ARM's Way
Behind most technological successes lies a turbulent saga. Companies set up to do one thing often become successful doing something quite different.
IBM started with adding machines but became successful with computers. Intel started with memories but only became a major company with its microprocessors. ARM, founded to make personal computers, has had its biggest success selling intellectual property.
It takes visionaries at the top to change a company's direction and refocus it to playing on its strongest assets. IBM, Intel and ARM all had visionary leadership.
Earlier this month a stock market float valued ARM at $600m - in marked contrast to GEC-Plessey Semiconductors (GPS) which was sold in February for only $225m.
While GPS had revenues of $315m, ARM had revenues of $45m. While GPS assets, including four factories, had a book value larger than its sale price, ARM's had assets of $21m. While GPS employs 1,600, ARM employs 270.
So why was ARM valued nearly three times more than GPS? The answer, of course, is the expectation gap:- whereas GPS seems established, familiar, stable, not credited with great expectations, ARM seems to have a great future ahead of it - and not for the first time.
Visionary though they were, ARM's founders Dr Hermann Hauser and Dr Chris Curry could hardly have foreseen how their foundling would develop. Their intention in 1979, when they set up ARM's parent - Acorn Computer - was to make personal computers.
Their first product was the the Acorn Atom. Two years later came the product which made the company's fortunes - the 1981 BBC Micro. Over 150,000 were sold in the next two years.
In 1983, came the Acorn Electron. That year Acorn, growing exponentially, floated on the stock market with a valuation of £134m. The company was on a roll and looking around for its next product.
Almost accidentally, it stumbled onto the track that was to lead, 15 years later, to another stock market bonanza.
The Atom, BBC Micro and Electron had used an off-the-shelf 8-bit microprocessor, called the 6502, made by the Western Design Centre of Phoenix, Arizona.
"16-bit and 32-bit microprocessors were beginning to appear and the LISA had been launched which made us realise that we needed a Windowing system, so some of my more academically inclined colleagues, like Steve Furber, argued that we needed a new architecture," recalls Chris Turner, chief hardware designer of the BBC Micro.
His boss agreed. "We came to the realisation that you can’t build computer systems without having control of a microprocessor core", remembers Hauser, "I asked Intel for the 286 core and they said we're not interested in selling cores - only microprocessors."
There were other 16-bit microprocessors. "We looked at National's 16032 and Motorola’s 68000 but they didn't suit our design style. They had very complicated instruction sets giving poor interrupt response. Basically they were too slow," explains Steve Furber, now Professor of Computer Science at Manchester University.
Acorn's engineering team considered developing their own microprocessor, but there were tremendous obstacles.
"We knew that it had taken National 200 years of development time for the 16032 and Acorn couldn’t afford that - we only had 300 people at the time," recalls Furber.
"Then we came across the Berkeley RISC. A group of graduate students had built a microprocessor with only a tiny percentage of the resources used by National. It was simple, it addressed the interrupt problem, and it seemed something worth looking at," recollects Furber, "in late 1983, I started working closely with Sophie Wilson who had developed all the versions of BASIC for the BBC Micro."
"The general view was that microprocessors had a mystique - that they were designed by very special people. I'd never designed a microprocessor and everything I knew about computers I'd learned at the Cambridge University Processor Group where people met to make computers for fun," recollects Furber.
"Sophie and I went on a trip to Phoenix to the Western Design Centre which was working on a 16-bit version of the 6502. We found it to be a cottage industry working in a bungalow in a back street. That gave us confidence," recalls Furber, "Sophie started playing with instruction set design. Our mentality was let's have a go at building a microprocessor." The next problem was to persuade the boss.
"Hermann was a great guy to work for," says Furber, "if he had confidence in you technically, he'd back a crazy idea. Building our own microprocessor was a crazy idea - but he backed it."
"Steve is one of the brightest guys I've ever worked with - brilliant," says Hermann Hauser, "but when we decided to do a microprocessor on our own, I made two great decisions - I gave them two things which National, Intel and Motorola had never given their design teams: the first was no money; the second was no people. The only way they could do it was to keep it really simple."
Furber set to work defining the architecture, while Wilson developed the instruction set. Hauser remembers: "While IBM spent months simulating their instruction sets on large mainframes, Sophie did it all in her brain."
Once the hardware architecture was established, the actual design of the silicon was done by Robert Heaton and Jamie Urquhart. There were three peripherals chips: a video controller designed by Tudor Brown, an I/O controller designed by Mike Muller and a memory controller designed by Furber. "They were a remarkable group of people. They did phenomenally well," says Furber.
Work started in October 1983 and first silicon was run on April 26th 1985. It had taken 18 months of calendar time and five man years of work time.
The microprocessor, christened Acorn RISC Machines (ARM), was first used in the company's Archimedes computer. However, Archimedes was launched while Acorn was suffering financial problems.
Sales operations in Germany and America had proved too expensive, while technical problems with its chips hit the Electron, curtailing sales. In 1985, Olivetti came to the rescue.
Five years later came the decision to spin out the microprocessor team as a separate operation. "We felt that ARM could not become a successful microprocessor while it remained within a systems company, because no computer company would use a microprocessor which was controlled by a competitor who could block them out," explains Hauser.
So, in 1990, backed by Apple, VLSI Technology and Nippon Investment and Finance - with Acorn retaining a 38 per cent stake - the microprocessor group was spun out under the name Advanced RISC Machines.
At the time, there was no clear idea how the ARM microprocessor would be exploited. "We gave no instructions to our management team, we left it to them to decide the business direction," says Hauser, "fortunately we have an absolutely brilliant CEO in Robin Saxby who defined a new business model by inventing the chipless, fabless company."
The rest is history. Managing his virtual chip company from a virtual office at 30,000 feet courtesy of Messrs Boeing, Saxby has criss-crossed the planet in his search for licensees - 26 so far, and growing. No one in the global electronics industry is unaware of ARM, and Saxby's personal shareholding is estimated to be worth between £10m and £15m.
The ARM alumni
Source: Electronic
Weekly
