DAGonCAPIO project with the research group belonging to the laboratory High-Performance Scientific Computing (HPSC) of the Department of Science and Technology (DiST). The research activity of the HPSC group focuses on HPC, cloud computing, and GPU virtualization applied to computational environmental sciences and to AI for environmental applications.
UNP’s contribution to the project
The UNP Team, in addition to the leadership of the project, will coordinate the design and development activities of the workflow engine and the use case. The dissemination activity of the results will be partly managed by the UNP Team and then have a predominance of the contributions of the corporate partners starting from the release of the prototype of DAGonCAPIO.
The UNP team involved in the DAGonCAPIO project will leverage the experience gained in the RAPID, ADMIRE, and MytilAI projects.
RAPID
The RAPID project’s (H2020) main focus was the exploitation of high-performance accelerators (GPUs) on low-computing power devices through the split-driver offloading paradigm. The project aimed at a secure and unified model to accelerate heterogeneous devices via private, public, and hybrid cloud.
ADMIRE
The ADMIRE project (H2020, EuroHPC) seeks to overcome the typical performance limits of current storage technologies in HPC systems. It focuses on computational and storage elasticity by proposing a software-only solution to create an active I/O stack that dynamically adjusts computation and storage requirements. During ADMIRE, the HPSC Team curated the aspects related to the synthesis between workflow engines and ad-hoc file systems, actively contributing to the co-creation of various software components depending on the use case.
MytilAI
The MytilAI project’s (Research Agreement with the Campania Region) aim was to develop a modeling system for the accumulation processes of pollutants in mussels based on AI and HPC technologies. The Team will benefit from the experience gained in the MytilAI project to develop the use case.