HPC & Data Center

The world of TeraFlops and PetaBytes

For many years Pyramid has been working in the field of Cluster technology. Our highly power saving and highly efficient node platforms, based on Intel or AMD, in connection with years of experience in the area of Linux and networking, have made Pyramid become market leader. In 2008 Pyramid internationally reached a milestone of success. After years of investment and growing cluster installations for various research institutes, on May 22nd, 2008 the Albert-Einstein-Institute in Hannover put ATLAS on stream, the worldwide fastest supercomputer of the category Gigabit-Ethernet-Interconnect. Not only had Pyramid built the fastest Ethernet-based cluster, but could we also offer this customer an impressive price value ratio with € 55.000 per Teraflop/sec.

Pyramid Cluster in the MPI for Gravitational Physics Golm

Our HPC-business unit derives from a very strong engagement for Linux from the 90s. Linux is the exclusive no. 1 operating system regarding HPC on a technicalscientific and commercial level. Pyramid has reached international recognition with some successful HPC installations, at a size of up to 32-TeraFlop/sec. Especially regarding the price-value-ratio, Pyramid has an advantage through manufacturing in-house and omitting unnecessary node-components. Innovative power saving concepts and the early engagement in the field of the 80 Plus energyefficiency-initiative have been other major elements leading to Pyramid’s success in this field. Pyramid tailors offers for clusters to the individual needs and requirements of customers and therewith builds an ideal basis for cluster concepts used in research institutes and universities.

• Individually scalable Linux Cluster
• Project and budget orientation
• Cooling liquid on request
• Virtual-server-technology
• Latency optimized 10 GBit-Ethernet-
• Interconnect
• Infiniband Interconnect