
Scientists are pushing innovation to the limit, trying to shape chips that are smaller but more powerful. Take a quick look at the chip's transition in the last 30 years and what the future holds.
Vipin Kumar
SUMMER, 1972. A fledgling company called Intel releases a microprocessor named `8008'. Two computer freaks, teenagers Bill and Paul see an opportunity in that which no one else could: that computers in future would be built around these kinds of microprocessors. When the industry overlooked the potential of the chip, Bill and Paul foresaw that it would make computers small and computing-affordable. They decided to write software to run on such machines.
Years later, Bill wrote in his book, `The Road Ahead': ``Paul and I realised that this first microprocessor was very limited, but he (Paul) was sure that the chips would get more powerful, that computers on a chip would improve very rapidly. This insight of Paul's was the cornerstone of all that we did together later, including the founding of Microsoft.''
It's been almost three decades now since Bill Gates and Paul Allen dreamt of writing ``new and innovative'' software for machines based on newer and more powerful chips. They went on to create history. But the quest for more and more power from the tiny chip still continues. In this story, we look at the future of microprocessors, some issues therein and a few recent developments that promise to -- what else, deliver more powerful chips!
Moore's Law -- for how long?
Gordon Moore, who would later co-found Intel with Bob Noyce, predicted in 1965 that the number of transistors on a chip would double every 18 months. This forecast, which later became an industry axiom -- `Moore's Law' -- has proved right till date. ``Imagine what would happen to automobiles if Moore's Law is applied,'' jokes Patrick P. Gelsinger, Vice-President and Chief Technology Officer, Intel Architecture Group. ``You need to fuel your car only once in its lifetime, every time you take it out for a spin, you will get a second car and it will also be invisibly small!'' Fortunately or unfortunately, Moore's Law pertains only to computer chips, but now the question doing the rounds is: for how many more years would it hold good?
Gelsinger believes that Moore's Law would be relevant even in 2010, and predicts that by then a single piece of silicon would contain one billion transistors, delivering a power of 30 GHz, up from today's 1.8 GHz! While that kind of power offers mind-boggling computing power, it also throws up a serious challenge. Such power, coupled with the shrinking size of the chip, would generate extreme heat. At these levels, the temperature in a system exceed that in a nuclear reactor and be equivalent to that in a rocket nozzle! Gelsinger is hopeful that new methods will emerge to address this issue and already a micro-radiator, built into the silicon, is being developed to cool down the processor. Also, the search for a substitute for silicon, which is the fundamental material at the heart of microprocessors, is on. But people like Gelsinger believe that ``there is nothing on the horizon to replace silicon.''
IBM's punch -- altering silicon
Against this backdrop, IBM's recent announcement that it has altered silicon to increase chip speed up to 35 per cent, assumes significance. This new technology, called ``Strained Silicon,'' stretches the material to speed up the flow of electrons through transistors, leading to greater performance. It takes advantages of the natural tendency of atoms inside compounds to align with one another. In the strained silicon, electrons experience less resistance and flow up to 70 per cent faster, which leads to chips that are 35 per cent faster without having to shrink the size of transistors. IBM expects this new technology to be used in products by 2003.
Itanium arrives
After seven years of joint development with Hewlett-Packard, Intel launched its much-touted Itanium processor towards May end as a surprisingly low-key affair; the company just issued a press release announcing the launch. The world's largest chipmaker believes Itanium will help it further strengthen its dominance in the industry, though many feel that the $1-2 billion that the company apparently invested in the project would go down in history as ``one of the world's worst investments.''
However, besides HP, a number of companies such as Compaq, Dell and IBM have pledged their support to Itanium and have come out with a new range of servers and workstations. Besides, Intel is investing millions in encouraging software developers to write applications specifically for the new architecture. Meanwhile, Intel's rival, Advanced Micro Devices (AMD) is also developing a new 64-bit chip to take on Itanium, though it is not expected before next year.
When all these things take place, it is safe to assume that another Bill and Paul, somewhere in this globe, would be wondering how they can come up with new technologies to match such computing power.