Chip Letter Links No. 16 : Gordon Moore's Notebooks, Nvidia on Acquired, Assembly Language, Jeff Hinton, RISC-V on Jupyter
Great links, images and reading for 16 April 2023
Hi everyone and thanks for subscribing. This is one of a regular series of posts with links, images and articles of interest, inspired by Adam Tooze’s excellent Chartbook.
Each edition starts with a beautiful die image. This week it’s an Intel Core 2 Duo courtesy of Martijn Boer.
The Chip Letter is 100% reader supported, so if you’d like to support us and get access to all of our content, then please consider taking out a paid subscription. Thanks to everyone who has taken out a paid subscription already!
Gordon Moore’s Notebooks
In our post on the Intel iAPX432 we looked at Gordon Moore as a risk taker in his role as CEO of Intel. Today we focus on Moore’s work in his role as a research scientist during his years at Fairchild. ‘Patent notebooks’ were used by researchers to record their work in Fairchild’s laboratories. Moore’s notebook’s from this period are held in the archive of the Computer History Museum (CHM). From the museum’s website:
In 2012 Texas Instruments’ donated five of Gordon Moore’s Fairchild patent notebooks to the Computer History Museum collection. Spanning the period 1957 through 1968, they are hand-numbered 1 through 6.
The notebooks provide a fascinating look into Moore’s work at Fairchild, during the period when he was Director of Research and Development.
The first notebook opens with a memo to Moore, dated 2 August 1962, on the development of parts for the Minuteman 2 Intercontinental Ballistic Missile (ICBM), reflecting Fairchild’s focus on military and defense procurement.
The material in the notebooks ranges over materials science, component design, electronic circuits, circuit timing, logic design, project planning, manufacturing, testing and commercial negotiations.
Here is one example of a sketch for a ‘Pentode’ replacement.
Full pdfs of the five notebooks can be found here.
There is a key gap, though, for the period running up to the publication of his famous 1965 paper published in Electronics magazine, setting out what would become known as Moore’s Law. From the CHM website again:
Number 5, that would have covered the years from 1963 to 1965 when he was gathering data for the Electronics article, is missing. We would appreciate any information as to its current location.
Nonetheless, we can get a sense of Fairchild’s focus on reducing the size of components on integrated circuits from this memo from 1962, which talks about ‘Maintaining our leadership in the photolithographic field’ and which sets out work to reduce resolution sizes (0.3 mil to 0.2 mil to 0.1 mil - where a mil is a thousandth of an inch).
This was an era when Fairchild and its competitors had to do everything themselves, including creating their own photolithography equipment.
The notebooks show just how busy developing this technology Moore was over this period. Gordon Moore wasn't just a ‘visionary’ who observed the increases in component density over time and extrapolated them to derive ‘Moore’s Law’. He was a leader in making them happen.
Assembly Language Poll
Our recent post ‘The Unnecessary Obscurity of Assembly Language’ had a poll on whether readers had used 8-bit or 64-bit assembly languages. The results are in and 8-bit ‘won’ by a significant margin!
Thanks to everyone who voted!
I think this confirms that assembly language was much more popular (or perhaps, more accurately, it was more necessary) for 8-bit machines when compared to today’s leading 64-bit architectures.
It’s interesting too that at least 78% of readers who voted have used an assembly language of some type.
There are some great comments on the post too, so thanks to everyone who commented, including those who disagreed with the post’s stance!
Nvidia on the Acquired Podcast
I was thinking of writing about Nvidia’s early years, only to discover that Ben Gilbert and David Rosenthal at the ‘Acquired’ Podcast had have made an episode covering this period already. It’s available on YouTube or on Spotify and Apple Podcasts.
It’s a really entertaining and informative listen (or read). Some gems:
Nvidia CEO Jensen Huang started his career at now arch-rival AMD;
They popularised the term ‘Graphics Processing Unit’ or ‘GPU’ with the GeForce 256;
The name Nvidia comes from the latin ‘invidia’ meaning envy.
Most significant of all though is the link to Moore’s Law yet again. Nvidia only survived because they understood that they needed to align their business with Moore’s Law.
Ben: Yeah. And the magic was that Jensen really figured it out early that they were in a business that was totally at the mercy of Moore's Law. In having that initial realization as early as they did with the proliferation of competitors, and everyone doing the triangles, DirectX and what, that taught them the lesson early enough that, oh, we are in a business where we must be reinventing. There is no way to stay ahead other than ruthless self-examination and completely ending and rebounding the business.
David: Yup, ship faster and reinvent.
Ben: Yeah. That, to me, is why they survived.
Geoff Hinton
An interesting and wide ranging interview with Geoff Hinton, the ‘Godfather of Artificial Intelligence’, made by CBS and now on YouTube.
It includes an interesting comment on how he came to be interested in AI:
I was interested in psychology originally then I decided we were never going to understand how people work without understanding the brain. The idea that you could do it without worrying about the brain that was a sort of fashionable idea back in the 70s but I decided that wasn't on. You had to understand how the brain worked …
Look out for a future post exploring the links between the history of AI and research into how the brain works.
Substack Notes
I’m planning to post regularly on Substack Notes, which, at the moment, looks like a promising place for short form posts and already seems quite active.
Please head over to the Notes tab on the Substack App for updates, some previews of upcoming posts and more topical commentary. Look forward to seeing you there!
After the paywall is a fun way of exploring RISC-V assembly language in the browser without installing anything on your machine.