Keeping up with Moore's Law

December 6, 2018 by David Bradley, Inderscience
Plot of CPU transistor counts against dates of introduction; note the logarithmic vertical scale; the line corresponds to exponential growth with transistor count doubling every two years. Credit: Wikipedia

These days, Moore's Law is not so much a scientific law as an aspiration. The notion that there is a doubling every year of the number of components that can be squeezed on to the same area of integrated circuitry was first observed in the mid-1960s by Gordon Moore, the co-founder of Fairchild Semiconductor and Intel. Ever since the microelectronics industry has strived to Moore's Law although in some periods that annual doubling seems to occur over a period of 18 months if not longer.

Nevertheless, it still offers a rule-of-thumb for how rapidly advances and posits a guideline as to what technology industries might aim for. Now, a paper in the International Journal of Technology Management asks whether technology improvement rates in knowledge industries, microelectronics, , and genome-sequencing technologies might follow this law.

Yu Sang Chang of Gachon University, in Seongnam, Jinsoo Lee of the KDI School of Public Policy and Management, in Sejong, and Yun Seok Jung of the Institute for Information and Communications Technology Promotion, in Daejeon, Korea, have tracked technology developments to see whether Moore's Law held over the period 1971 to 2010. Their study shows that indeed it did, moreover they suggest that an analogous exponential law also applies to mobile cellular and genome-sequencing technologies.

While there has been no in transistor density, the team has found that the improvement rate in microprocessor clock speed has not been sustained. That said, for genome sequencing technology which is essentially still in the early stages of , developments continue apace.

The team points out that the 5-nanometre limit on the quantum tunnelling effect will represent a barrier to the further shrinking of transistors and that we are fast approaching that limit. However, developments in nanotechnology might still allow the industry to sustain Moore's Law in microelectronics even into its centenary year.

Explore further: Processing power beyond Moore's Law

More information: Yu Sang Chang et al. Are technology improvement rates of knowledge industries following Moore's law? An empirical study of microprocessor, mobile cellular, and genome sequencing technologies, International Journal of Technology Management (2018). DOI: 10.1504/IJTM.2018.095629

Related Stories

Processing power beyond Moore's Law

April 20, 2018

In 1965, businessman and computer scientist Gordon Moore observed that the number of transistors in a dense integrated circuit doubles approximately every two years, which means a doubling of computer processing power. The ...

Moore's Law is 50 years old but will it continue?

July 20, 2015

It's been 50 years since Gordon Moore, one of the founders of the microprocessor company Intel, gave us Moore's Law. This says that the complexity of computer chips ought to double roughly every two years.

Silicon Valley marks 50 years of Moore's Law

April 24, 2015

Computers were the size of refrigerators when an engineer named Gordon Moore laid the foundations of Silicon Valley with a vision that became known as "Moore's Law."

A $1000 genome could be reached by 2013

July 21, 2011

(PhysOrg.com) -- A new report published in the journal Nature describes the new machine created by Jonathan Rothberg of Ion Torrent Systems which uses semiconductors to decode DNA and takes them one step closer to being able ...

Recommended for you

Uber filed paperwork for IPO: report

December 8, 2018

Ride-share company Uber quietly filed paperwork this week for its initial public offering, the Wall Street Journal reported late Friday.

9 comments

Adjust slider to filter visible comments by rank

Display comments: newest first

V4Vendicar
1 / 5 (4) Dec 06, 2018
The Graph shown has no horizontal axis, and hence has no meaning.

This is typical of the kind of thing seen coming from scientifically illiterate, technically incompetent, Americans these days.
691Boat
5 / 5 (4) Dec 06, 2018
The Graph shown has no horizontal axis, and hence has no meaning.

This is typical of the kind of thing seen coming from scientifically illiterate, technically incompetent, Americans these days.

One can infer from the title of the plot what the X-axis is. Also, anybody familiar with Moore's Law should have no problem figuring out what is on the X-axis. If you can't figure out what would be on the X-axis of a Moore's Law plot, you may be the technically incompetent one.
Eikka
2 / 5 (1) Dec 06, 2018
The notion that there is a doubling every year of the number of components that can be squeezed on to the same area of integrated circuitry was first observed in the mid-1960s by Gordon Moore


Not Moore's Law.

The Graph shown has no horizontal axis, and hence has no meaning.


The typical depiction of Moore's law by taking a bunch of different CPUs from different years, and observing that the number of transistors in them fit roughly on a line in a lin-log plot, is also cherrypicking and misunderstanding what the Moore's law is about.

That's because it's always possible to add more transistors to a chip - it just costs more. How many there are in each CPU depends on which market segment and purpose they're aiming for, and the choice to include them in the graph is arbitrary.
Da Schneib
5 / 5 (1) Dec 06, 2018
The original Moore's Law was a doubling of transistors on a single chip every year; it's been changed a few times since, and misquoted once by an Intel exec who should have known better.

As for @Eikka's accusation of cherry-picking, anyone who's been working in electronics for the past several decades will recognize just about every one of the chips listed here as "best-of-breed" for its time.

And in fact, anyone who knows what Moore's law is knows what's on the horizontal axis.

Finally, as to the article, I wonder what metrics might apply to some of these industries.
V4Vendicar
1 / 5 (1) Dec 06, 2018
"And in fact, anyone who knows what Moore's law is knows what's on the horizontal axis. "

Yes. It is Non-Existent to hide the fact that the plot is fake.
Nik_2213
4 / 5 (1) Dec 06, 2018
This graph and its ilk always make me shiver. My home computers have progressed from a treasured 1 MHz 6502 Apple ][+ 48 KB, to current AMD 8-core, 4GHz, 32 GB self-build.
My next computer will probably be a minimal ChromeBook...
;-)
axemaster
4 / 5 (1) Dec 06, 2018
If only the software and operating systems didn't have so much bloat, maybe we could actually enjoy having fast computers...

But on another note, Moore's Law is far from over. The real barrier right now is because of heat - that's the biggest obstacle preventing us from making highly 3D processors. But work continues apace, particularly in materials science research (which I have contributed to). Magnetic/optical computers are in our future, and they will be truly amazing.
Eikka
2 / 5 (1) Dec 07, 2018
As for @Eikka's accusation of cherry-picking, anyone who's been working in electronics for the past several decades will recognize just about every one of the chips listed here as "best-of-breed" for its time.


That's an arbitrary distinction. Moore's law isn't about being "the best" - whatever that means for anyone.

Is it the most popular?
The most power-efficient?
The most cost-efficient?
The fastest?

None of that has anything to do with Moore's law, which is concerned about the minimum price per transistor per area. A particular CPU doesn't have to use the optimum chip size, and most likely won't because of other considerations.

The original Moore's law stopped working in the 70's, and the re-defined version stopped in the 90's, and now whatever is taken as "Moore's" law is simply a matter of cherry-picking data points to fit on a logarithmic curve and/or making up new definitions ad-hoc to fit the extrapolations made by hypesters.
Eikka
2 / 5 (1) Dec 07, 2018
The real Moore's law is about the multi-variable optimization problem of semiconductor chip manufacturing.

There's the chip size. There's an optimum size for the silicon wafer where the least waste is produced for the largest number of chips.

There's the feature size. There's an optimum feature size where the defect rate is balanced with the number of transistors you can lay down.

Moore's law is about finding the optimum size for a chip, stuffing the optimal number of transistors in it, and seeing how many you got. This is a theoretical number: there's no use in stuffing the chip just full of transistors - you need wires, capacitors, diodes... etc. as well so any real chip would have less, the number depending on the application. You might also need bigger or smaller chips.

This is why the number of transistors in particular CPUs has nothing to do with Moore's law. It's just about the industry choosing to double up the number, and the reporters ignoring where they didn't.

Please sign in to add a comment. Registration is free, and takes less than a minute. Read more

Click here to reset your password.
Sign in to get notified via email when new comments are made.