Advertisement

Keil Buys ARM, Intel

November 09, 2005

Jack Ganssle-November 09, 2005

This week Keil Software, Inc., the company that makes compilers, simulators and debuggers for the 8051, C16x and ARM processors announced they purchased ARM Holdings plc (designers of the popular 32 bit CPU) and Intel Corporation, original purveyor of the 8051 plus lesser-known processors like the Pentium, Xeon, etc.

A Keil spokesperson noted: “These acquisitions further strengthen our ability to support our customers. By providing both the tools and the silicon we can insure the utmost compatibility in all phases of the development process.”

“Though Intel doesn’t sell 8051s anymore, they have unparalleled fab capability plus deep knowledge of microcontrollers.”

The spokesperson acknowledged they plan to sell Intel’s 32 bit business to AMD and retool the fabs to crank out 8051s using 65 nanometer geometry. A new campaign will encourage OEMs to brand their embedded products with “Keil Inside” labels.

On a more serious note…
ARM bought Keil last week. Keil is one of the great embedded tool companies. For nearly 20 years they’ve been a major force in the 8051 business, and now support other architectures as well. Their simulation environment is simply breathtaking.

The company’s press release states they’ll “[continue] to support our 8051 and C16x compilers within the uVision environment.” And I’m sure they will. But “support” could mean a lot of things.

Keil is a relatively small outfit. I know nothing of their financials, but having operated a tool company in this excruciating-difficult market, I can imagine that with more cash they could innovate even more. That has got to be good for consumers of ARM tools.

The companies aren't saying what their new strategy will be. Is it possible the new owner will discontinue or stop supporting the non-ARM tools? We saw something similar happen when Motorola bought Metrowerks.

Certainly, the revenue stream from a small operation like Keil can't be that important to ARM. Cutting off all 8051 support won't have much financial impact, and frees the company’s developers to focus more on ARM tools.

ARM has announced new forays into microcontroller versions of their processor, clearly a move to seize some of the venerable 8051 market. How does one both support a market and attempt to take it over?

I know the Keil people, and they're passionate embedded folks, smart and fun to be around. The 8051 is in their blood. But the decisions will be made by people who live and breathe ARM, whose compensation packages and bonuses stem from increased sales of the ARM design.

We'll see what the future brings. I sure hope that includes Keil’s traditional commitment to 8051 and C16x developers, with support of new variants and new toolchain features.

What do you think? Is this acquisition good for 8051/C16x developers? For ARM developers?

Jack G. Ganssle is a lecturer and consultant on embedded development issues. He conducts seminars on embedded systems and helps companies with their embedded challenges. Contact him at jack@ganssle.com. His website is www.ganssle.com.


My guess would be that ARM likes the Keil tools and brand loyalty and wants some of that for the ARM product.

The quote in the press release "As the MCU applications shift from 8/16-bit to 32-bit solutions,..." does not give me warm feelings that supporting the 8051 will continue to be a priority.

The death of 8/16-bit processors has been predicted for more than 20 years now. I wonder if strategic marketing people will ever learn that most products simply don't require more computing power.

- Richard Fellows


Pretty funny, you had me checking the calandar to see if it was April Fools day. Then again 30 years ago who'da thunk that a software company would be worth far more than GM, and probably be able to buy them if they were moronically inclined.

- Ken Rebitz


I think it stinks! Whenever a larger company buys a smaller independent company like Keil things always get worse for the other guys. There is hardly any justification for this other than the suppression of competition real or imagined. Of course the backlash from all this is that other compiler companies may gain new friends as Keil/ARM begins its anti competitive behavior. As others have pointed out there are many applications that simply do not require the complexity or power of an ARM processor and never will. ARM should just get over that fact and concentrate on serving those applications that can benefit from their higher powered offering. I guess if I want to get the full blown Keil compiler instead of the limited version I have now that I had better act fast.

- Don McCallum


If my guess (and I'm thinking your too) is right, ARM will orphan Keil's non-ARM portfolio for all but the largest customers and force new designs towards ARM.

One of the first things I do when selecting a processor is look for toolsets, and the fewer options there are, the less likely it is that I'll spec that chip.

Based on the debacle that is Metrowerks, I won't hold my breath for superior tools from a company controlled by a chip vendor :-)

- Ralph Hempel


As a recovered 8051 addict, I urge you to consider the following. Although there was some resistance at the time, the old dependable horse and buggy was displaced by the automobile. Now that newer processors run circles around the dependable old 8051 for less money and on less power, it's time to follow Keil's lead and embrace change. Maybe I'm out of step, but I haven't had too many requests lately for a cell phone, a fingerprint scanner, an MP3 player, a blood tester, or anything else for that matter, built on 8051, but I sure do see a lot of ARM, DSP, and nano power 16-bit applications. For those who still resist change, I have available a stockpile of 2716s, 2708s, and even some 1702s and 2102s you might like, and a dusty Data I/O 19 to go with 'em!

- Lindon Baker


Keil already has ARM compilers. Similarly IAR also has ARM compilers.

The new ARM chips like LPC2xxx from philips "looks" like any 8051 and development/debug processes are the same. I stopped caring about assembly instructions long ago. I will learn to ignore RISC/CISC differences too if the power and price is right.

- Sameer Cholayil


I hope ARM sells the non-ARM portion of Keil and allows the 8-bit products to continue to flourish. Wasn't it about a month ago we saw a new product announcement out of Japan for an automotive 4-bitter? 8 and 16 have a few more decades of life in them.

- Andy Kunz


Although, ARM devices is rapidly expanding in low end Embedded System market, there is hard to find a substitute for 8-bit 8051 architecture.

Therefore ARM will or will have to support Keil compiler for 8/16-bit 8051 architecture or in future you may find something called ARM-8051 architecture.

- Shridhar Pophali


A comment with a Teutonic flavor. A few months ago I drove with my colleagues on A4 autobahn in VW Passat about 120 MPH. At one point we were passed by the latest BMW M5 with a differential speed of approximately 50 MPH. My engineer-passenger said “My friend *had* one like this. Wouldn't buy another BMW ever.” I asked “Why?” He replied ”The car died on autobahn. Full tank of gas, battery fully charged. The guy was stranded and pissed. Got his money back, though.”

That car has well over 80 microprocessors on board. But none of them did anything useful for the guy calling on a cell phone (with ARM core) for help.

My point is this: does a consumer buy product by choosing what is inside (with the exception of Intel marketing people efforts)? The answer is -mostly- “no”. Do you care what is inside a router, mouse, toaster, alarm clock, bank's ATM? You expect it to route, “mouse”, toast, keep the time and dispense cash. And make the car go. That's all. However, the technological evolution is uncompromising, regardless of our habits. How many embedded systems engineers will design anything based on 8051, if the ARM-based uC loaded with memories and peripherals will cost $1? How many commerce web sites are built with COBOL?

What will happen to ARM/KEIL's in regards to 8051 business model in the future remains to be seen, for a very few people know the facts that led to this acquisition. And if KEIL would stop supporting 8051, there are fortunately other vendors who would be gladly taking orders for their '51 compiler. In this context, free market works well, IMHO.

- Roger Lynx

Loading comments...