June 2012 was a pivotal moment in the history of computer development. It was the time when IBM Sequoia was fully deployed and became wholly operational, landing the top spot as the world’s fastest supercomputer.
IBM Sequoia was designed for the use of the National Nuclear Security Administration. As part of their Advanced Simulation and Computing Program, the organization was looking for a computer model that could run different applications for their experiments and research in astronomy, nuclear weapons, human genome, climate change and energy. These tasks require a great deal of power, speed and memory so the organization tapped IBM to do the job.
As early as 2008, talks were in order and IBM was looking into various possibilities. They first built a prototype called Dawn that utilized the Gene/P design. It was pretty successful although it was a bit slower. Thus, they began working on the actual model, including many variations and advancements in it. The whole system is built with over a million and a half processor cores. It has around 96 racks of computer nodes and runs on 1.5PiB memory. The computer runs on Linux and has a registered speed of 16.32 petaflops, the fastest on record. The reason behind this is the number of processor cores it contained.
What makes IBM Sequoia so highly-applauded is the fact that it can run real-time simulations and all the research and experiments on it can be run sustainably. It was perfect for the use of the Lawrence Livemore National Laboratory. To date, it has run several applications like the HACC cosmology simulation and the Cardioid code studying the electrophysiology for the heart. Despite all these capabilities, IBM Sequoia is said to be energy efficient. It consumes less power than let’s say, the previous holder of the world’s faster supercomputer – K-computer.
Steven's Computer Blog 