Hacker News Reader

Comment by codeulike | Hacker News Reader

The people that designed and built the first ARM processor (in 1985), for those that dont know. Among other things.

ARM originally meant Acorn RISC Machine.

The Acorn Atom, its successor, the BBC micro family, then the Archimedes, the first 32-bit RISC personal computer (for which the ARM was designed)...

Their impact cannot be overstated.

I had an Archimedes! It was amazing.

Weird to think how many billions of ARM chips there are now. Another 40 million made every day apparently.

IIRC, at launch time, its floating point performance was close to a Despro 386 with a 387 coprocessor.

Is that good or bad?

That was pretty good as it didn't have a floating point coprocessor at launch and was entirely software implemented. You could buy a card with an AT&T WE32206 copro for it which made it pretty damn fast. I think they had dedicated copro socket later on. I didn't buy one because all of them were stupid expensive.

The Archimedes absolutely blew the popular 16-bit 68000 machines (Ataris and Amigas) out of the water, that's for sure. Although to be fair it was more expensive.

I had BASIC games I'd written for the BBC Model B and when I ran them on the Archimedes everything moved around the screen too fast to see. It was amazing.

But wasn't the Atom a copy of the MIPS? I was under the impression MIPS did it first but ARM did it better?

You mean ARM not Atom? I dont think the ARM was a copy of anything, but ARM and MIPS both use a RISC approach.

If anything, I remember hearing a presentation where they claimed the ARM was influenced by the 6502 they used in the Atom an the BBC. The 6502 is an absolutely delightful machine to program for.

The jump delay slots, I think, are something they share with MIPS, but I'm not aware of anything else.

There's some stuff about it in this Sophie Wilson interview, which I found an interesting read: http://www.computerhistory.org/collections/catalog/102746190

It influenced their decision to do their own, first and foremost. Basically Bill Mensch at Western Design Centre provided them evidence by example that a tiny team could build their own CPU.

> a tiny team could build their own CPU

I almost did that in college. It was a beauty - a stack based CPU that could run an almost decent FORTH on metal. I wonder if it would have worked if actually built.

Of course, ARM is a much more complex thing than my toy.

Atom means the Acorn Atom:

https://en.wikipedia.org/wiki/Acorn_Atom

Nope. ARM and MIPS used concepts from RISC but that's where the similarity ends.

And the low budget Acorn Electron, a kind of stripped down and partially compatible BBC.

As I understand it, the Electron was a bit of a commercial failure. However this caused remaining stock to be dumped (at least in the Netherlands) at rock bottom prices. My primary school teacher was a big fan of them, and installed a couple of them at our primary school. I got to borrow one for a summer break (1990?). My first forays into Basic programming, as well as trying to understand German, since that was the language in which the accompanying Basic manual was written.

Reply on news.ycombinator.com