Moore’s Law takes its name from Gordon E. Moore, co-founder of Intel, who in 1965 estimated that the number of transistors on a processor would double every two years. This was later amended to 18 months.
The uncanny accuracy of Moore’s Law is at least partly due to a self-fulfilling effect; it’s been used by the industry to structure targets effectively.
The importance of Moore’s Law to the web hosting industry cannot be underestimated since it allows providers to boost server CPU, RAM and operating performance at lower costs which is passed back to the customer in the form of lower prices. You thus get more for less. What follows is a review of the Intel chip/transistor technology and where the technology is heading.
Moore’s Law is made possible by innovations in technology and chip design. Significant milestones have included the development of CMOS and advanced photolithography techniques. Advances in chip fabrication technology have allowed more and more ambitious designs to be realized.
A key development in recent years has been the introduction of chip designs featuring multiple cores: first dual-core chips, then quad core and upwards. Server chips in particular make heavy use of multi-core designs.
The latest multi-core server chips may have eight, ten or even more cores on a single die, allowing faster performance without overheating. Such chips require specialized programming: software must be written in a multi-threaded or multi-processor fashion.
Another major advance used in Intel’s server chips is the introduction of three-dimensional transistor technology.
Intel’s 3D Tri-Gate Transistor Designs use transistors that stand up vertically from the silicon rather than lying flat. This increases the number of transistors that can fit on a single die. Currently, tri-gate transistors can be made as small as 22 nanometers.
Difficulties arose around the problems arising from increasing numbers of transistors. These included more numerous flaws in Intel’s chips and the increasing power needed to run them. These issues were remedied to a great degree by the introduction of multi-core technology. An ongoing source of debate is the fact that components can only be made so small before problems arise at an atomic level.
Intel’s development roadmap uses a “tick-tock” model. A tick release consists of the same architecture as previous chips but on a smaller die; a tock release represents a change in architecture. Intel aims to release a tick or tock each year.
Through the use of innovative technologies such as three-dimensional transistors and multi-core chips, Intel has succeeded in bucking Moore’s Law. Current projections suggest that transistors are set to shrink still further, with 14nm transistors in development for 2013 and 10nm transistors expected to appear sometime after 2015. To continue this process, Intel will need to find new and innovative solutions to the problems ahead. While many have predicted the end of Moore’s Law, Intel always seems to be finding ways to make transistors smaller.
Author Bio: Jason Stevens from jason-stevens.com / Freelance web developer, tech writer and follower of cloud computing trends. Follow him on Twitter: @_jason_stevens_.
*UK2.net reserve the right to agree or disagree with our guest bloggers. Call it freedom of speech, but our guest bloggers are entitled to have an opinion. If you wish to agree or disagree, then feel free to leave a comment. Thanks for visiting our blog! If you wish to become a Guest Blogger for UK2, please contact our marketing department.