IT Infrastructure · Data Centre

HPC VEGA

Efficient Infrastructure for EuroHPC Supercomputing

Lokacija
Maribor, Slovenia
Sector
Research Infrastructure · HPC
Client
IZUM – Institute of Information Science, Maribor · Eviden
NTR Role
Turnkey Infrastructure Engineering, Integration & Lifecycle Support

NTR designed, engineered and built the complete system and utility infrastructure required for HPC Vega, one of the first EuroHPC petascale supercomputers. The scope covered the data centre, supercomputer installation, networking, supporting power and cooling infrastructure, commissioning and ongoing maintenance, with a strong focus on operational reliability, energy efficiency and waste-heat utilisation.

01
The Challenge

HPC Vega combines exceptional computing performance with a peak electrical load of 830 kW, placing significant demands on power delivery, heat removal and year-round operational reliability. The project also had to be integrated into an existing facility with spatial, structural, acoustic and architectural constraints, while meeting a tight six-month delivery schedule.

02
The Solution

NTR delivered an integrated infrastructure architecture centred on direct water cooling for 95% of the compute load, supplemented by high-temperature air cooling. External heat rejection is provided by adiabatic dry coolers operating without compressors, while the system enables waste-heat recovery for building heating. Power distribution, short-term UPS support, 2N-redundant storage and support systems, and the in-house NTR DCIM HPC monitoring platform complete the critical infrastructure environment.

03
The Result

The completed infrastructure enables stable operation of HPC Vega while limiting cooling-related electrical consumption to approximately 50 kW despite an 830 kW peak compute load, resulting in a PUE below 1.06. Waste heat is recovered for building heating, while centralised monitoring supports efficient operation and lifecycle management. The complete infrastructure, including stress testing and operational handover, was delivered within six months.

Key Project Facts

830 kW
HPC compute-section power supply capacity
< 50 kW
Cooling-related electrical consumption
Year-round operation
< 1.06
Power Usage Effectiveness (PUE)
95% direct water cooling
High-temperature liquid-cooling architecture
5% supplementary air cooling
Waste-heat recovery
Recovered heat used for building heating
6 months
Infrastructure delivery
Including stress testing and operational handover

More about the project

Want the full technical picture?

Download the complete HPC Vega case study for detailed system architecture, equipment specifications, cooling and power infrastructure, redundancy, security and monitoring.

Download case study
PDF · Case Study
All references

What can we do for you?

Planning a similar project?

Tell us what you need to achieve. We'll help you define the right infrastructure for your environment, workloads and future growth.

No detailed specification required. A challenge, an idea or an upcoming project is enough to start the conversation.