Semiconductor Industry To Officially Abandon Moore’s Law

[Megalith]

Heat death and other fundamental limits have made Moore’s law impossible to sustain.

A rule of thumb that has come to dominate computing, Moore's law states that the number of transistors on a microprocessor chip will double every two years or so — which has generally meant that the chip's performance will, too. The exponential improvement that the law describes transformed the first crude home computers of the 1970s into the sophisticated machines of the 1980s and 1990s, and from there gave rise to high-speed Internet, smartphones and the wired-up cars, refrigerators and thermostats that are becoming prevalent today.

 

[westrock2000]

Repealed

 

[Ducman69]

I call bullshit. I think they have witnessed a decrease in demand for additional performance, a market push for smaller and smaller devices (thinner phones being more important than faster ones), while simultaneously being met with a collapsing competitor that would leave Intel at least as a global monopoly which would force world governments to break the company apart.

This is them simply saying, screw it, desktops and gaming laptops are fast enough, and nobody cares about fast phones... heck, they love the iPhones despite the crappy outdated hardware that is inside them, so lets change the focus to lower power and lower heat.

Sucks for people like myself though, that would be happy with a much larger physical CPU, and would readily mount a 520mm radiator to the top of my case.

 

[jimmyb]

I call bullshit.

As someone who works in semiconductor engineering I can confidently say that your comment is mostly wrong. Physics is the primary reason that Moore's law is failing - just as the Nature article states.

 

[Hornet]

I call bullshit. I think they have witnessed a decrease in demand for additional performance, a market push for smaller and smaller devices (thinner phones being more important than faster ones), while simultaneously being met with a collapsing competitor that would leave Intel at least as a global monopoly which would force world governments to break the company apart.

This is them simply saying, screw it, desktops and gaming laptops are fast enough, and nobody cares about fast phones... heck, they love the iPhones despite the crappy outdated hardware that is inside them, so lets change the focus to lower power and lower heat.

Sucks for people like myself though, that would be happy with a much larger physical CPU, and would readily mount a 520mm radiator to the top of my case.

Moore's law has more to do with transistor density, rather than the increase in CPU size to accommodate more transistors. It has been held true by the rapid advancement in shrinking the size of transistors (hence increasing density).

This of course is an observation that cannot be forever true, as we cannot shrink the size of atoms, we can only use fewer atoms but as dictated by the law of physics, there's a limit to how small you can go to have a functional transistor.

That's one limitation to Moore's Law. Another limitation that will come sooner, is the difficulty in manufacturing these chips with smaller transistor. You need increasingly expensive and complex manufacturing tech to fabricate them and the industry can no longer advance at the same rate as it did 10 years ago. Since Moore's Law observation is with respect to time, this will obviously mean the observation will no longer hold true.

 

[Ehren8879]

A refocus on more efficient electronics and coding is welcome. Then, when the next breakthrough in transistor medium is developed be can benefit from those principals.

It's like tick-tock reimagined.

 

[pxc]

Firstly, thank you for linking to a good information source. That doesn't happen too often in news here.

But by the early 2020s, says Paolo Gargini, chair of the road-mapping organization, “even with super-aggressive efforts, we'll get to the 2–3-nanometre limit, where features are just 10 atoms across. Is that a device at all?” Probably not — if only because at that scale, electron behaviour will be governed by quantum uncertainties that will make transistors hopelessly unreliable. And despite vigorous research efforts, there is no obvious successor to today's silicon technology.

The end is in sight for FETs, but that's still several years away. You can't really call a 60+ year streak for almost consistent scaling bad...

 

[-Dragon-]

You can't really call a 60+ year streak for almost consistent exponential scaling bad...

Yeah it's hard to argue with 9 orders of magnitude improvement.

 

[USMCGrunt]

A refocus on more efficient electronics and coding is welcome. Then, when the next breakthrough in transistor medium is developed be can benefit from those principals.

It's like tick-tock reimagined.

This...we have so much untapped potential in our current hardware because of lazy or inefficient coding. I'm really not too concerned though with advances in quantum computing, optical circuits, and carbon nanotech, the next medium or "generation" of computers is only a decade or two away.

 

[butterfliesrpretty]

Who the hell is the "worldwide semiconductor industry". I really doubt we'll be still using silicon 10 years from now anyway, so many promising technologies supposedly around the corner.

 

[westrock2000]

Who the hell is the "worldwide semiconductor industry". I really doubt we'll be still using silicon 10 years from now anyway, so many promising technologies supposedly around the corner.

It's still coming down to physical size being the limiting factor. As we move down in size, electrons can "leak" across barriers because the barriers just aren't thick enough to contain the electrons. It doesn't matter WHAT the structure is made out of, once you get down to a couple atoms thick, you just run into problems. Leakage is the big problem.

 

[westrock2000]

Who the hell is the "worldwide semiconductor industry". I really doubt we'll be still using silicon 10 years from now anyway, so many promising technologies supposedly around the corner.

Also don't forget, Silicon is not the first thing we made semconductors out of. It is the best AND most practical element for our needs. People have been looking for something better, but it's hard to do better.

Kind of like convincing people that gasoline is bad. It's been very hard to present an alternative that's better in all ways.

 

[TechLarry]

Moores law was dead 5 years ago. Might as well make it official.

I'm sure it will be replaced with "Intels Law". "They get faster when WE say they get faster"

 

[pxc]

Who the hell is the "worldwide semiconductor industry".

Well, when a mommy semico and a daddy semico want to make a processed wafer, they rely on various tool manufacturers to equip their fabs with lithographic tools, mask making tools, etching tools and several other tools.

For decades there has been an alignment on Moore's Law progress and delivery of appropriate tools. What is about to happen is those tool makers will release revised roadmaps that no longer fall on the Moore's Law curve due to physical limitations.

 

[westrock2000]

Well, when a mommy semico and a daddy semico want to make a processed wafer, they rely on various tool manufacturers to equip their fabs with lithographic tools, mask making tools, etching tools and several other tools.

There is a variant to this that been popular lately.

Wait for a smaller semiconductor company to finally go bankrupt and then buy all their assets for a song. And bitch and moan at the vendors for wanting to charge for service and repairs. It's a very profitable model, and the only requirement is that at least 2 companies exist.

 

[Jagger100]

Lack of competition probably has moore to do with it.

 

[Gatto]

Wow i didnt know this ! Look everyone has problem this year !