REBOOTING MODRE'S LAW
Is the end nigh for Moore’s Law? You could certainly make that case. After all, it’s true that the computer chip industry has had a tricky few years of late, keeping up with the pace of progress that Moore’s Law implies.
Lest you have forgotten, Moore’s Law is simply the observation that transistor densities in integrated circuits tend to double every two years. Put another way, Moore’s Law says computer chips either double in complexity and thus performance, or halve in cost — or some mix of the two, depending on what you’re after — every couple of years.
It dates back to the 1960s or 1970s, depending on the part of the technological folklore you want to focus on. But the point is that it undeniably proved prescient for several decades. More recently, it’s come into doubt. Whether Moore’s Law has really hit a wall will only become clear several years from now. What we can say is that big players like Intel and TSMC — the latter being the foundry responsible for making most AMD and Nvidia GPUs in recent memory — have missed several self-imposed targets when it comes to new production nodes and shrinking their transistors.
The situation has become acute enough for Intel to abandon its so-called TickTock cadence of a new CPU architecture and production process in alternating years, as dictated by Moore’s Law. Instead, Intel has switched to a three-year cadence of process, architecture and optimisation.
Meanwhile, it’s also undeniable that CPU operating frequencies have hit something of a wall around roughly 4GHz. In the good old days when MHz ruled, CPUs would benefit from both process shrinks and increased frequencies. Not any longer.
It’s just possible, however, that 3D XPoint’s chalcogenide underpinnings could reboot Moore’s Law, at least for a time. The structure is completely different, so the fact that transistor gate sizes are running up against the laws of physics suddenly becomes irrelevant. 3D XPoint’s chalcogenide devices don’t use transistors. Likewise, chalcogenide devices have at least the theoretical potential to hit insanely high frequencies, well beyond those seen in commercial silicon chips. Whether any of this becomes a practical reality, only time will tell. But it’s fun to know that it’s possible we’re on the cusp of a new wave of computing.