A new breakthrough in photonics gives a massive boost to Moore's Law and could increase processing speed by twenty-fold while reducing power consumption.
Moore's Law was one of the most important phenomena of the last 50 years; it made the IT revolution possible; without it, we would still live in an analogue age, but it is dead, or so they say. But wait, is that new life? Has Moore's Law merely suffered a near-death experience? Photonics could change everything with a breakthrough. But consider this opportunity in the context of convergence with other technologies such as 5G and graphene. That is why people who look at the slowing pace of computing advances as evidence of terminal decline are wrong and why we are, in fact, on the cusp of a remarkable, exciting but scary revolution.
To remind you, Moore's Law is maned after Gordon Moore, co-founder of Intel, who in 1965 observed that the number of transistors on an integrated circuit was doubling every two years. And from that, popular mythology twisted and tweaked, and Moore's became known as computers doubling in speed every 18 months. But there are these things called atoms and electrons, and in a world of atoms, there is a limit to how small you can get, and electric signals have limitations. And so we hit a limit — instead, we saw the age of specialist chipsets, with the likes of Nvidia becoming the most valuable chip company in the world.
But as Ray Kurzweil says, we have been here before. We tend to see paradigm shifts, such as the shift from vacuum tubes that once underpinned computer technology to silicon-based transistors. With each shift, it all changes. Silicon-based integrated circuits might be a tad passe, but new paradigms are afoot.
Photonics is a new paradigm of extraordinary potential — the technology uses light (photons) rather than electronics (electrons) to transmit data, and it has several major advantages. For one thing, it is faster as its speed is limited solely by the speed of light. For another thing, it is less energy-intensive as circuits don't suffer from the same heating problem as silicon integrated circuits; thirdly, whilst phonetics can mean more transistors per unit of space, it can also mean larger computers without significant loss of efficiency. Science fiction was envisioned as super-computers the size of a mountain, and early computers were indeed the size of s small building. But consider a computer or a neural network the size of s mountain with just as many transistors per unit of space as a current state-of-the-art computer, transmitting data at the speed of light. Is your mind-boggling yet? If not, check yourself into the mind-boggling clinic fast; your boggling genes might not be working properly.
The photonics opportunity
The most well-known player in the photonics field is Lightmatter, a company which says, "We've created a photonic processor and interconnect that are faster, more efficient, and cooler than anything else on earth (or anything ever experienced before) to power the next giant leaps in human progress."
But now, a company called CogniFiber says its newest photonics system "far surpasses NVIDIA and Lightmatter's leading servers."
It explains: "soon to be launched pure-photonic system is expected to reach a staggering speed of 100 million tasks per second, far surpassing NVIDIA's DGX A100's estimated 5 million tasks per second or Lightmatter's Envise Server of 24 million tasks per second (est.*)."
Dr Eyal Cohen, Co-founder & CEO of CogniFiber, said: "Our initial findings are outstanding, displaying the capabilities to support the mega data centres of tomorrow," said "Operating at room temperature without dissipating significant heat to its surroundings means data centres can offer customers greater uptime reliability with lower cost."
Techopian makes no comment on this specific technology and how it compares with lightmeter.
But this does illustrate an important point. And consider the implications of more powerful server technology in the age of 5G. And in time, of course, it will mean more powerful personal devices.
The AI revolution is here
Jan 25, 2023
The impossible conclusion about technology becoming less disruptive and why it is so dangerous
Jan 20, 2023
Tech bubble! Are you kidding?
Jan 06, 2023