Moore’s Law; Will it stop?

Harpreet Thandi

In 1965, Gordon E Moore, an electrical engineer from America, wrote an article in Electronics magazine. It suggested that every two years the capacity of transistors would double. Later his prediction was updated to processor power doubling every two years and is now known as Moore’s Law. He then became the co-founder of one the biggest creators of microprocessors that figure the speed of laptops and PCs.

This law has wider implications than simple processing power. Devices have become smaller and smaller. We went from a large mainframe to smartphones and embedded processors. This has resulted in a more expensive process where chips have become smaller.

In the larger scheme of things this two-year evolution is the underlying model for technology. It’s resulted in better phones, more lifelike computer games and quicker computers which we use every day. Maybe this effect came from goal setting: we must make processing power double every two-years, or maybe it was just a natural progression? Either way, Brian Krzanich-chief executive of Intel suggested this growth could be coming to an end but he still supports this; “we’ll always strive to get back to two years”. However, the firm still disproves the death of Moore’s Law, as future processors won’t be made so quickly. Technology users might realise their new phone or laptop is only a bit superior than the older model. There is a drastic need for Moore’s Law to be met again as this speed of development leads to more effective processors and save us so much money with efficiency.

To keep up with Moore’s law there have been some major compromises. Now we are at a crossroads, microprocessors are getting smaller and smaller but now they are reaching a fundamental limit due to their size. Transistors are a certain size for quantum effects to take place. “The number of transistors you can get into one space is limited, because you’re getting down to things that are approaching the size of an atom.”

A problem that started in the early 2000’s is overheating. As the devices have shrunk the electrons are more restricted and the resistance goes up dramatically in the circuits. This creates the heating problem in things such as phones and laptops. To counteract this the ‘clock rates’- the speed of microprocessors has not increased since 2004. The second issue is that we are reaching a limit the size and limit of a single chip. The solution is to have multiple processors instead of one. This means rewriting various programs and software to accommodate this change. As components get smaller they must also become much more robust and stronger.

Four and eight are standard quantities when it comes to the processors in our laptops. For example, “you can have the same output with four cores going at 250 megahertz as one going at 1 gigahertz” said Paolo Gargini-chair of the road mapping organisation. This lowers the clock speed of the processors also solving both problems at once. There are more new innovations being undertaken. However, many of these are simply too expensive to be effective.

According to the International Technology Roadmap for Semiconductors (ITRS) transistors will stop getting smaller by 2021. Since 1993 they have predicted the future of computing. After the hype in 2011 of graphene and carbon nanotubes, ITRS suggested it would take 10-15 years before these combine with logic devices and chips. Germanium and III-V semiconductors are 5-10 years away. The new issue is that transistors will not get smaller and move away from Moore’s Law.

Intel is struggling to make new breakthroughs. If they have not been resolved and they fall of the 2-year doubling target. However, there will be strong competition from their competitors. IBM have also started challenging them; a processor seven nanometres wide, 20 billion transistors and 4 times than today’s power. This will be available in 2017. “It’s a bit like oil exploration: we’ve had all the stuff that’s easy to get at, and now it’s getting harder, … we may find that there’s some discovery in the next decade that takes us in a completely different direction”-said Andrew Herbert who is leading a reconstruction of early British computers at the National Museum of Computing.

There is a new future for quantum computing. This works with qubits-quantum bits with values of 0 and 1. The nature of quantum mechanics can be to have multiple states in a system. We could get a quantum computer to work on multiple problems at once and come up with solutions in days that would naturally take millions of years traditionally.

  In May 2015 Moore spoke in San Francisco at an event celebrating the 50th anniversary of his article. He said “the original prediction was to look at 10 years…The fact that something similar is going on for 50 years is truly amazing…someday it has to stop. No exponential like this goes on forever.” At the time this was completely unknown that the total transistors in a computer chip would double every year. This has continued for a lot longer than expected and is now a major part of popular culture- Moore’s Law has become the underlying physical standard of the future that society has lived up to and has driven to meet.

#HarpreetThandi #Physics #Technology