What is the world’s fastest computer? Like most people, you are probably thinking it is the latest Mac. The fact is, consumer computers are nowhere near being the world’s fastest computers. That mantle is for supercomputers. Supercomputers have high levels of performance compared to general-purpose computers. They are capable of performing millions of operations per second. Mostly, government research organizations fund the developments.
Until recently, China’s Sunway TaihuLight supercomputer was the fastest in the world. The US could not have that. IBM and the Department of Energy collaborated to create IBM Summit. As of the time of writing this article, IBM Summit is the fastest computer in the world.
IBM Summit is a supercomputer specially designed for data and AI. The US Department of energy commissioned IBM to build the supercomputer in 2014. The goal was to create a computer up to 10 times faster than Titan (a 2012 supercomputer that was once the world’s fastest computer).
During the 4 years of developing the supercomputer, IBM helped to figure out and find solutions to a significant number of technological barriers. The result was a machine that can be defined as an AI with deep learning capabilities, which perform quadrillion operations per second (200 petaflops).
IBM Summit’s Specifications
The performance of supercomputers is measured in floating operations per second (FLOPS). IBM summit is capable of 200 petaflops. Thus, putting it in less technical jargon it’s 200 quadrillion calculations per second. To put that into perspective, the previous record holder, Sunway TaihuLight, is capable of about 100 petaflops.
So, what makes Summit so fast?
The speed is attributed to its 9,216 Power9 IBM CPUs and 27,648 NVIDIA Tesla GPUs. Summit is capable of transferring 25 GB of data per second between nodes. And, it has 250 petabytes (250,000 TB) of storage.
The Summit supercomputer, located at Oak Ridge National Laboratory, was built to tackle some of the world’s most significant challenges. It will be used in the fight against cancer and help in the understanding of diseases. Its deep learning capabilities will help researchers identify next-gen materials to be used in semiconductors, batteries, building materials and so on.