The infrastructure also hosts user-developed AI-based codes, alongside computational resources supporting a broad range of numerical codes—particularly for materials science, enabling simulations across multiple scales and accuracy levels for material study and characterization.

This capacity has enabled ENEA’s participation in numerous national and international projects. The CRESCO Computing Lab forms part of ENEAGRID, a distributed HPC infrastructure serving several hundred users across ENEA research groups and external partners, spanning fields such as alternative energy, nuclear fusion/fission, climate science, new materials, AI/ML, manufacturing, bioinformatics, transport, security, and cultural heritage.

Node interconnection runs via a Mellanox-NDR network at 200 Gb/s bandwidth. The system comprises three partitions: a GPU partition (1920 CPU cores + 983040 GPU cores) with 15 nodes featuring dual 64-core Intel Xeon 8592+ processors, 512 GB DDR5 RAM, and 4 Intel Max Series 1550 GPUs (128 GB HBM2e, 3.2 TB/s bandwidth, 636 GB/s Xe Link interconnect); an HBM-CPU partition (1680 CPU cores) with 15 nodes of dual 56-core Xeon 9480 processors, 1024 GB DDR5 plus 128 GB HBM; and a standard CPU partition (97280 cores) spanning 760 nodes with dual 64-core Xeon 8592+ processors and 512 GB DDR5 RAM.

ENEA-CRESCO Computing Lab grants access to CRESCO8, an HPC facility ranked 228th in the June 2025 TOP500 list with 5.9 PFlop/s peak performance.