It was never a question of whether, but only of when and why. When would microprocessor manufacturers be forced to slow the primary propulsive force for their industry--namely, the biennial release of new chips with much smaller transistors and much higher clock speeds that has made it so attractive for computer users to periodically trade up to a faster machine? And would it be fundamental physics or simple economics that raised the barrier to further scaling? The answers are: in 2004, and for both reasons.
Production difficulties bedeviled almost every major semiconductor firm this year, but none were more apparent than the travails of Intel, the flagship of the microchip business. The company delayed the release of "Prescott," a faster version of its Pentium 4 processor, by more than six months as it worked out glitches in the fabrication of the 125-million-transistor chip. When Prescott did finally arrive, analysts were generally unimpressed by its performance, which was only marginally superior to the previous, 55-million-transistor Pentium 4. The company recalled defective batches of another microchip, postponed the introduction of new notebook processors, and pushed to next year a four-gigahertz Pentium model that it had promised to deliver this autumn.
This article was originally published with the title A Split at the Core.