Processors use billions of tiny transistors, electric doors that turn on and off to make calculations. They take the power to do it, and the smaller the transistor, the less power is needed. "7nm" and "10nm" are measurements of the size of these transistors – "nm" being nanometers, a tiny length – and are a useful metric for judging the power of a particular processor.
For reference, "10nm" is Intel's new manufacturing process, which will begin in the fourth quarter of 2019, and "7nm" generally refers to the TSMC process, which is based on AMD's new processors and the A12X chip. Apple.
So, why are these new processes so important?
Moore's Law, an old observation that the number of transistors on a chip doubles each year, while the costs are halved, long retained, but slow recently. In the late 1990s and early 2000s, the size of transistors was halved every two years, resulting in significant improvements on a regular basis. However, the additional reduction has become more complicated and we have not seen any reduction of Intel's transistor since 2014. These new processes constitute the first significant reduction in a long time, especially from Intel, and represent a brief reminder of the law from Moore.
With the delay of Intel, even mobile devices have had the chance to catch up, with Apple A12X chip manufactured on the 7 nm process of TSMC and Samsung with its own 10 nm process. And with AMD's next processors on TSMC's 7-nm process, this gives them a chance to surpass Intel's performance and bring healthy competition to Intel's monopoly on the market, at least until that Intel's "Sunny Cove" 10nm chips are commercially available.
What the "nm" really means
Processors are made using photolithography, where an image of the CPU is etched on a piece of silicon. The exact method used is usually called the process node. It is measured according to the reduced size of the manufacturer for the manufacture of the transistors.
Since smaller transistors consume less power, they can perform more calculations without overheating, which generally limits processor performance. It also reduces the size of the dies, which reduces costs and can increase the density to the same sizes, which means more cores per chip. 7 nm is twice as dense as the previous 14 nm node, allowing companies like AMD to release 64-core server chips, a significant improvement over the previous 32 cores (and Intel's 28 cores).
It is important to note however that Intel is still on a 14 nm node and that AMD is about to launch its 7 nm processors very soon, that does not mean that AMD technology will be twice faster. The performance does not match exactly the size of the transistor and, on such a small scale, these numbers are not so accurate. The way each semiconductor foundry measurement can vary from one source to another, so it is better to take them more as marketing terms used to segment the products rather than as exact measurements of the power or size. For example, Intel's next 10nm node should compete with TSMC's 7nm node, despite the fact that the numbers do not match.
Mobile chips will see the most important improvements
Node reduction is not just about performance; this also has huge implications for laptop chips and low-power laptops. With 7nm (versus 14nm), you can get 25% more performance with the same power or performance for half the power. This means a longer battery life with the same performance and much more powerful chips for smaller devices, because you can integrate twice as much performance with the limited power target. We have already seen the Apple's A12X chip crushes old Intel chipsAlthough they are only passively cooled and packaged in a smartphone, this is the first 7-nm chip to hit the market.
Knot reduction is always good news, as faster, more energy-efficient chips affect almost every aspect of the world of technology. 2019 will be an exciting year for technology with these last nodes, and it's good to see that Moore's Law is not quite dead yet.