Megware

from Wikipedia, the free encyclopedia
Megware Computer Vertrieb und Service GmbH

logo
legal form GmbH
founding February 1, 1990
Seat Chemnitz , Germany
Number of employees approx. 50 (as of 2019)
Branch Supercomputers , computer clusters , high-performance computers , IT system solutions
Website www.megware.com

Megware Computer GmbH has its headquarters in Chemnitz-Röhrsdorf since 1996
Knot insert for the SlideSX®-LC
HPC cluster CooLMUC-3 with ColdCon® hot water direct cooling

The Megware Computer Vertrieb und Service GmbH (proper spelling: MEGWARE) is a developer and supplier of solutions in the field of high performance computing and cluster . It is based in Chemnitz (Saxony) and serves customers from all over Europe.

Company history

Megware was founded in 1990 in the garage of one of the three company owners. In the years that followed, the company, with its retail, wholesale and system house divisions, developed into a provider with 35 specialist retail branches in Central Germany. From 1991 to 1995 there was a very close partnership with Escom. With a few exceptions, all Escom branches in Central Germany were run by Megware. The demand for high-quality computers, servers and network technology in the years after German reunification had a significant impact on the company's economic development. Megware manufactured computer systems and other information technology devices under its own brand name. At the end of the 1990s a saturation of the private customer market became apparent, to which the company reacted and concentrated more on the development and manufacture of servers .

A major project with the Chemnitz University of Technology (TU Chemnitz) initiated the change in corporate strategy from 2000. Computer scientists from the TU Chemnitz and employees from Megware developed a computer cluster for high-performance computing in research and teaching. The Chemnitz Linux Cluster , also known under the name CLiC , reached 126th place in the Top500 list of the most powerful supercomputers in the same year . Hans-Werner Meuer , the initiator of this ranking, confirmed that this computer system had the best price-performance ratio in the world.

Business activity

Based on the success of this project and the demand resulting from it, today's business field of Megware established itself, which mainly comprises the development and sale of computer clusters . For these high-performance computers , the company develops racks (ClustRack®), chassis (SlideSX®, SlideSX®-LC), multiple power distributors (ClustSafe®), direct liquid cooling systems (ColdCon®) and systems for cluster management (ClustWare® appliance) from hardware developed in-house and advanced software tools that are supplied turnkey as a computer cluster.

On February 7th 2007 the successor of the CLiC was handed over to the TU Chemnitz. Megware installed the Chemnitz high-performance Linux cluster ( CHiC ) in partnership with IBM. In the Top500 list from June 2007, this system reached number 117.

In the summer of 2011, Megware delivered the Vienna Scientific Cluster 2 ( VSC-2 ) in Vienna. The VSC-2 was Austria's most powerful supercomputer with 135.6 tera FLOP / s and in June 2011 it was 56th in the TOP500 list. Also in summer 2011 the supercomputer named CooLMUC was put into operation at the Leibniz Computing Center (LRZ) of the Bavarian Academy of Sciences in Garching. Equipped with the ColdCon cooling concept, it is possible, due to the very high return temperatures, to both freely cool and reuse the resulting waste heat . The system is therefore extremely energy efficient . It is the world's first AMD cluster system with hot water direct cooling and is considered to be the forerunner of the SuperMUC . The computer system contains processors of the type AMD Opteron Magny Cours 6128 HE with a clock rate of 2.0 GHz; a total of 2816 cores are available.

In 2012, Megware delivered a powerful HPC system to Johannes Gutenberg University Mainz . The cluster was named MOGON and reached 81st place in the June edition of the TOP500. With 33,792 processor cores and 205 TFlop / s it was ranked 8th among the German HPC systems. Another TOP500 cluster was put into operation in Norway. ABEL is at the University of Oslo and, with 10,080 processor cores and 178.6 TFlop / s, achieved 96th place in the ranking.

The University of Bayreuth in 2013 awarded the contract for the supply of current high-performance technology to Megware. The high-performance computer btrzx3 consists of a total of 12 server cabinets. With a performance of 10224 cores and 97.6 TeraFlop / s, the HPC cluster was one of the most powerful computers in the world in June 2013 and entered the TOP500 list with 486th place. The HPC cluster is located in the laboratory and internship building of the University of Bayreuth. Due to the size of the system and the continuous operation, the temperatures are very high, so that this waste heat can be used to heat the entire building.

In the following year, Megware installed the ForHLR1 system equipped with powerful Intel Xeon processors at the Karlsruhe Institute of Technology (KIT). With a computing power of 206.3 TFlop / s and 10800 cores, the supercomputer reached number 242 in the Top500 list of the world's fastest HPC systems published in June 2014. Strategically important research areas of KIT, such as the environment, energy, nanostructures and technologies as well as materials science, will use the new research high-performance computer to deal with highly complex application problems on a new scale.

The bwForCluster MLS & WISO Production is a joint project of the University Computing Center (URZ) and the Interdisciplinary Center for Scientific Computing (IWR) of the University of Heidelberg and the Computing Center of the University of Mannheim (RUM) and was delivered and set up by Megware in 2015. The production part, with which permanent computing capacity is provided, is distributed between the computer centers of the Universities of Mannheim and Heidelberg and, with 241.1 TFlop / s, reached place 297 in the TOP500 list, which was presented in July 2015 at the International Supercomputing Conference (ISC) in Frankfurt was presented on the Main. The bwForCluster MLS & WISO is a computing cluster for the molecular life sciences (MLS) and economics and social sciences (WISO) and is part of the Baden-Württemberg state concept for high and high performance computing bwHPC.

In 2015, Megware was awarded the contract for the procurement of the Hummel HPC system at the Regional Computing Center (RRZ) of the University of Hamburg . The high performance computer is required for a wide variety of scientific projects. It is considered a central resource for high-performance computing needs, which is used by the scientific working groups of the University of Hamburg and operated by the Regional Computing Center (RRZ). The configuration of MEGWARE is based on the latest Haswell CPU generation . The computing power is 195.6 TFlops with 6304 processor cores, with which the HPC system took position 387 in the TOP500 publication from June.

In 2015, Megware, with the high-performance computers bwForCluster MLS & WISO Production , MOGON , ForHLR1 , Hummel and Abel, was the only German HPC company to place five systems in the global ranking among the 500 fastest computers in the world and thus ranked 9th in the world.

In 2017, Megware delivered the successor to the CooLMUC to the Leibniz Computing Center (LRZ). The CooLMUC-3 was characterized by high energy efficiency. Compared to its predecessor, in which various processor technologies were already cooled directly with hot water, all compute and login nodes, power supplies and Intel® Omni-Path switches have now been cooled with hot water.

A year later, Megware installed two more large systems. The Goethe HLR for the Goethe University in Frankfurt reached 449th place in the Top500 list with its 472 computing nodes and a computing power of 1.45PFlop / s Rpeak. At the University of Münster , PALMA II now supports users from the research areas of physics, chemistry, mathematics and biology with complex application problems of new dimensions. The computing power of the Münster supercomputer with approx. 1.1 quadrillion individual calculations per second was made possible by a network of 412 individual computing nodes.

In 2019, the CooLMUC-3 of the Leibniz Computing Center (LRZ) in Garching was expanded. The high-performance computer now consists of a total of 9 server cabinets, in which all computing nodes including all integrated memory modules, network interfaces and power supplies as well as all Intel® Omni-Path Switches are 100% directly cooled with hot water using the ColdCon® liquid cooling technology developed by Megware.

Megware installs computer clusters in industry, research and teaching with several hundred computer nodes, but also systems with 8 or 16 nodes. These systems are used there for numerical flow analyzes, crash tests and many other computationally intensive computer simulations in which natural phenomena are simulated with mathematical models and are used to predict changes in the system under investigation.

In addition to the main business HPC and Compute Cluster, MEGWARE is a provider of IT system solutions for banks, educational institutions, regional councils and other public institutions.

Selected references

  • Abel at the University of Oslo. The high-performance computer reached number 96 in the TOP500 list.
  • MOGON: The supercomputer from Johannes Gutenberg University in Mainz reached number 81 in the TOP500 list and was one of the fastest in Germany at the time of delivery. Users are mainly institutes and working groups from all over Rhineland-Palatinate, mainly from the field of natural sciences.
  • Vienna Scientific Cluster 2 (VSC-2): In 2011 the VSC-2 was the most powerful high-performance computer in Austria and reached number 56 in the TOP500 list of the fastest supercomputers in the world. The HPC system is equipped with over 1300 servers developed by Megware , each with two processors of the type AMD Opteron Magny Cours 6132 HE. In total, more than 21,000 processor cores are available for scientific calculations. With 136 teraflops, i.e. 136 trillion arithmetic operations per second, it provides five times as much performance as its predecessor, the VSC-1.
  • Max Planck Institute for Gravitation Technology Golm: Gigantic collisions of black holes are being simulated on the Damiana supercomputer in order to obtain predictions for the first time about the properties of the resulting gravitational waves.
  • Ludwig Maximilians University Munich (at the Leibniz data center ): Tier-2 of the LHC Computing Grid , D-Grid , et cetera
  • Max Planck Institute for Iron Research (MPIE): The scientists at the MPIE can use the computer to determine the structure and dynamics of newly developed materials, such as B. Examine steel, shape memory alloys, biomaterials and implants.
  • Gaz de France : The clusters are used to track down oil reserves, to drill and to pump natural gas back into the earth.
  • CERN (European Organization Of Nuclear Research Geneva (Switzerland)): Numerical simulations of quantum chromodynamics (QCD), whereby the properties of pions, nucleons and other particles and their interaction are calculated from the basic equations of the theory.
  • Pal Computer System in Abu Dhabi: The computer cluster Hydra was installed by Megware in 2005 and used in international tournaments as part of the competitive sport of computer chess .

Web links

Individual evidence

  1. ↑ CLiC: Overview ( Memento from December 5, 2006 in the Internet Archive ) In: tu-chemnitz.de .
  2. ↑ CLiC PIII 800 MHz In: top500.org .
  3. ↑ Chemnitz high-performance Linux cluster in operation In: heise.de .
  4. ↑ VSC-2 In: top500.org .
  5. ↑ The world's first AMD cluster with hot water cooling at the LRZ In: heise.de .
  6. ↑ MOGON In: top500.org .
  7. ↑ ABEL: New computing facilities for research at the University of Oslo In: uio.no/english .
  8. ↑ ABEL In: top500.org .
  9. ↑ The super brain in the basement In: frankenpost.de .
  10. ↑ btrzx3 In: top500.org .
  11. ↑ First stage of the parallel computer ForHLR is in operation In: kit.edu .
  12. ↑ ForHLR1 In: top500.org .
  13. ↑ bwForCluster MLS & WISO Production ' In: top500.org .
  14. ↑ New high-performance computing cluster in the top 500 In: pro-physik.de .
  15. ↑ Hummel: Hardware Configuration In: uni-hamburg.de .
  16. ↑ Hummel In: top500.org .
  17. ↑ List Statistics In: www.top500.org
  18. ↑ "CoolMUC-3" New energy-efficient supercomputer at LRZ in: badw.de .
  19. ↑ Goethe-HLR In: www.top500.org

Coordinates: 50 ° 52 ′ 6.8 ″  N , 12 ° 51 ′ 11.8 ″  E