Milestone: Introducing Moore’s Law
Date: December 2, 1964
Location: San Francisco Bay Area
People: Gordon Moore
In a modest address to a local professional association in 1964, computer scientist and chemist Gordon Moore made a prediction that would define the world of technology for more than 50 years.
The final version of this prediction became known as “Moore’s Law” and would drive progress in the semiconductor industry for decades.
you may like
Although they were called laws, they were predictions based on economic trends and industry trends rather than physical laws of nature.
At the time of his talk, Moore was director of research and development at Fairchild Semiconductors, and his ultimate goal was to sell more chips. Back then, computers were huge machines that took up entire rooms, and the practical use of integrated circuits known as microchips was somewhat limited.
The silicon transistor, the mainstay of computing in computers, had been invented just a decade earlier, and the integrated circuits that allowed computers to be made smaller had been patented just five years earlier. In 1961, electronics company RCA built a 16-transistor chip, and by 1964 General Microelectronics had built a 120-transistor chip.
Moore witnessed this dramatic progress and realized that mathematical laws seemed to govern it. This mathematical correlation was later given the name “Moore’s Law” by others.
Moore explained this principle to the Electrochemical Society in 1964, but it received widespread attention when he was asked to write an editorial in Electronics Magazine the following April. In it, he boldly predicted that as many as 65,000 components could be packed onto a single chip. This was an unprecedented number at the time. Considering that one company announced 4 trillion transistor chips in 2024, this is now an attractive small number.
In 1968, Moore co-founded chipmaker Intel. There his rule of doubling went from casual observation to motivation for innovation.
Despite its name, Moore’s Law was never an ironclad law. In 1975, Moore reduced the rate of transistor doubling to every two years instead of every year. This more modest multiplication rate became the official Moore’s Law and would remain in place for many years to come. This constant drive toward more computing power and miniaturization has made possible virtually every modern electronic device, from personal computers to smartphones.
you may like
For years, people predicted that this law would become obsolete, but for quite some time this law has proven to be surprisingly resilient.
“The fact that I have been able to continue, [Moore’s law] “This has surprised me more than anything else in a long time,” Moore said in a 2016 Electrochemical Society interview. “There always seems to be an impenetrable barrier ahead, but as we approach it, people come up with solutions.”
But eventually that principle no longer holds true. It is not clear exactly when Moore’s law died out. In its standard form, the standard could have died in 2016, as it took Intel five years to move from 14 nanometer size technology to 10 nanometers. Moore saw this coming years before his death in 2023 at the ripe old age of 94.
In the end, Moore’s “law” had to disappear because it contradicted the actual laws of physics. As transistors became smaller and smaller, quantum mechanics, the physics that governs the very small, began to play a larger role. The world’s smallest transistors could face the problem of “quantum tunneling,” where electrons in one tiny transistor tunnel into another tiny transistor, allowing current to flow through a transistor that is supposed to be in the “off” position.
As a result, chip manufacturers are considering designing chips using new materials and new architectures. The following Moore’s law could also apply to quantum computers, which use quantum mechanics as a feature rather than a computational bug.
Source link
