The Barbora NG supercomputer was installed in 2026 with a theoretical peak performance of 1,858 TFlop/s.
The computing power consists of:
- 144 standard computational nodes; each node is equipped with two 96-core Intel Xeon 6952P processors
The Barbora supercomputer is built on the Bull Sequana XH architecture, and for cooling its standard compute nodes, the direct liquid cooling technology is used. The computing network is based on Infiniband NDR technology. The SCRATCH data storage system for computations has a capacity of 500 TB.
technical information of the barbora NG supercomputer
| put into operation | summer 2026 |
|---|---|
| Theoretical peak performance | 1,858 TFlop/s |
| operating system | RHEL 9 |
| compute nodes | 144 |
| cpu | 2 x Intel Xeon 6952P; 96jádrový, 2,1 GHz |
| RAM per compute node | 768 GB |
| storage | 25 TB / home, 500 TB / scratch (NVMe, 187 GB/s sequential write performance, 384 GB/s sequential read performance) |
| interconnect | Infiniband NDR 400 Gb/s |
Learn more at docs.it4i.cz.
cluster utiliSation of the barbora NG supercomputer
View current Barbora NG supercomputer data: DASHBOARD Barbora NG
Gallery
