editorial

Free the PC! Web anarchists attempt to challenge both Intel's dominance and the notion that complex CPUs are billion-dollar projects. by Jonah McLeod

I got a call from James Lee, senior consulting engineer at Seva Technologies. He told me about an ad hoc Web-based group that has a plan to collectively create--in cyberspace--a microprocessor that could rival Intel's Pentium. The idea is that a bunch of engineers would collaborate in their spare time to specify and then design the Pentium-rivaling product, and then make it available for free on the Web. Prophetically, the project is called the Freedom CPU Project. You can find out more about it at http://f-cpu.tux.org.

As James spoke, I thought about the similarities between the characters in William Gibson's Neuromancer--portrayed as fierce anarchists that operate in unfettered cyberspace--and our friends trying to create the public-domain "Pentium killer." But notwithstanding my science-fiction bent, their project appeals to me on a number of levels. First, there's the sheer scale of the project--I have to say I like big undertakings. Lately, for instance, I've been trying to convince various tool companies to simulate, synthesize, and place and route the Torch processor, a public-domain, MIPS-like CPU developed at Stanford University under an ARPA contract.

Another appeal of the project is its sheer audacity: It's trying to replicate the R&D effort of a billion-dollar semiconductor company, produce a superior processor, and then essentially give it away on the Web. What's more, according to the schedule, the processor should be unveiled at about the same time as Intel's Merced--sometime in 2000. Such boldness shouldn't go unheralded.

For edification, here's how the Freedom CPU Web site outlines the concepts behind the project:

"Hardware and software are really two instances of the same concept--an information-processing mechanism. Intellectual freedom, as the underlying idea behind Free Software and the GNU/GPL license, should be extended to encompass the entire desktop.

"From these two basic ideas followed the mission of the F-CPU Project--to design an advanced, high-performance 64-bit architecture that will easily take us well into the next century, and to work towards the first implementation of this new architecture: the F1 processor, compatible with x86 commodity hardware.

"Not only will the architecture and its implementation be entirely placed under the GNU/GPL (or a license very similar to it), but also the development process itself is designed to follow the principle of freedom: freedom to create, freedom to contribute, freedom to learn."

Just who is behind this project? The Web site reveals a primary team of four working to make the Freedom CPU a reality. Based in France, Andrew Derrick Balsa is a computer science major from the University of Geneva who claims to have designed a microprocessor at age 15--though his most recent work history is in information technology. Richard Gooch is an Australian astrophysicist who developed the Pentium Pro MTRR support in the Linux 2.1.x kernels and is also a hardware developer. The background of EDA tool designer Rafael Reilova, who is pursuing a master's degree in computer engineering at the University of Cincinnati, includes VLSI design, reconfigurable and adaptive computing, FPGAs, operating systems, and computer architecture. Brion Vibber is a film student and system administrator with no formal training in computers and a smattering of experience with Perl and Linux kernel hacking.

These guys are smart and educated, certainly; idealistic and brave, unquestionably. But are they able to challenge the hundreds of experienced designers and multimillion-dollar budgets at Intel? The natural reaction is to dismiss them as being unrealistic in their expectations, if not altogether loony. Yet their madness depends not only on the time-honored method of intellectual collaboration, but also on a recent technology--the Internet--that has already made fundamental changes in the way human beings communicate. Take note: Even if only two out of every hundred designers offers to help design the F1 CPU, the resulting team would far outstrip any single company's forces. With a little help from their friends, these four dreamers just might turn their harebrained scheme into the most glorious example of the power of the Web.

Can they succeed? Are other engineers likely to contribute to their efforts? Are you? Write to me at jonah@isdmag.com and let me know your thoughts.

To voice an opinion on this or any Integrated System Design article, please email your message to miker@isdmag.com.

integrated system design  December 1998