Looking to the Future:
International Technology Roadmap for Semiconductors 2.0

The ten year anniversary of Conroe comes at a time when the International Technology Roadmap for Semiconductors report into the next 10-15 years of the industry has been officially launched to the public. This biennial report is compiled by a group of experts in the semiconductor industry from the US, Europe and Asia and is designed to help the industry dictate which path to focus R&D for the next 10-15 years, and runs for nearly 500 pages. While we could go into extensive detail about the contents, we plan to give a brief overview here. But for people interested in the industry, it’s a great read for sure.

The report includes deep discussions regarding test equipment, process integration, radio frequency implementations (RF), microelectromechanical systems (MEMs), photolithography, factory integration, assembly, packaging, environmental issues, improving yields, modeling/simulation and emerging materials. With a focused path to a number of technologies, the hope is that leading contenders in each part of the industry can optimize and improve efficiency in directional research and development, with the possibility of collaboration, rather than taking many different routes.

Obviously such a report is going to make successful and unsuccessful predictions, even with a group of experts, based on the introduction of moonshot style features (FinFET) or unforeseen limitations in future development. For example, here is the first roadmap published by the Semiconductor Industry Association in the first report in 1993:


Original 1993 Semiconductor Industry Association roadmap

As we can see, by 2007 it was predicted that we would be on 10nm 100nm chips with up to 20 million ‘gates’, up to 4GB of SRAM per chip and 1250mm2 of logic per die. Up to 400mm wafers were expected in this timeframe, with 200W per die and 0.002 defects per square cm (or 5.65 errors per 300mm wafer).

Compare that to 2016, where we have 16/14nm lithography nodes running 300mm wafers producing 15 billion transistors on a 610mm2 die (NVIDIA P100). Cache currently goes up to 60-65MB on the largest chips, and the power consumption of these chips (the ASIC power) is around 250W as well. So while the predictions were a slow on the lithography node, various predictions about the integration of components onto a base processor were missed (memory controllers, chipsets, other IO).

What makes the most recent report different is that it is listed as the last report planned by ITRS, to be replaced by a more generalized roadmap for devices and systems, the IRDS as the utility of semiconductors has changed over the last decade. In this last report, a number of predictions and focal points have been picked up by the media, indicating a true end to Moore’s Law and how to progress beyond merely shrinking lithography nodes beyond 7nm. Part of this comes from the changing landscape, the move to IoT and the demand for big data processing and storage, but also the decrease in the profitability/performance gain of decreasing node sizes in comparison to their cost to develop is, if believed, set to put a paradigm shift in integrated circuit development. This applies to processors, to mobile, to DRAM and other industry focal points, such as data centers and communications.

I do want to quote one part of the paper verbatim here, as it ties into the fundamental principles of the future of semiconductor engineering:

“Moore’s Law is dead, long live Moore’s Law”

The question of how long will Moore’s Law last has been posed an infinite number of times since the 80s and every 5-10 years publications claiming the end of Moore’s Law have appeared from the most unthinkable and yet “reputedly qualified” sources. Despite these alarmist publications the trend predicted by Moore’s Law has continued unabated for the past 50 years by morphing from one scaling method to another, where one method ended the next one took over. This concept has completely eluded the comprehension of casual observes that have mistakenly interpreted the end of one scaling method as the end of Moore’s Law. As stated before, bipolar transistors were replaced by PMOS that were replaced by NMOS that were also replaced by CMOS. Equivalent scaling succeeded Geometrical Scaling when this could not longer operate and now 3D Power Scaling is taking off.

By 2020-25 device features will be reduces to a few nanometers and it will become practically impossible to reduce device dimensions any further. At first sight this consideration seems to prelude to the unavoidable end of the integrated circuit era but once again the creativity of scientists and engineers has devised a method ‘To snatch victory from the jaws of defeat’.

Core: Performance vs. Today Looking To The Future: 450mm Wafers in 2021, and Down to ‘2nm’
Comments Locked

158 Comments

View All Comments

  • fanofanand - Thursday, July 28, 2016 - link

    Ian, your reviews are always too notch. You have incredible knowledge, and your understanding of both CPU and memory architecture etc is unparalleled in journalism. Ignore the trolls, this was a fantastic article.
  • extide - Wednesday, July 27, 2016 - link

    Different editors for different content. Honestly I thought this was a great piece. I think this site is not quite up to what it was back then, just go read the articles for Fermi, or when Bulldozer released and stuff, much more deep dives into the architecture. I realize that Intel and the other manufacturers may not always be willing to release much info, and they seem to release less these days but I don't know -- the site feels different.

    Honestly, and I am pretty forgiving, being as late as this site has on the recent GPU reviews is pretty inexcusable. Although, that is obviously nothing to do with Ian, I always like Ian's articles.
  • Ian Cutress - Wednesday, July 27, 2016 - link

    Thanks! :)
  • fanofanand - Thursday, July 28, 2016 - link

    The Pascal review was pretty damn deep, not sure how much farther you expect them to dive. That said, it was very, very late.
  • Michael Bay - Thursday, July 28, 2016 - link

    ADHD millennial detected.
  • Notmyusualid - Thursday, July 28, 2016 - link

    Hey Rain Cloud!

    I enjoyed it, as did many others here - try reading the friendly discussion!
  • tipoo - Wednesday, July 27, 2016 - link

    I like articles like these. Sometimes certain processors stick around as the baseline in my brain even after a decade (holy hell!). Core 2 Duo is always a reference point for me, so is a 3GHz P4.
  • rocky12345 - Wednesday, July 27, 2016 - link

    Yea I still have a Q6600 Core2Quad running strong in the front rooom OC to 3700Mhz been running like that since day 1.
  • wumpus - Thursday, August 4, 2016 - link

    Wow. I've assumed that they would at least burn out so that they would need replacement (like my old super celeries). I'm sure you can measure a speed increase between a modern i5 and yours, but it would be hard to notice it.
  • chekk - Wednesday, July 27, 2016 - link

    Nice, thanks Ian. Interesting to look back and then ahead.
    I still use my E6400 in a media playback machine using the first "good" integrated graphics, the NVidia 9300. Since it runs at stock frequency @ 1V (VID spec is 1.325), it's pretty efficient too.

Log in

Don't have an account? Sign up now