If you own a smartphone or a tablet the chances are that it’s processing package is based on ARM’s energy efficient architecture. Android, iOS and Windows Phone all use ARM based chips, which makes these operating system agnostic processors some of the most ubiquitous devices on the planet. That doesn’t mean ARM doesn’t have any competition, however. Intel is trying very hard to break ARM’s dominance, and Android can also run on Imagination’s MIPS-based processors. But for the time being ARM is the king of mobile computing. And now with the advent of its 64-bit ARMv8 architecture, ARM might start to make in-roads into the server room, a market which until now has been one of Intel’s traditional strong-hold. So how did ARM get to this place of dominance? Where did it come from? What makes it so special? And where is it going? Come with me on a journey through the history of ARM, from its very beginnings to 2014 and beyond.
In 1983 the company started its Acorn RISC Machine (ARM) project and the resulting Reduced Instruction Set Computing (RISC) processor would eventually become known as the ARM1.
It all started with the British Broadcasting Corporation (BBC). Back in the late 70’s/early 80’s, the BBC created an educational project called the “BBC Computer Literacy Project.” It was envisioned as a series of television programs based around a common computer. After the bidding process a Cambridge based company in the UK called Acorn Computer won the right to build what would become known as the BBC Micro. The BBC Micro became the standard educational computer for schools across the UK and Acorn became a household name. By the mid-80’s Acorn was posting profits in the millions and as a result of its new found financial freedom, Acorn set about to create the successor to the BBC Micro.
As part of the project to replace the BBC Micro Acorn started looking into creating its own processor to replace the 6502. In 1983 the company started its Acorn RISC Machine (ARM) project and the resulting Reduced Instruction Set Computing (RISC) processor would eventually become known as the ARM1. Then came the ARM2 and in 1987 the first consumer computer based on an ARM chip, the Acorn Archimedes. By 1990, Apple along with VLSI Technologies (the company who actually fabricated the processors for Acorn) became interested in new ways of using the ARM concept and as a result Acorn spun off the design team into a new company named Advanced RISC Machines Ltd. This was shortened to ARM Ltd when its parent company, ARM Holdings, floated on the stock exchange 1998.
At this point ARM has established itself as a Intellectual Property (IP) licensing company. What that means is that ARM creates processor designs but it doesn’t actually produce any physical products. Instead the technology (the intellectual property) is licensed to other companies who design and build their own processors based on ARM’s design. By 1997 ARM had licensed its technology to dozens of other organizations including Cirrus Logic, Texas Instruments, Samsung, DEC, Hyundai, Lucent, Philips, Rockwell and Sony.
At around this time a historical irony occurred, one which I was personally involved in. ARM had licensed its technology to the Digital Equipment Corporation (DEC – the company that brought us Ethernet, the PDP-11, the VAX, and the 64-bit Alpha chip). DEC in turn produced the StrongARM processor which at 233 MHz only drew one watt of power. DEC had lots of projects in development for the StrongARM processor. I was working for DEC at the time and I was partially involved in one of those projects.
Ironically Intel still holds an ARM architectural license, the highest level license granted by ARM, which it retained when it sold XScale to Marvell.
One morning when I came into work we were told to stop everything and to give back all our StrongARM based equipment. It later transpired that DEC gave the StrongARM technology to Intel as part of a lawsuit settlement. Intel used StrongARM to supplement its i960 line of processors and later developed its own high performance ARM-based implementation named XScale, which it sold to Marvell in 2006. Fast-forward eight years and Intel are desperate to get into the mobile space using its x86 architecture, but ironically the chip giant still holds an ARMv6 architectural license, the highest level license granted by ARM, which it retained when it sold XScale to Marvell.
After joining the London FTSE 100 index, ARM has gone from strength to strength. In addition to acquiring smaller companies, it has also licensed its IP to some more big names including Mitsubishi, Motorola, Fujitsu, Seagate, and Broadcom. The latter is the company which makes the processor used in the popular Raspberry Pi. By 2002 ARM’s partners had shipped over 1 billion ARM based microprocessors and in 2007, the year it announced the Cortex-A9 design, the company said that 5 billion ARM powered processors had been shipped. And that number just kept on growing. In 2008 it reached 10 billion and now in 2014 it is a staggering 50 billion. If each processor shipped was a LEGO brick you could make a row that went around the world 7.5 times!
ARM chips have provided the processing power for some of the most iconic devices released in the last decade. The Nintendo DS has sold more than 155 million units worldwide since it was launched in 2004, it uses two ARM processors. Canon’s EOS 5D Mark II was launched in 2008 and it uses an ARM processor in its Digital Imaging Core (DIGIC). The Fitbit uses an ARM Cortex-M3 processor, while the Nest Thermostat uses an ARM Cortex-A8 processor, and so on.
ARM chips have provided the processing power for some of the most iconic devices released in the last decade.
What of the future?
The smartphone and tablet revolution is well underway but there are new challenges for ARM up ahead. Wearables and the Internet of Things (IoT) require smaller and more power-efficient processors than those used in smartphones. At the other end servers, laptops and desktops require higher performance processors, with energy-efficiency, but without the constraints of running off a battery. The amazing thing about ARM’s processors is that they are very versatile.
That means that ARM is already meeting these challenges. Many of today’s wearable devices are already using ARM-based processors while at the other end you find ARM processors in Chromebooks and now in servers. AMD is already sampling four or eight 64-bit ARM Cortex-A57 core based processors, specifically designed for the server market. Also, Amazon is rumored to be looking to use ARM-based processors in its next generation of cloud computing.
So from wearables to smartphones, from smartphone to tablets, from tablets to Chromebooks, and from Chromebooks to servers, ARM has its processors in just about everything. The only market where ARM isn’t really present is the desktop market. That remains an Intel strong-hold, however it seems that the desktop market is shrinking, and although it won’t disappear, it certainly won’t be the cash-cow is has been previously.
It is estimated that 60 percent of the world’s population, or some 4.3 billion people, interact with a device using an ARM chip every day. If ARM can dominate in wearables, IoT and in the energy-efficient server room then that number is going to increase, substantially. What was the first ARM powered device you used? Please tell us in the comments section below.
What was the first ARM powered device you used?
This article was brought to you by ARM.