Geekly Articles each Day
📜 ⬆️ ⬇️

In addition to Moore - who else formulated the laws of scaling computing systems

We are talking about two rules that are also beginning to lose relevance.


/ photo Laura Ockel Unsplash

Moore's Law was formulated more than fifty years ago. Throughout this time, for the most part, it remained fair. Even today, when moving from one process technology to another, the density of transistors on a chip increases about two times . But there is a problem - the speed of developing new technical processes is slowing down.

For example, in Intel for a long time postponed the mass production of its 10-nanometer processors Ice Lake. Although the IT giant will begin shipping devices next month, the architecture announcement took place about two and a half years ago. Also in August last year, the manufacturer of integrated circuits GlobalFoundries, who worked with AMD, stopped developing 7-nm process technology (we described the reasons for this decision in our blog on Habré).
')
Journalists and leaders of large IT companies have been predicting the death of Moore’s law for more than a year. Even Gordon himself once stated that the rule he had formulated would cease to act. However, Moore's law is not the only law that is losing relevance and on which processor manufacturers are equal.

Dennard's scaling law


It was formulated in 1974 by DRAM dynamic engineer and developer Robert Dennard (Robert Dennard) along with colleagues from IBM. The rule is as follows:

"By reducing the size of the transistor and increasing the processor clock frequency, we can easily increase its performance."

Dennard's rule enshrines the reduction of the width of the conductor (process technology) as the main indicator of progress in the microprocessor technology industry. But the law of scaling Dennard ceased to operate around 2006. The number of transistors in the chips continues to increase, but this fact does not give a significant increase in device performance.

For example, representatives of TSMC (manufacturer of semiconductors) say that the transition from 7-nm process technology to 5-nm will increase the processor clock frequency by only 15%.

The reason for the slow growth of the frequency are leakage currents, which Dennard did not take into account in the late 70s. When reducing the size of the transistor and increasing the frequency of the current begins to heat the chip more, which can disable it. Therefore, manufacturers have to balance the power allocated by the processor. As a result, since 2006, the frequency of mass chips has been set at 4-5 GHz.


/ photo by Jason Leung Unsplash

Today, engineers are working on new technologies that will solve the problem and increase the performance of microcircuits. For example, experts from Australia are developing a metal-air transistor, which has a frequency of several hundred gigahertz. The transistor consists of two metal electrodes that act as a drain and a source and are located at a distance of 35 nm. They exchange electrons with each other due to the phenomenon of field emission .

According to the developers, their device will allow to stop “chasing” the reduction of technical processes and concentrate on building high-performance 3D structures with a large number of transistors on a chip.

Kumi rule


He was formulated in 2011 by Stanford Professor Jonathan Koomey. Together with colleagues from Microsoft, Intel and Carnegie Mellon University, he analyzed information on the power consumption of computing systems since the ENIAC computer, built in 1946. In the end, Kumi made the following conclusion:

"The volume of calculations per kilowatt of energy with a static load doubles every one and a half years."

However, he noted that the power consumption of computers over the years has also increased.

In 2015, Kumi returned to his work and updated the research with new data. He found that the trend he described slowed down. The chip's average performance per kilowatt of energy began to double about every three years. The trend has changed because of the difficulties associated with cooling the chips ( p. 4 ), as with the reduction in the size of the transistors it becomes more difficult to remove heat.


/ photo by Derek Thomas CC BY-ND

Nowadays, new technologies for cooling chips are being developed, but it’s not yet possible to talk about their mass implementation. For example, developers at a university in New York have proposed using laser 3D printing to apply a thin thermally conductive layer of titanium, tin, and silver to a crystal. The thermal conductivity of such a material is 7 times better than that of other thermal interfaces (thermal paste and polymers).

Despite all the factors, according to Kumi , the theoretical energy limit is still far away. He refers to the study of the physicist Richard Feynman (Richard Feynman), who in 1985 noted that the energy efficiency of processors will increase 100 billion times. At the time of 2011, this figure increased only 40 thousand times.

The IT industry has become accustomed to the high growth rates of computing power, so engineers are looking for ways to extend Moore's law and overcome the difficulties dictated by the rules of Kumi and Dennard. In particular, companies and research institutions are looking for a replacement for traditional transistor technology and silicon. We will tell you about some of the possible alternatives next time.


What we write about in the corporate blog:



Our reports from VMware EMPOWER 2019 on Habré:

Source: https://habr.com/ru/post/454326/

More articles:


All Articles