To use this website fully, you first need to accept the use of cookies. By agreeing to the use of cookies you consent to the use of functional cookies. For more information read this page.

Jamie Balfour'sPersonal blog

As I'm sure anyone who read my blog for technology related stuff will know, Moore's Law is a fundamental 'law' that defines that the speed of computers will double every two years. It's not entirely the case but it holds true for the majority of systems produced.

The law is more of a theory of a computer scientist called Gordon Moore, one of the founders of what is now Intel. It was theorised in 1965 and what it really stated was that the number of transistors that can be crammed in to one integrated circuit will double every two years. 

Intel call this a tick in their 'tick-tock' cycle. Examples of Intel CPUs include the Sandy Bridge range (tick) when compared with the Ivy Bridge range (tock). Both of these ranges were based on the Sandy Bridge architecture. The Haswell architecture which was the next tick could fit twice as many transistors in the same size of integrated circuit, following Moore's Law. 

But on the release of Broadwell, which was based on the Haswell micro-architecture and was the successor 22nm Haswell, we have arrived at transistors that are only 14nm in size, compared with Haswell's 22nm transistors this change is huge. The next step after 22nm according to the International Technology Roadmap for Semiconductors will come in at 10nm. Currently, Skylake, which is the current range of Intel APUs and is a tick in the tick tock cycle, is facing several problems with going further. For the very first time in the history of Intel's tick-tock cycle, there are going to be two ticks (tick-tick-tock). Why you may ask?

The answer is that Moore's Law no longer holds true with current fabrication techniques. In fact 10nm is posing such problems that it has been delayed until 2017. Cannonlake (formely Skymont), which will be the tock in the cycle will succeed the successor of Skylake, codenamed Kaby Lake. It will drop the size to 10nm. From here on however, there is considerable worry about whether or not we can go any further. We may see for a few years that computers cannot get any more powerful. What worries me is that the companies may use this to make money out of us at no extra cost to them (since the technology will change but the systems will be no more powerful).

So what's the next step then? Quantum computing? Chemical based computing? Biological computing? Good question. 

For the foreseeable future I would imagine that quantum computers will be the future, since they currently already exist. What worries me about the future is how will devices we currently use (such as the world wide Internet) interface with these new devices? I worry greatly about this and how the transition will turn out.

Posted by jamiebalfour04 in Tech talk
moore
law
moore's
gordon
intel
cpu
transistor
logic
gate
Powered by DASH 2.0 (beta)
Code previewClose