Chip Letter Links No. 13: Andy Grove and the 80386
Great links, reading and images for 29 January 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 we have an Intel 80386 (photo by Pdesousa359, CC BY-SA 3.0, via Wikipedia).
The Chip Letter is 100% reader supported, so if you’d like to support us then please consider taking out a paid subscription. Full year subscriptions can get a discount of over 40% - but this is a limited time offer! Thanks to everyone who has taken out a paid subscription already!
Andy Grove
Today’s edition focuses on Andy Grove and the Intel 80386 microprocessor.
Andy Grove (born András István Gróf in Budapest in 1936) was one of the most important figures in the semiconductor industry in the 20th century. He achieved this position whilst coming from a background that was incredibly challenging, to say the least:
By the time I was twenty, I had lived through a Hungarian Fascist dictatorship, German military occupation, the Nazis' "Final Solution," the siege of Budapest by the Soviet Red Army, a period of chaotic democracy in the years immediately after the war, a variety of repressive Communist regimes, and a popular uprising that was put down at gunpoint...
Grove escaped from Hungary in 1956 and shortly afterwards made his way to the United States. Unable to speak English when he left Hungary, by 1960 he had gained a Batchelor’s degree in Chemical Engineering from City College of New York and by 1963 a Ph.D. in Chemical Engineering from University of California, Berkeley. He would write about his experiences in his memoir Swimming Across.
Soon after Grove was working with Robert Noyce and Gordon Moore at Fairchild Semiconductor, and followed them to Intel in 1969. He would become Intel’s president in 1987 and then CEO in 1987, remaining chairman until 2004. Grove passed away in March 2016.
Although Grove was an expert in his technical field (whilst at Fairchild he wrote the textbook ‘Physics and Technology of Semiconductor Devices’) it’s really for his approach to management, developed during his time at Intel, that he has become famous.
CPU Of The Week: Intel 80386
There is only one real option in a post dedicated to Andy Grove, and that’s the Intel 80386, which was introduced in 1985. Technically the was a major step forward when compared to its predecessor, the Intel 80286, adding true memory protection, 32 bit arithmetic and a much larger memory addressing space (up to a maximum of 4 Gigabytes, which would prove sufficient until the mid 2000s).
The instruction set, of course, forms the base for all later Intel / AMD x86 instruction sets, and 386 code can execute on modern Intel / AMD x86 processors.
During 1985, at around the time that the 80386 was introduced, Grove and Gordon Moore, then Intel’s CEO, took the decision to exit the memory business that had earlier been the cornerstone of Intel’s business.
They also took the decision not to ‘second source’ the 80386, or in other words to have Intel as the sole manufacturer, and not to cross license the technology to another firm. (AMD effectively became a second source for the 80386 and later Intel CPUs because of their own efforts and a lengthy legal dispute with Intel).
At the time this was seen as a brave decision as having a second source available had been a key factor in getting customers to sign up for a new microprocessor design.
Here Richard Tedlow, who wrote Andy Grove’s biography, discusses the business case for the 80386.
Intel 386 Oral History Panels
There is a fascinating Oral History on the decision not to second source the 80386.
Amongst many gems, there is the reminder that the 80386 wasn’t the only (or even the main) 32 bit architecture being developed by Intel in the mid 1980s.
House: In fact, when I went to launch, well my team under me launched a 386 development there was push back within the company from Bill Lattin and the 432 group that that was their territory. They were going to handle 32 bits. And I had to position the 386 as, “Well this is just another interim product. I mean we can get this out quickly and we’ll just fill the gap until you come out with your real 32-bit machine.”
Jenkins: I wish I had a nickel for every 32-bit architecture we developed.
House: Most of them are worth a nickel, except for one.
Jenkins: I know. Have you ever counted them all up? I can’t tell. I mean, it’s like five or six isn’t?
For insight into the development of the 80386 there is another Oral History created by the Computer History Museum. Grove’s total focus on the 80386’s successors is alluded to at one point:
I bumped into one of the marketing guys years after I had left Intel and talked to him about what's it like at Intel now. "Well, there's only one product, and Andy Grove's the product manager." So total devotion to it.
The discussion also highlights the origins with the 386 of Intel’s ‘tick-tock’ approach to developing a new microarchitecture followed by a die shrink:
… because to come out with a new generation chip, you really don't have time to do everything that you'd like to do to optimize the layout . What happened was that a new chip would come out on the current technology and it would be an expensive die, but the users then would redesign new machines, new generation of PC's and that would take a while. And by the time they ramped into full production, Joe Schutz would have his compaction ready for it for mass production at a small die size and low cost. And it fit so well.
And Intel’s current CEO, Pat Gelsinger, even get’s a mention:
One of the difficulties of dividing the chip into units, which was probably more pronounced on the 386 than on previous chips, was that the majority of the engineers knew their unit but very few people understood how the chip as a whole really worked. So when we got to the debug on the first silicon we only had maybe a few people who really could get their overall view of things and figure out where to go and what to do. I remember Pat Gelsinger and Greg Blanck were in particular very helpful because they understood more of what was going on between the various pieces of the chip ….
Gelsinger, who was lead architect of the 80386’s successor, the 80486, was himself mentored by Andy Grove. It’s remarkable that he’s leading Intel almost 40 years after he was working on the development of the 80386!
Computer Chronicles and Eric Schmidt on the 80386
The introduction of the 80386 was covered by the TV program the Computer Chronicles. In the linked clip we can see a young Eric Schmidt (later of course CEO of Google) discuss why Sun, where he then worked, was porting its NeWS windowing system (interesting but long since discontinued) to the 80386.
With hindsight one can see the start of the end of the road for specialised workstation CPU’s like Sun’s own SPARC designs (and of course Google would be built on top of servers using Intel processors rather than Sun servers).
Grove on Strategic Inflection Points
We can see Grove speak about his approach to business at MIT in 1996 (when he was still CEO of Intel). It’s impossible to summarise what is a wide ranging talk, but it’s particularly interesting to hear him discuss globalisation and China, topics that are still absolutely central in 2023.
Book of The Week: High Output Management
Many readers will be familiar with Grove’s most famous book Only The Paranoid Survive which is about how to deal with major change and which was informed by Grove’s experience at Intel and especially with the decision to leave the memory business and to focus on the 80386 and its successors.
Equally deserving of attention, though, is High Output Management which sets out Grove’s approach to day-to-day management and which often features at the top of ‘recommended' reading’ lists for startup founders.
It’s a relatively short book - you can read it in a day or so - and it’s remarkably accessible, for example, starting with the example of a company whose business is delivering breakfasts!
There is an audio version of the book on YouTube for those who prefer listening to reading.
This week’s paid subscriber links focus on getting the best performance from the Intel CPUs that have followed the 80386. Please consider supporting The Chip Letter and becoming a paid subscriber.