Prometheus – the fastest Polish supercomputer

Prometheus, a new supercomputer with its 2,4 Pflops of theoretical performance has been listed at 39th position. This top-efficient supercomputer located in Poland has been built by Hewlett-Packard according to requirements and partial design provided by Cyfronet. To date, Prometheus is one of the biggest installations of this type in the world and also the first one in Europe, based on the latest technology of direct liquid cooling.

Prometheus has been launched this April. It consists of more than 2,200 servers based on HP Apollo 8000 platform, combined with super-fast InfiniBand network with 56 Gbit/s capacity. Its energy saving and high-performance Intel Haswell latest-generation processors offer more than 53,000 cores. These are accompanied by 279 TB RAM in total, and by two storage file systems of 10 PB total capacity, and 180 GB/s access speed. Prometheus has been also equipped with 144 NVidia Tesla GPGPUs.

Prometheus and Zeus tandem on TOP500 again!

Prometheus and Zeus tandem on TOP500 again! Photo credit – Cyfronet

To illustrate Prometheus’ speed, one would have to harness the power of more than 50,000 first-class PCs, in their strongest configuration, additionally connected to super-fast network and managed by special software, to catch up with its abilities – says prof. Wiatr, Director of ACC Cyfronet AGH. Thanks to the innovative technology of direct liquid cooling, Prometheus is also one of the most energy-efficient computers in its class in the world. Its PUE factor is comparable with some of the largest data centres in the world, such as Google or Facebook. Furthermore, liquid cooling allows for extremely high installation density, therefore Prometheus, weighing of more than 30 tons, covers 18m2 area and is placed on 20 racks only. To achieve the same computing power in case of Zeus (Prometheus’ predecessor), it would have to take about 160 racks. Baribal, the predecessor of Zeus, with computing power of 1,5 Tflops was placed on 8 racks. To achieve the computing power of Prometheus it would take as many as 12 000 Baribal’s racks – points Cyfronet’s Director.

Prometheus has been installed in a high-tech computing room, exclusively adapted for its operation. The supercomputer’s proper functioning is additionally supported by the accompanying infrastructure, including such systems as guaranteed power supply with an additional generator, modern air-conditioning and gas extinguishing. Prometheus can serve for:

  • data results analysis,
  • numerical simulations,
  • (big) data processing
  • advanced visualisations provision.

Zeus is still among the best

Zeus is still noted on a high, 386th position on the TOP500 list. This gives it the 5th position in Poland. As a reminder – Zeus consists of four partitions adapted to diversified needs of users:

  • a “classic” supercomputing cluster,
  • a set of servers equipped with GPGPU and FPGA graphics accelerators,
  • a virtual SMP computer with a large pool of shared memory running vSMP software,
  • a cluster of nodes with large memory pools (so-called “fat nodes”).

Zeus has got more than 25,000 computing cores with total theoretical computational power of 374 Tflops. This is accompanied by 60TB of RAM and 2.3PB of disk storage. Zeus is built of over 1,300 individual blade servers, interconnected via high-speed 40Gb/s InfiniBand network. A significant part of Zeus is equipped with NVIDIA GPGPU cards. This allows selected algorithms used in scientific applications to benefit from multiple acceleration. Zeus uses Scientific Linux as an operating system.

The two serving scientists together

Access to the Prometheus and Zeus resources is done via the PLGrid infrastructure – Polish nationwide computing infrastructure for scientific research in silico (performed with the help of computers). Dedicated computing environments, so-called domain grids, and specialised IT platforms facilitate conduction of increasingly complex research problems. Modelling of energy demands, drug and new material design, or simulation of complex metallurgical processes is just a fraction of research issues studied within nearly thirty domain grids operating in the PLGrid infrastructure.

The research portfolio carried out with the help of the Zeus and, recently, Prometheus is quite reach. It includes: prediction of 3D protein structures, study of semiconductor nanostructures and catalytically activity molecules as well as effective biosensors. Computations are used to study the behaviour of galaxies in a wide range of electromagnetic spectrum, for nuclear magnetic resonance modelling for the purposes of structural analysis of molecular systems, antidots in quantum world, for structural characteristics of human telomeres and complexity of the financial markets. Scientific computations do not include simulations only. Computing power is utilised by Polish scientists also within international projects like CTA, EPOS, LOFAR and Large Hadron Collider in CERN. With the help of dedicated software packages the supercomputers perform analyses of large and dispersed data sets as well as provide advanced visualisations.

In 2014, on Zeus only, almost 8 million jobs have been executed topping nearly 13,000 years of CPU time. The vast majority of computational tasks, performed until recently on Zeus only, were multicore, thus significantly reducing time of computations. Now, thanks to Prometheus, it is possible to run simulations also for those research issues for which the computing power of Zeus has so far been insufficient.

Supercomputers from Poland on TOP500 list

The November 2015 edition of TOP500 spots also other supercomputers installed in Polish computing centres and research institutes. The full list of Polish supercomputers on TOP500 is as follows:

  • 39 – Prometheus, ACC Cyfronet AGH,
  • 128 – Bem, WCNS, Technical University Wrocław,
  • 163 – Tryton, TASK, Technical University Gdańsk,
  • 218 – NCNR, Świerk
  • 387 – Zeus, ACC Cyfronet AGH.

Tianhe-2 still fastest in the world

Tianhe-2 (Milky-Way-2) kept its first position on TOP500 for the sixth time with 54,9 Pflops computational power (theoretical). The list is dominated by supercomputers from the US (199 units, 40 %) and China (109 units, 22%). On subsequent spots are: Japan (37 units, 7%), Germany (33 units, 7%) and UK (18 units, 4%). Poland (5 units, 1%) is 10th in country rank.

Source: Academic Computer Centre CYFRONET AGH