default header

ASTRON boosts performance of the LOFAR radio telescope with a new GPU-cluster

Published: Fri, 12/07/2019

We have successfully replaced the GPU-cluster of the Low Frequency Array (LOFAR). The new powerful cluster, COBALT2.0 is now ready for use and replaces the old ‘heart’ of LOFAR. This makes new applications with LOFAR possible and makes LOFAR ready for the future. COBALT2.0 runs the next-generation correlator and beam-forming application of LOFAR, this requires a powerful cluster. The design is future proof and COBALT2.0 is amongst the most energy efficient (greenest) High-Performance correlators.

In February, ASTRON acquired COBALT2.0, Bossers & Cnossen supplied the cluster in partnership with Dell Technologies. After a period of software-, conformity- and compliance tests COBALT2.0 is now ready for use and enables the LOFAR Mega Mode (LMM) for the telescope.

René Kaptijn, about the project: "I’m really proud of how smooth the tender and project went and all the hard work the team has done. They have put in a lot of effort to make this happen. Because of the good people and their work, the project ran very smoothly and the system is now in production!"


LOFAR Mega Mode

LOFAR is the world’s largest radio telescope, and is addressing key astronomical questions in a unique way. It observes the sky at very low radio frequencies (10 – 250 MHz). LOFAR tracks the sky using software, unlike traditional telescopes that physically need to point their large dishes. It is one of the scientific cornerstones of international radio astronomy. Thanks to COBALT2.0, LOFAR will be able to carry out simultaneous observations to address multiple scientific questions in parallel. This new capability is known as the LOFAR Mega Mode. COBALT2.0 will simultaneously process data streams from LOFAR stations spread across Europe.


Technical specifications of COBALT 2.0


  • 26 Intel Xeon Gold Skylake-SP 6140 CPUs
  • 96 GByte RAM per CPU
  • 26 NVIDIA Tesla V100 GPUs
  • 100 Gbps EDR InfiniBand Technology

Compute Power:

  • 63 TFLOPS of CPU compute power
  • 360 TFLOPS of GPU compute power


COBALT2.0 can receive more than 1 Terabits per second (more than 30 billion samples) from the LOFAR stations and process them in real time. It offers significant extra capacity both in terms of network bandwidth and compute power. The extra capacity will be used for data quality improvements.

The increased efficiency and capabilities of LOFAR resulting from COBALT2.0 will lead to new discoveries to help us unravel the mysteries of the Universe.