The quantum computing team at the CRS4 Centre for Research, Development and Advanced Studies in Sardinia, composed of researcher Giuliana Siddi and technologists Marco Moro and Michele Muggiri, has succeeded in emulating the operations of a 30-qubit quantum computer thanks to the computing infrastructure acquired with the recent investment of over €3 million.
Giacomo Cao, sole administrator of CRS4: 'We can proudly say that it has been possible to emulate the operations of a 30-qubit quantum computer thanks to the use of the computing centre's new infrastructure, which reaches an installed power of 4.3 PetaFlops and a storage space of 6.3 PetaBytes. This result is comparable with those recently achieved by other national and international research centres. Naturally, we hope to have a 100-qubit quantum machine at our disposal as soon as possible, which would enable us to further increase our computing performance and become one of the best-performing centres internationally in this field'.
Cao said: "On the one hand, CRS4 would like to thank the Region of Sardinia for its continuous financial support for investments that significantly highlight the island, and on the other hand, it responds with concrete facts to the community's need for growth and development.
As is well known, quantum computing could be applied in many fields, the most important of which include: machine learning; artificial intelligence; chemistry; pharmacology; finance; health; and cybersecurity. CRS4 therefore aims to continuously improve the performance of the algorithms it develops, in terms of operations performed in the shortest possible time and with real-time results.
Features of the CRS4 computing centre
Computing power 4.3 PetaFlops; disk space 6.3 Petabytes; more than 250 computational nodes for a total of more than 8000 cores and low-latency, high-bandwidth network connections; distributed file system with parallel RDMA access via Infiniband; hybrid CPU/GPU clusters for quantum computing simulation/emulation 16 nodes with 16 Nvidia T4 GPUs available; 4 nodes with 12 Nvidia A100 Tensor Core GPUs available; distributed file system with parallel access, implemented with GPU Direct Storage (Nvidia GDS) for direct access between GPU memory and storage; distributed file system with parallel access; 100Gb/s core ethernet; Infiniband low latency 100Gb/s network.