Clusters are a group of computers that work together on a common task. As the rapid growth of CPU and networking technology, more and more people are using clustered High Performance Computing (HPC) to replace traditional big iron. Compared to supercomputer, PC cluster features superior performance/price ratio and its easiness of assembling which has phenomenally lowered the threshold to enter the high performance computing world.

Cluster types

There are few types of Cluster include High Availability (Failover), Load Balancing and Parallel Processing. Our HPC system infrastructure is based on the Parallel (Distributed) Processing method. The advantage of Parallel Processing HPC is high availability and performance for applications. A large processing task is broken down into many smaller tasks and worked on by many small servers (nodes).

Where Clustered HPC can be used?

HPC offers high availability, load balancing, manageability and scalability. It can be used when mission is critical and processing is required to be quick. For example, industries which involve advanced engineering like Automotive, Aerospace, Biomedical, Chemical, Military, Oil, Power generation, Semiconductor, etc, they require sophisticated analysis tool such as Computational Fluid Dynamics.

Traditional supercomputers has always been used in seismic analysis. Large amounts of data collected, data mining, calculations, and complex renderings required by the bioinformatics industry (related to genetic research) are well-suited based on price/performance for the HPC.

Additionally, in the film industry, rendering quality animation/graphics/game and streaming media require high processing speed. Beowulf clustered HPC can, also , be used in Life science, Exploration Geophysics, Bioinfromatics, Finance and Internet to work on projects in fields such as protein folding, neural networks, quantum mechanics, fluid dynamics, astrophysics, evolutionary biology, molecular dynamics, genetic research, statistics and economics. Nix HPC provides supercomputer-level processing power at an affordable price.

Example: ChryslerDaimler has purchased 108 dual-processor Beowulf Linux workstations to run car-crash simulations