We’ve barely got the 14nm die shrink in our hands but Intel has claimed it already knows how to produce 7 nanometre chips. Intel’s continuing battle with Moore’s Law continues apace, with fabrication of central processing units using the 10nm process beginning next year, and the jump to the smaller 7nm node already planned after that.
The traditional methods of die shrinking for manufacturing microprocessors will eventually come to an end, but Intel believes there’s some life in it yet with these 7nm plans. “After 10nm, comes 7nm and we feel very confident that we have line of sight of how to manufacture that node,” said Renee James, president of Intel, during a Q&A session with stockholders at the company’s annual meeting. “Overall, we think, we are in a very good shape.”
In terms of when these will arrive; 10nm chips are expected in 2016, and 7nm should be coming our way in 2018. Obviously Intel is keeping details on how it intends to make the move a secret, but it’s thought the processor giant is considering moving away from silicon, and changing the structure of transistors altogether.
CPUs and chips in general have been integrated circuits created on a two-dimensional structure. This method is reaching near-capacity, and Intel has been making use use of three-dimensional FinFET architecture in order eke out more performance and efficiency. It’s thought optimisations of FinFET architecture are going to prove helpful in die shrinks, but ultimately Intel will be looking towards more advanced processes to solve the looming problem. This includes technologies such as multiple patterning and extreme ultraviolet lithography, both capable of printing transistor features smaller than the optics of the manufacturing system can even detect - a seeming impossibility.
Regardless of where Intel takes it from here though, it’s important that Intel is working on, and confident of, the 7nm fabrication process. This ensures we should be seeing a steady increase in processor power over the 3-4 years, as Skylake and then Cannon Lake surface.
0 comments:
Post a Comment