Red Hat Delivers First Integrated Linux-Based High Performance Computing Platform

Red Hat recently officially released its High Performance Computing Platform, a large-scale cluster system, with support through Red Hat Network. This follows its announcement last November that it would deliver the platform. It will expand the high-performance computing market, previously largely confined to education, further into business. The platform combines the processing power of computers on a dedicated network to perform computations that one server cannot finish quickly. Applications requiring this level of computing power include biological calculations, 3D rendering, engineering calculations, and many others.

In Thailand, high-performance computing is still used mainly in education and research. A few years ago, a center was established to provide this immense processing capacity under a project called “ThaiGrid”.

Abroad, these systems are already widely used. The best-known fastest cluster at present is Road Runner (peak speed 1375.8 Tflops*), which has decisively overtaken the previous champion, Blue Gene (peak speed 596.4 Tflops*). IBM supports both projects.

* This is the floating-point instruction processing rate per second. A rough comparison would be 1 Gflops = 1 Ghz.

When I have time, I will write about how to install and configure HPC, or High Performance Computing.