There are users who for different reasons maintain this almost transparent organization when discussing each performance. This is overridden in just as hot a topic as nanometer comparisons between Intel and AMD, or even in the companies themselves.
In the end, the average user concentrates precisely on nanometers as if it were a real measuring device that affects CPU performance. Let's look and think if this is true.
Nanometers vs Transistor Density vs Performance
Processor lithography is directly related to the number of transistors that each chip can accommodate, but as we all know, there are plenty of relationships that are added to this whole process.
A large number of transistors are required for certain types of wafers and lithographic processes to rearrange or a new type of transistor. We are currently experiencing it and its destination EUV and so it will be in the future with other pigeons.
Paradigm switches and transistors and their occurrence are notable changes to the instructions of each circuit capable of operating, turning on or off depending on the same design. But while everything said so far is true, CPU performance, so to speak and expand the mind, is not the result.
Good testing has been provided by Intel, where its 14 nm having a much lower density than AMD of 7 nm gets very similar results depending on the performance. In addition and further to the above example, within 14 nm we find in the architecture Broadwell-E
Do they rely directly? What are the influences?
So, even though there is an inevitable relationship, the size of the transistors is not directly related to performance and does not affect you as such. It involves a very large number of transistors per square centimeter reaching the technological advance and in many cases several problems that result in overconsumption, poor frequency waves or overuse.
Continuing with the idea of a normal and high level, we cannot assure you that the relationship between the two is directly dependent. Excellent performance is mainly due to advances in CPU architecture, which together with a higher frequency than the usual high-quality lithograph processes can cause startups.
But logically this does not mean that high performance (rather than non-CPI) is achieved in large quantities, but because the lithographic process is able to offer more in a small space, with less usage and higher speeds, but above all, higher performance is achieved in architecture.
You have to separate the terms well, because they are easy to fall into the trap of overcrowding and operation where the most significant improvement in the design of the structures is the reduction of the lithographic process. Something similar to what happened to Zen 2 with AMD: step at 7 nm (95 MTr / mm2) significant structural changes in architecture, medium waves and equitable use.
