If Gordon Moore were only known as the co-founder of Intel, he would still be considered one of the biggest forces in the PC industry. However, Moore, who passed away on Friday at the age of 94, was far more well-known for his prediction that later came to be called "Moore's Law". It was a prediction that shaped the entire PC industry in general, and Intel in particular.
Moore came up with his initial prediction before he co-founded Intel with Robert Noyce in 1968. While he was working at Fairchild Semiconductor as its head of R&D, Moore wrote an article in Electronics magazine in 1965. The article gave his thoughts on the semiconductor industry's future for the next 10 years. In the article, he wrote:
The complexity for minimum component costs has increased at a rate of roughly a factor of two per year ... Certainly over the short term this rate can be expected to continue, if not to increase. Over the longer term, the rate of increase is a bit more uncertain, although there is no reason to believe it will not remain nearly constant for at least 10 years. That means by 1975, the number of components per integrated circuit for minimum cost will be 65,000. I believe that such a large circuit can be built on a single wafer.
In 1975, Moore revised his forecast somewhat in a speech he made during the annual IEEE International Electron Devices Meeting. He stated that until 1980, transistor amounts on chips would double every year, and after 1980 they would double every two years. It's that last prediction that finally became popularized as "Moore's Law."
That push to add more and more transistors inside chips became Intel's mission for launching new computing chips for the next several decades. Indeed, Moore saw that the personal computer in the early 1980s, and in particular the ones that were based on the IBM design, would serve to help Intel innovate in the CPU space. That led to Intel's chips being installed on nearly all PCs as the rise of that industry happened in the next two decades.
Intel's chips, as Moore predicted, crammed more transistors into single chips and became more powerful as a result, and cheaper as well, from the 80286 processor to the 386 and 486 chips. In 1993, the Pentium was released, with 3.1 million transistors, a massive leap forward compared to Moore's prediction of 65,000 transistors for a chip in 1975. Today, Intel has its 13th Gen Core chips, which, although the company has not revealed its official numbers, is believed to have as many as 25.9 billion transistors inside.
Moore's Law allowed Intel to become the massive company it is today, and even now it still is the number one maker of PC GPUs, although AMD has been a highly skilled competitor for some time now.
For the past couple of decades, many researchers have stated that we are reaching the limit of how Moore's Law would work. As early as 2003, Intel researchers predicted that CPUs would stop doubling transistors by 2018, or maybe a bit further. The famous theoretical physicist Michio Kaku predicted in 2012 that the laws of physics would sooner or later prevent Moore's Law from allowing smaller transistors on CPUs.
However, Intel might find ways to get around some of those limitations so that Moore's Law could still work. In a 2020 article for the MIT Technology Review, it mentions Jim Keller, who at the time headed up Intel's silicon engineering team. The article stated:
He points out that there are probably more than a hundred variables involved in keeping Moore’s Law going, each of which provides different benefits and faces its own limits. It means there are many ways to keep doubling the number of devices on a chip—innovations such as 3D architectures and new transistor designs.
More recently, in a 2022 editorial, Dr. Ann B. Kelleher, Intel's Executive Vice President and General Manager of Technology Development, stated:
Intel’s next great architectural innovation is RibbonFET, our implementation of the gate-all-around (GAA) transistor, arriving with Intel 20A. RibbonFET represents our first new transistor architecture since FinFET. RibbonFET delivers faster transistor switching speeds with the same drive current in a smaller footprint.
Dr. Kelleher says Moore's Law will continue to be the baseline for making faster and more powerful CPUs for a long time to come:
Intel’s Components Research is focused on three key research areas to deliver the fundamental building blocks for more powerful computing well into the future. We have a full pipeline of research underway that gives us the confidence we will maintain Moore’s Law for the next decade or longer. Future innovations fueling Moore’s Law are limited only by our imagination.
It sounds like Intel hasn't given up on pushing CPU performance, and as a result, Moore's Law is still in effect, but perhaps with new engineering that Moore could not have foreseen when he first came up with his prediction nearly 60 years ago.
1 Comment - Add comment