Moore’s Law is the great legacy of Gordon E. Moore, engineer, doctor in physics and chemistry and co-founder of Intel who passed away last weekend and to whom we pay a well-deserved tribute here as another of the “fathers” of modern computing. But what exactly is this law and how did it decisively influence modern computing?
What is Moore’s Law
It is a simple statement, but very ambitious. An empirical law where Moore predicted that the number of transistors per unit area in integrated circuits would double every year, with a minimal increase in costs and that the trend would continue for the next two decades. This statement, formulated in April 1965, was revised by the engineer in 1975 when he verified that this progression was impossible to fulfill and he redefined his famous law establishing that the doubling of transistors would occur approximately every two years.
Beyond the increased time in its compliance, this law has been a maxim for the development of chips at Intel and throughout the industry, since it defined the business strategy in the semiconductor industry, allowed the appearance of the microprocessor and later the personal computer. It should be noted that Moore’s Law does not only refer to computers but to any type of integrated circuit. It is a vital component for the entire industry and has ended up elevating the technological branch to the world leader.
The effect of the “law” went far beyond a simple increase in the number of transistors that had been invented by John Bardeen at Bell Laboratories in the United States in 1947 by allowing the technology to become more efficient with each generation. Just as important as performance was the cost reductionwith an inversely proportional relationship and as a result, the industry was able to develop new products and services.
Compared to the first single-chip CPU, the Intel 4004, a modern chip can multiply by thousands of times its performance and energy efficiency, while lowering its costs to 1/60 thousandths of the 4004. And what to say about other important aspects such as its size. The first semiconductor transistors were the size of a fingernail, and the supercomputers of the 1970s took up an entire room. Today, to see a single transistor, we would have to enlarge a single chip to the size of a house. When it comes to performance, a simple smartphone has more computing power than those supercomputers.
The origin of Moore’s Law
If you are curious to know where this law came from, nothing better than going to the explanations that the same engineer has been carrying out and that put us on the track of its origin since he founded Fairchild Semiconductors along with seven other pioneers.
“In the early 1960s we were still developing semiconductor technology and making it increasingly practical. It was a difficult technology to implement with the tools we initially had available. I became the director of R&D at Fairchild Semiconductor, managing the lab and looking at what we could do as we improved the technology.
So Electronics Magazine asked me to submit an article for their 35th annual issue predicting what was going to happen in the semiconductor component industry in the next 10 years. So I took the opportunity to analyze what had happened up to that point. This happened in 1964 I guess. I looked at the few chips we had made and realized that we went from a single transistor on a chip to a chip with about eight elements, transistors and resistors.
The new chips that arrived had about twice the number of elements; about 16. And in the lab, we were making chips with about 30 elements, looking at the possibility of making devices with twice that many: about 60 elements on a chip. I took a piece of semilogarithmic paper, plotted this, and starting with the flat transistor in 1959, realized that we were essentially doubling every year.
I extrapolated the observation and said that we were going to continue doubling each year and go from about 60 items at the time to 60,000 in 10 years. At the end of 10 years, if we didn’t have 10 doublings of the number of elements on a chip, we at least had nine. So one of my colleagues—I think it was Carver Mead, a professor at Cal Tech—christened it “Moore’s Law,” a name that has stuck well beyond my calculations.”
Is there a future for Moore’s Law?
Various authoritative sources, including big Intel rivals like NVIDIA, they assured years ago that this law was dead. Moore himself did not ensure its duration forever due to the limitations of the use of silicon as the main material in the semiconductor industry. It must be insisted that it is not a concrete mathematical/physical law but an observation/prediction of the capacity of the industry that from its origin had an expiration date.
While acknowledging that the law has encountered natural barriers in the form of physical limits, Intel says it is working in the future of processing with alternatives to be able to maintain it that go through improvements in the lithographic manufacturing processes, in the packaging of the chips and in their architecture. With these techniques, Intel promises to deliver 1 trillion transistors on an integrated circuit by 2030.
Intel insists on this not only for technical reasons, but also for commercial ones, with an x86 architecture with which it has dominated world computing with an iron fist in past decades, but which no longer exists today, as shown by the work of companies like Apple with other companies. architectures like ARM in their personal computers or NVIDIA solutions with accelerators in data centers.
Also, the use of silicon is reaching its limits (if it hasn’t already). The exponential increase in the reduction in size and cost of semiconductors is long over and with it Moore’s Law. And maybe not because of technology but because of costs, as very few manufacturers can afford the exorbitant cost of researching and designing new generations of chips, let alone building and maintaining the factories to produce them. Intel itself has had serious problems transitioning to 10nm process technologies.
And for the future, completely new technologies must be imposed that must come from quantum computing. Here Intel is not alone, it is not at the forefront of its development, much less it has the dominance when its microprocessors opened the modern era of the personal computer. There are still decades left for its application at the end customer level, but quantum is the next frontier And it is a technology with enormous potential that will change everything known in computing. Rest in peace pioneer Gordon E. Moore and also the famous Law of him.
