Here’s the story of Apple’s failed attempt to build their own CPU architecture, starting in 1986 or so and ending in '89.
(There’s a modal to sign up to email updates but you can click “Continue Reading” to ignore it.)
The Hyperchannel links the Cray to several nets of Apple Macintoshes via a VAX-11/785, a VAX-11/780, and a Sun/2. The Macs are linked to the VAXes and the Sun via AppleTalk in series with Ethernet. Ethernet and AppleTalk are bridged by a converter system from Kinetics Inc.
…the Cray was a ‘vector machine’, built to run a certain class of programs quickly: those that could be broken up into smaller tasks and run in parallel. This wasn’t an approach that could be easily applied to the tasks that the Aquarius team needed. In due course, the team found that the Sun workstations that they attached to the supercomputer could run their software just as quickly as the Cray itself.
All the new features included in the Scorpius design meant that they lost one of the key advantages of the RISC approach: simplicity. Small teams at UC Berkeley and Stanford, and then at Acorn in the UK, had successfully designed and built working RISC processors in little more than a year. At Apple, even with a team of fifty, the design struggled to make progress.
It may be interesting to revisit the RISC-based 6502 project, we mentioned earlier, in this context, as this is the kind of thinking it would have had to compete with.
(The RISC-6502 project could have been indeed done by a small team and would have probably been somewhat opiniated. But we can also imagine, why this would lead nowhere and why Pete Foley “was taken for a walk”.)
This is not a new phenomena … There is a term “Alpinist” which was originally coined by a small group of mountaineers who abandoned the traditional “siege tactics” of conquering mountains (base camp, camp 1, camp 2, the final push, etc.) and just did it with the minimalist of equipment plus a lot of knowledge…
So maybe that’s what’s happened there? A small team just gets on and does it vs. big team, committees, meetings, large, expensive hardware thrown at it too, etc.
I am still unclear why they would ever consider a Cray for this. As the article notes, the Cray only has an advantage when its vector units are going, and there’s basically nothing in IC design that makes use of the vector units.
I guess the machines could be specced out with more memory and had fast access to it, but in 1986 one could also buy a VAX 8800 which I would guess would provide almost all of the performance for almost none of the price, assuming you needed larger memory than workstations of the era.
Certainly none of this should have been surprising to the group, especially considering they were a processor design group. One suspect they ended up with this machine just because no one in the company said no.
I do wonder if it was overenthusiastic management rather than engineering need - the article has a quote from Cray Channels in an article by Kent Koeninger:
our need for supercomputer capabilities was clear, and the lack of bureaucratic obstacles within the company hastened the process considerably
So, perhaps some status move whereby management knows best, and the most powerful managers Get Things Done. And it’s a PR event, to get a Cray.
There is one parallelisable task in chip design that I’m aware of, which was implemented in Plessey on a VAX using custom microcode, and that’s fault simulation (in a logic simulator.) But I don’t know if it’s the right kind of parallelism for a vector machine.
Edit: From the LA Times " A Peek at the Computing Future" - “It has even given Kent Koeninger the job of encouraging research on the Cray and a job title that only Apple would bestow–”Cray Evangelist.”.”
Edit: from Kent’s LinkedIn profile I see he worked for NASA using Crays, then for Apple where they bought a Cray, and then for Cray.
Edit: reading the above-linked article from Cray Channels, I see two interesting case studies, one being an exploration of minimum necessary arithmetic precision in some chip, and the other being computational fluid dynamics for disk head design. Also noteworthy that the Cray was to support Apple’s needs for supercomputing, not necessarily just for chip design.