facebook rss twitter

Moore's Law is not dead insists Intel's Jim Keller

by Mark Tyson on 1 July 2019, 14:31

Tags: Intel (NASDAQ:INTC)

Quick Link: HEXUS.net/qaebe6

Add to My Vault: x

In the world of tech, Moore's Law pops up in discussions about the industry, progress, and what the future holds rather frequently. In brief, the law, which 'celebrated' its 50th anniversary in April 2015, is a prediction that transistors will decrease in cost at an exponential rate and increase in performance. In recent years, however, more and more tech commentators have asserted that Moore's Law is dead and they seek to amend or redefine technological progress.

With the above in mind, and the well known stalling by Intel which seems stuck on variations of the 14nm process, it was interesting to see Intel's Jim Keller was at the Silicon 100 Summit yesterday, delivering a keynote speech. The keynote was provocatively dubbed 'Moore’s Law isn't dead and if you think that you’re stupid'.

During his presentation Keller argued that Moore's Law is relentless, and presented slides showing plans for a path to 50x gate density. Of course the argument over whether Moore's Law is dead depends on quite a lot of things, especially your interpretation of the law to begin with. Many distil the law to a doubling of transistors every 18 months due to process improvements…

In his main slide, 'Path to 50x gate density' you can see Keller highlights various new technologies coming together to result in that goal, each contributing in its own way. However, special attention is placed on die-on-wafer stacking here.

Unfortunately Keller's presentation and slides aren't shared anywhere other than Twitter that I can find. The Silicon 100 Summit is described by Raja Koduri as "basically our silicon design and transistor engineers nerding out with SF locals". That must be the reason for the lack of press and official blog coverage, and the peculiar cocktails on offer.

Transistor anniversary

I don't know if the Silicon 100 Summit is timed to coincide with the anniversary of the announcement of the first transistor - but it was indeed on the same day this year. Bell Labs announced the first working transistor on 30th June 1948. However, the first working transistor actually came to life on 16th December 1947, and it was demonstrated working in an amplifier on Christmas Eve that same year.

HEXUS Forums :: 11 Comments

Login with Forum Account

Don't have an account? Register today!
“Moores law isn't dead, Intel just bonked it over the head for the past 5 years”

But isn't claiming it isn't dead by building (stacking) vertically with layering basically cheating Moores law?
“However, special attention is placed on die-on-water stacking here.”

I think ‘die-on-wafer’ has been autocorrected based on the slide photo…
Die-on-water sounds so good though!
Sounds like the “orange” drink I had in a pub the other day …. and may taste like it too. It certainly bore no more than a passing resemblance to anything with actual fruit.
“, is a prediction that transistors will decrease in cost at an exponential rate and increase in performance”
No Moore's law doesn't say that but:
“Moore's law is the observation that the number of transistors in a dense integrated circuit doubles about every two years.”

Nothing to do with cost.
Do you really write for a tech site?