Moore’s Law has governed progress in the computer industry for 40 years. Quite simply, it states that computational bang-for-buck doubles every 18 months. That means you will have twice the processing power at you fingertips in one-and-a-half years for the same outlay. It’s been achieved in various ways, but until about 5 years ago it was achieved almost solely by virtue of shrinking microprocessor architecture, cramming ever more, ever-smaller transistors on a slice of silicon. Then along came dual core, quad-core, hexa-core CPUs to keep the momentum going once the transistor-minituarization reached maturity and heat issues brought that other eternal law into play – the law of diminishing returns.
Lately SoC designs – such as Intel’s Sandy Bridge inclusion of the graphics core on the same wafer as the main CPU – have helped the electronics industry keep pace with the ultimate high-tech benchmark standard, laid down by the co-founder of Intel more than forty years ago. Less distance between the main system components means less time spent transferring data, and also frees up the processor bus – which fetches data from other parts of the system – for other tasks.
Relatively simple refinements such as larger caches also improve performance, since a good 80% of normal processor instructions or data are already stored in the cache from recent use. Add to this the discovery of new silicon doping agents to improve the efficiency of the base materials, and you have a recipe for progress that will last long after Intel’s current Ivy Bridge 22nm architecture has been replaced by 14nm and even 10nm technology, already under development in Intel’s labs.
But now theoretical physicist party-pooper Michio Kaku has predicted that in ’10 years or so we will see the collapse of Moore’s Law’. Kaku noted that Pentium chips today have miniaturized the transistor to 20-odd atoms’ width and once the architecture reaches 5 atoms across, the law of diminishing returns will kick in as thermal noise makes the components too unreliable.
Thermodynamics? Quantum physics? Or just hot air? Kaku expects the microprocessor industry will need to embrace new approaches such as molecular or quantum computers. Similar to the way the new Ivy Bridge architecture revolutionizes traditional binary-switching to use new tri-gate technology, whereby a single transistor can now record three individual states rather than two. Gordon Moore himself accepts Moore’s law will eventually fail,
‘but ever time I turn around, I’m fascinated how we’ve been able to extend it.’
Here is Mr Big Think Michio Kaku in video:
No related posts.
Sign up to the Gadgetmix Newsletter (free) for news and reviews mailed directly to your mailbox CLICK HERE