Creating the Digital Future: The Secrets of Consistent Innovation at Intel - Hardcover

Yu, Albert Y.c.

Hardcover

ISBN 10: 0684839881 ISBN 13: 9780684839882

Publisher: Free Press, 1998

Synopsis

Nowhere are the perils of success more threatening than in the fast-paced personal computer and semiconductor markets. Here today's exciting innovation quickly becomes yesterday's news. In this unforgiving climate, Intel Corporation has established itself as the undisputed trailblazer of microprocessor and silicon technologies. What exactly are the secrets of consistent innovation at Intel? Top microprocessor expert and Intel insider Albert Yu reveals that the key lies in Intel's ability to reinvent itself. In this fascinating and instructive book, Dr. Yu shows how Intel "obsoletes" its own products and relentlessly raises the bar to the next level. He brings the reader into the results-oriented, hyper-innovative, creative Intel culture that thrives on fresh ideas, risk-taking, and learning from failure. Yu shows how volume is a key to profits and describes how interactions between customers, marketers, and engineers often generate sparks that spawn great products. Above all, Dr. Yu demonstrates that Intel has prevailed by learning lessons from its mistakes in fierce, bitter competition with Motorola and Sun's SPARC for strategic leadership of the high-technology world.

About this title

About the Author

Albert Yu is Senior Vice President and General Manager of the Microprocessor Products Group at Intel Corporation. Dr. Yu is the author of many technical and professional articles and frequently lectures throughout the world on the future of microprocessors, computing, and the management of technology. He lives in Los Altos, California.

Reviews

The invention of the microprocessor at Intel in 1971, as Yu tells it, was instigated by a customer request and represented a leap into the unknown. Since then, he writes, the California-based microchip giant has tried to make "obsolete" its own products with better ones before competitors do it first. While Intel senior v-p Yu offers details from the front of Intel's furious chip-building competition with Motorola, IBM and Sun Microsystems and explains how Intel's breakthroughs have affected the computer and electronics industries, most of this concise report is a savvy, straightforward primer for managers, business and computing professionals. While some business maxims are peculiar to high-tech (take Moore's "law," which states that the "number of transistors on a semiconductor chip doubles approximately every 18 to 24 months"), many of Yu's recommendations are drawn from business basics: focus on delivering measurable results; nurture a creative, risk-taking atmosphere; provide nonstop on-the-job training; develop a cohesive product line. Yu does, however, flesh out each strategy with pertinent case examples, making his manual a useful springboard for those designing or developing high-tech products in many fields. Editor, Robert Wallace.
Copyright 1998 Reed Business Information, Inc.

Excerpt. © Reprinted by permission. All rights reserved.

CHAPTER ONE: PLAN THE COMPUTING FUTURE

At the Microprocessor Forum in San Jose, California, on October 14, 1997, in front of 1,200 key players in the microprocessor and personal computer industry, top Intel architect John Crawford and his colleague at Hewlett-Packard unveiled the future of computing. A totally captivated standing-room, only audience learned for the first time the technical details of the new EPIC technology for Intel's upcoming 64-bit microprocessor product line. Several microprocessors incorporating this technology are under development, and the first product, codenamed Merced™, is expected to be on the market by the turn of the century. Merced was labeled the "Killer Chip" by Fortune magazine. These 64-bit microprocessors represent the next giant leap in performance and capability that will allow the microprocessors to power the full spectrum of computing from its current base of personal computers to mainframes, supercomputers, and beyond. Several leading companies such as Microsoft, Hewlett-Packard, Compaq, IBM, and Sequent presented their respective software and hardware product plans to take full advantage of the power of the Merced chip. In addition to the 64-bit disclosure, Fred Pollack, Intel's head of microprocessor planning, detailed our complete microprocessor product road map, including both the 32-bit and the 64-bit products, all the way out to the year 2001. How do we plan and develop our product road maps so far into the future? What are the technological and economic trends for the future of digital products? What does computing look like into the next millennium?

MICRO 2001

Charles F. Kettering, former director of research at General Motors, once said, "I am interested in the future, because I have to spend the rest of my life there." My key managers and I spend lots of time planning future generations of microprocessors. No matter how good a processor Intel has just introduced, the next one -- usually the next two -- are well underway, and that's where our attention is focused. In 1988, three Intel colleagues and I mapped out our view of what the leading microprocessor would be in the year 2001, and called it Micro 2001. We predicted that the leading-edge microprocessor would have nine million transistors on a chip by 1997 and 100 million by 2001, due to silicon technology advances. We were close: the Intel Pentium II processor in 1997 has over eight million transistors, and I believe that we are well on our way to making the 100-million transistor microprocessor a reality by 2001. In addition, we projected that the microprocessor of 1997 would perform eight times faster than the 486, based on the increased transistor count. The actual Pentium II processor performance turns out to be a lot higher: about 25 times faster than the 486. Clearly there has been an enormous amount of innovation in computer architecture over the last few years that we did not foresee in 1989. This includes the concepts of dynamic execution, a novel idea that speeds up microprocessor performance, and the Dual Independent Bus, both of which are featured in the Pentium II processor and will be described in more detail later. From 1997 to 2001, microprocessor performance is expected to increase by another factor of 10. When we published our projections in 1989, many people thought we were being too optimistic. It turned out that our forecasted performance jumps were way too conservative!

MOORE'S LAW REIGNS

How could we predict in 1988 the enormous increase in the number of transistors that could be placed on a chip by 2001? In 1965, back in the early days of silicon integrated circuits, Gordon Moore, one of Intel's founders and then head of Fairchild Semiconductor Research and Development Laboratory, set forth a principle concerning the pace of semiconductor advances that has since become known as Moore's Law. This "law" states that the number of transistors on a semiconductor chip doubles approximately every 18 to 24 months. For microprocessors, the doubling of transistors has occurred about every 24 months, as shown in Figure 1. This simple statement, one of the most significant in the semiconductor industry, has held true since 1965 and promises to be a good predictor for semiconductor advances well into the twenty-first century. Now chairman emeritus of Intel, Gordon has been the industry's visionary for decades. He is an unassuming man with incredible insights on technology and business. At meetings he usually sits quietly listening to various discussions and arguments until suddenly he comes up with a clear and crisp statement that hits at the heart of a matter and establishes a guiding principle. When Gordon speaks, everyone listens.

Let me explain how Moore's Law came about and what it means. One of the most significant inventions of the twentieth century -- before the microprocessor -- was the integrated circuit. Integrated circuit technology made the semiconductor memory and microprocessor possible. Robert Noyce, another founder of Intel, helped invent the integrated circuit at Fairchild Semiconductor in 1957. Noyce discovered that instead of hooking up individual transistors to build an electronic circuit, it was possible to build the whole circuit on a single chip of silicon by connecting the various elements, such as transistors, capacitors, and resistors, with aluminum interconnections on the chip itself. As the circuitry was all integrated on a single chip, it was called an integrated circuit, or IC. The benefits were immediately apparent. A single IC package could replace an entire circuit board, saving space. It was much more reliable and cost far less than the board. The ability to build complex electronics functions on a single piece of silicon was a major technological breakthrough and became the foundation of semiconductor memories and microprocessors that have fueled the digital explosion.

THE MIRACLE OF SILICON

It is the unique electrical and chemical properties of silicon -- silicon dioxide and thin aluminum layers -- that make possible the dynamics of Moore's Law. It turns out that as one builds smaller MOS (metal oxide simiconductor) transistors in silicon, the transistors perform faster and consume less power. Silicon dioxide is a perfect insulator because, as it gets thinner for smaller devices, it continues to perform predictably and reliably. Aluminum is a perfect conductor to form electrical connections; it just happens to stick to silicon and silicon dioxide really well to form very reliable interconnections. The combined properties of these elements make the evershrinking transistor and more complex integrated circuits possible.

I know of no other technology that gets better as the element gets smaller. For example, neither internal combustion engines nor electric motors get better as they get smaller. The miracle of silicon provides the foundation for the whole high-technology industry. This new industry grew rapidly from 1957, and Fairchild Semiconductor was one of its leaders.

As silicon technology advances and transistors get smaller, more of them can be packed on a single chip to perform more complex electronics functions. For example, let's say the next generation of silicon technology allows us to pack 2,000 transistors on a single chip, instead of the maximum of 1,000 that we could place in the previous generation. What happens? First of all, one can build a 1,000-transistor chip on the new technology that is half the size of the previous-generation chip. Because the cost of a chip is roughly proportional to its size, the new chip will cost only about half what the old chip costs. In addition, the smaller chip will perform at higher speeds and consume less power. Therefore, it can be sold for less than the old chip, while performing better in every way! This clearly is more attractive to custome

Edition details

• Publisher: Free Press
• Publication date: 1998
• Language: English
• ISBN 10: 0684839881
• ISBN 13: 9780684839882
• Binding: Hardcover
• Number of pages: 224