Moore or Less?

Here at Zeitgeist we find supercomputers rather fascinating and the developments in computing power never cease to amaze us. Back in 2010 we wrote about the military investment to advance technology to the next level via a step change and the grants provided by the US Military to build so-called Exascale computers.
We had to look that one up too. An Exascale computer is one that works at a level entitled an Exaflop – the equivalent of one million, trillion calculations per second. How fast is that? Very. Current supercomputers are only managing a paltry 1,000 trillion calculations per second currently charmingly known as a Petaflop.
Advancing computing power is rather evolutionary currently and follows Moore’s Law – the rule that states that the number of transistors which can fit on a single piece of silicon will double every 18-24 months. The limitations of the current evolution approach are the mushrooming power, management and structural issues that crop up as components shrink.
The secret to the solution therefore is to develop chips that use substantially less power per calculation and a step change in energy efficiency. An Exaflop supercomputer — roughly 30 times more powerful than today’s fastest machines — will need to rely on new technologies. One using CPUs alone would suck up 2 gigawatts of electricity. That’s the output of the Hoover Dam….
Last year President Obama created the National Strategic Computing initiative (NSCI) because the world recognizes that the ability to manage vast quantities of data is becoming increasingly important and the US needs to keep pace with the likes of China and India. So it’s full steam ahead to encourage innovation in the design and production of the new chips.
Building ever faster machines will require rethinking the way these machines are built. We hear of step change often in business, usually as a result of some dramatic need to turnaround a performance or to leap frog the competition. However, although conceptually people understand step change, they invariably end up ‘polishing the same apple’ and working in increments because it is known, understood and frankly, safer in terms of risk.
The challenge laid down by the NSCI is to think radically and actually develop a step change in performance – something we can learn from in business. Forward thinking, progressive companies establish incentive programs to encourage innovation and creative solutions, and capture the corporate capability to achieve change.
The European Commission’s Human Brain Project hopes to use Exaflop computers to better understand the workings of the human mind and so despite all of these staggering advances, it is clear that even the most powerful computer fails to come close to the basic ability of the human brain – that brain that we put to good use on a daily basis……