About


CITerra is a facility that houses High Performance Computing (HPC) Clusters. These clusters are a coordinated group of commodity server computers running the Linux Operating System. This configuration (a Linux-based HPC Cluster built using commodity hardware) is often called a Beowulf.

You can learn more about CITerra on the following pages:

History


Pangu


In 2004, Caltech division of GPS built their first cluster for division use. This was a small cluster of machines handbuilt by members of the GPS division. This cluster proved to be extrememly useful for calculations and simluations in the division.

With the success of the first cluster, it was decided to go bigger. A new cluster, named pangu, was built. This new cluster started as 160 Dell Poweredge 1850s. This was in place while a new facility was built. Soon after the cluster was exapnded further to 1024 compute nodes. This cluster was rated as the 41st fastest supercomputer in the worls according to the Top 500 list. This cluster was active for a couple of years and was extrememly successful.

During this time, however, technology had gotten better. Multicore technologes had become the standard. To increase the performance and decrease total power usage a new cluster was designed.

In 2007, the older cluster was sold back to Dell, and newer quad core machines were purchased to replace it. The new cluster would be 512 Poweredge 1950s with 2 quad core sockets. The high speed interconnect and filesystem remained the same. This new cluster gave the division 3 times the performance of the previous cluster while using a little more than half the total energy than previously. This new cluster rated 29th on the top 500 list.

This cluster was in place until December of 2011. It was used in the research that led to over 140 published papers. All told, It was an extreme success for the division.

Pangu was named for the first living being and creator of all in Chinese mythology.



Fram



In 2010, a new cluster started to be designed. This time interconnect, filesystem, and servers would all be replaced. Once again technoloy had changed over the years. GPGPUs were now on the landscpe for doing hugh performance computation. The new cluster would be designed with that in mind.

In Decmember 2011, the older pangu was removed and the new cluster was installed. This new cluster consists of 314 HP DL390sl G7s. 60 of these nodes contained 3 Nvidia GPGPUs for a total of 180. At the same time, The high speed interconnect was also changed. This new cluster would useQuad data rate Infiniband. This interconnect increased the bandwidth greatly, and decreased the latency. The filesystem was replaced with a Lustre filesystem connected via Infiniband. This filesystem could perform at 9.5 GB/s.

Fram was the name of a ship commissioned by Fridtjof Nansen for exploration of the arctic. Nansen was a Norwegian explorer, scientist, diplomat, humanitarian and Nobel Peace Prize laureate. Nansen believed he could reach the North pole by starting in the east and using the natural currents to take him there. The Fram was specifically commissioned to be frozen into the ice without damage and capable of housing enough fuel and provisions for 12 men for up to 5 years.

Fram was used by Otto Sverdrup (related to the famous oceanographer Harald Sverdrup) from 1898 to 1902 to chart the Arctic islands and to sample the rocks, animals, and plants. Then from 1910 to 1912 it was used by Roal Amundsen on the expedition that was the first to reach the South Pole.