Moore's Law and disorder Monday, 13 June 2016

News article written by Corbett Communications. The statements made or opinions expressed do not necessarily reflect the views of Engineers Australia.

It is forecast that the computer industry will be a much more complicated place following the relative predictability of Moore’s Law and its demise, according to The Economist's Technology Quarterly in March.

But let’s go back in time before we go to the future. In 1971, a small company called Intel released its first microprocessor, the 4004, a chip measuring 12 square millimetres containing 2300 transistors. Six years earlier, Gordon Moore, who would become one of Intel's founders, wrote a paper about the number of electronic components crammed into an integrated circuit doubling each year. A decade later it was noted that the rate was reduced to every two years which came to be known as Moore's Law which held for 44 years - until 2014. Intel's Xeon Haswell E-5 has over five billion, just 22 nanometres apart.

Pentium chip designer Dr Bob Colwell told The Economist that things like clever design and cunning programming will be useful but a collection of one-off ideas won't make up for the lack of an underlying exponential. Progress is set to become less predictable, less rapid and narrower than the industry has gotten used to over the last four decades. Chip designer at ARM Greg Yeric is quoted as saying, "as Moore’s law slows down, we are being forced to make tough choices between the three key metrics of power, performance and cost". Not all end uses will be best served by one particular answer.”

Computers have moved from entire rooms to desktops, laptops to mobile phones in the palm of your hand but will be in everything from automated homes and cars, all forms of electronic entertainment to wearables including clothing. And some in Silicon Valley say a lot of creativity will be unleashed over the next decade seeing performance improved and existing technology used in different ways.

But the chips in the computers we use or wear won't be in the devices themselves but in the data centres of the rapidly growing 'cloud'. One large  US investment bank has revealed that cloud computing investment grew by 30% in 2015 and looks to continue expanding at that rate for the next few years. The race for market share could well disrupt the industry but to what extent? Big companies like Amazon and Facebook design their own data centres but use hardware from firms like Cisco and Intel while Microsoft has started designing its own chips.

Moore's Law was about speed and making computers faster, but with that trajectory splintering there are other ways to make better computers laying other fertile ground for innovation. Colwell believes consumers don't care or even know about Moore's Law, they simply want products they buy to keep improving and be useful. But we may see a return toward more sustainability and longevity in products as consumers won't be lining up for the next iteration of their smartphone within the two-year period framework of Moore's Law as next year's model will cost about the same and store the same amount of stuff as the last one you had.

Electrical engineer Andrew Huang wrote in IEEE's Spectrum about the slowing down of Moore's Law as good news for small businesses while big businesses have been profiting from the two year leaps in technology by having the researchers and funding to anticipate several iterations of a product. The deceleration could also allow for standardisation of platforms - something undreamt of a decade ago - making life easier for software engineers.

"Personally, I’m looking forward to the changes—including the return of artisan engineering, where elegance, optimisation, and balance are valued over raw speed and feature creep," he said.

Even if Moore’s Law ends soon or abruptly, Huang believes that electrical engineers and consumers alike should stop worrying and be happy about what’s in store.


Image: Computer scientist Gordon Moore in a screenshot from the Scientists You Must Know video, by the Chemical Heritage Foundation, in which he briefly discusses Moore's Law.