The exploits in the field of microprocessors continue, despite the fact that Moore’s famous laws – which call for a doubling of the number of transistors every 18 months – are reaching their limits. Recently, IBM introduced a new functional processor with 50 billion transistors on a surface the size of a fingernail, making it the smallest and densest computer architecture chip ever produced. If adopted for data centers and consumer electronics, it could reduce overall power consumption.
Note that the computer chip manufacturing industry names the generations of chips by size class, although this does not refer to the actual size of the chip. Newer commercial devices tend to use either 7-nanometer or 5-nanometer chips. For example, Apple’s M1 chip, which powers its current line of computers, and the A14 chip, built into current-generation iPhones, both use 5-nanometer chips.
IBM now claims to have produced the very first functional 2 nanometer (nm) chip. The transistors of this prototype are 12 nanometers wide, or only 24 atoms of silicon. By decreasing the size of transistors, chips become smaller, faster, and more efficient. For mobile devices, this can translate into longer battery life, as necessarily less power is consumed.
Global use on all electronic devices
According to IBM, these new chips can be used in all devices, from phones and tablets to very high performance server chips, to supercomputers. They could achieve 45% better performance and provide 75% lower power consumption compared to the 7 nanometer chips currently in production.
According to Steven Freear of the University of Leeds in the UK, denser chips would bring a whole host of benefits beyond energy efficiency, including improved reliability through the possibility of including more security. and error checking logic.
IBM’s Mukesh Khare explains that the prototype chip was manufactured at the company’s headquarters in Albany, New York, and that the process involves “hundreds and thousands of steps” in a room. white of 9300 square meters, active 24 hours a day, every day of the year. IBM says the chips will go into production in late 2024, but will outsource the task to factories elsewhere.
Moore’s Law and the Physics Challenge
Chips as we know them today date back to 1959, at Bell Labs in the United States, where transistors were made by carefully oxidizing a piece of silicon. These tiny devices are the building blocks of modern electronics, and all progress since then has essentially been about miniaturizing similar technologies. Chip miniaturization is governed by Moore’s famous laws, according to which the number of transistors in a processor doubles every 18 months or so. This law has been proven correct for decades, but it has slowed down in recent years.
According to experts, the improved chip design focused on the transistor is quickly approaching fundamental physical limits. ” It is common knowledge that Moore’s Law is collapsing. Basically, you can’t keep going down smaller and smaller scales. You can’t change physics Says Kieran Brophy of Imperial College London.
The global chip industry is releasing a timeline of expected technology leaps in miniaturization so that companies can collaborate to develop the necessary manufacturing and testing processes. Currently, this roadmap indicates that the 2-nanometer chips will be launched in 2023, with companies like Taiwan Semiconductor Manufacturing Company expected to start manufacturing commercial products using them in 2024.
The European Union is also pushing for 2 nanometer technology and investing 145 billion euros in research and development, as well as in building production capacities.
The adoption of this technology could have a dramatic impact on climate change by reducing the energy demand of the vast data centers that power email, social media, banking and video streaming services. Data centers of this type currently account for over 1% of global electricity consumption.