Intel Ice Lake processors tile a silicon wafer.
Stephen Shankland / CNET
Moore's Law, the observation that the number of transistors on a computer chip doubles every 24 months, has caught on as progress in miniaturization of circuits comes to a standstill. Chip giant Intel, however, has set a course to keep the idea alive and plans to pack 50 times as many transistors on processors as is possible today.
The advancement of Moore's Law, named for Intel co-founder Gordon Moore, carried chips from expensive mainframes to personal computers in the 1980s and now to smartphones, watches, cars, televisions, washing machines, and pretty much anything powered by electricity.
For more of that
Subscribe to the CNET Now newsletter to receive the most important stories of the day from our editors.
Moore's Law worked by shrinking transistors, the computing elements on a chip. Intel plans to downsize them further, but also increase density by stacking chips in multilayer packages.
"We firmly believe that there will be a lot more transistor density ahead of us," said Intel chief architect Raja Koduri in a speech on Monday at the Hot Chips conference about the latest revelations of processors. "The vision will play out over time – maybe a decade or more – but it will play out."
Koduri's optimism reflected the excitement of many other companies at Hot Chips, a technical conference where researchers detail the progress. AMD, Nvidia, Google, Microsoft, IBM and a number of startups have shown how they are developing both general-purpose chips and those for tasks such as artificial intelligence, graphics and networking.
How Intel expects chip advancement
Koduri described several steps to packing more transistors into one chip as possible, with 10nm chips like the Tiger Lake processor arriving in laptops this fall. First there will be the most traditional approach of shrinking transistors and pushing them closer together. That will triple the transistor density, Koduri predicted.
Next up are new transistor designs that continue the current transformation of transistors from flat circuit elements into 3D structures. These steps, called nanowires and stacked nanowires, should quadruple the density.
Then there are packaging innovations with chips that are stacked to form a layer of processor elements. That should quadruple the density again. The mathematics as a whole increases the density by about a factor of 50.
Years of Intel trouble
Intel's optimism contrasts with troubled times when Moore's Law keeps going.
Intel, once the undisputed leader in chip manufacturing, has struggled in recent years. The transition from a manufacturing process with transistor properties of 14 nanometers to a later 10 nm took five instead of two years. A nanometer is a billionth of a meter, and with 14 nm wide circuit elements, Intel can pack about 7,000 across the width of a human hair.
Next, Intel delayed the transition from 10 to 7 nm for six months, and Apple is dumping Intel chips from its Macs. To make customization easier, Intel introduced a more flexible design process that allows it to rely more heavily on other chipmakers like its top rival Taiwan Semiconductor Manufacturing Corp. can leave.
Moore's Law, but at what price?
TSMC, which moved to 7nm manufacturing about two years ago and made Apple's iPhone chips, declared last year, "Moore's Law is good and alive." But unlike in the past, Moore & # 39; s Law's moves are now creating new costs for companies that want to adopt the most advanced manufacturing techniques.
Intel's Tiger Lake chip will vastly improve laptop performance in 2020.
Microsoft's Xbox One in 2013, Xbox One X in 2017, and Xbox Series X this year all have chips around the same size, which in the past would have meant that the chips would have cost about the same price . Now "it's significantly more expensive for the newest one," said Microsoft chip designer Jeff Andrews.
Another challenge besides the costs is that new chips often only accelerate certain computing processes. This is useful for tasks like artificial intelligence and graphics, but it makes life harder for software programmers who have to deal with processors that work in different ways.
Intel is trying to bridge this chip gap with a new layer of software called oneAPI. It's a remarkable move: Intel is a hardware specialist, but software is an essential step in making its chips useful.
"Hardware architecture teams must increasingly consist of software experts," said Koduri.
New chip ideas
At Hot Chips, the processor manufacturers have also presented a number of innovations. To the largest:
- Intel's Tiger Lake processor uses a new incarnation of energy saving technology called DVFS, or dynamic voltage and frequency scaling. Different parts of the chip can run faster or slower for high priority tasks to save power. Intel is now juggling the priorities between its multiple processor cores, the storage system and the communication structure that connects everything.
- AMD's competing Ryzen 4000 series chips, code-named Renoir, now available in PCs, are the first eight-core chips for super-thin laptops. AMD originally planned for a six-core design, but realized that a careful design could accommodate eight for better performance in tasks like video and photo editing, architect Sonu Arora said. They use half the power for a certain level of performance like their predecessors.
- IBM's Power10 processors, which have 18 billion transistors and will be available on massive Unix servers next year, can be combined into a single powerful server with up to 240 processor cores. In addition, a "pod" of interconnected servers can share up to 2 petabytes of storage. This is useful for massive business computing challenges such as data mining and managing inventory databases.
- Startup Lightmatter presented its Mars chip to accelerate AI work such as image recognition. It connects roughly a billion traditional transistors with tens of thousands of components that use light instead of electricity to carry data and do calculations. The idea behind this photonic technology is to reduce power consumption.