Intel Centrino/Pentium-M Notebook Roundup: Dell, FIC and IBM Examinedby Matthew Witheiler on March 12, 2003 11:22 AM EST
- Posted in
Before we jump into looking at new notebook solutions made possible by Centrino and the Pentium M, let's take a quick look at the preexisting solutions that notebook producers had when designing thin and light and ultraportable notebooks.
The chip that really began the thin and light and ultraportable revolution was the Transmeta Crusoe. Announced in 2000 the chip was received with an unusual amount of hoopla. Not only did the Crusoe promise to run at a fraction of the power required by competing mobile processors, it had the support of a computer icon: Linus Torvalds. On paper the Crusoe looked to be an exciting product. Using a Very Long Instruction Word (VLIW) engine with a code morphing wrapper to convert x86 code into VLIW, the chips specifications indicated that it would be powerful and not power hungry.
A number of notebook manufacturers turned to the Crusoe when designing new thin and light and ultraportable computers. Early adopters of the chip included Sony and Toshiba, both of whom offered full Windows machines in sizes not previously possible. The looks were impressive but the performance was not.
As we saw in our
More recent iterations of the Crusoe processor have helped increase performance but only marginally. The Crusoe TM5800 improves upon the TM5600 we originally tested by offering a new version of code morphing software and a smaller manufacturing process (.13 micron) which scales in clock speed up to 1.0GHz currently. This helps to increase performance somewhat but with initial performance along the lines of 533MHz Celeron means that quite a bit of speed must be made up in order to make the processor competitive.
Along with the slow performance of the Crusoe processor came another caveat: battery life, although good, is impacted by the slow performance of the processor. Think about it this way: if it takes you twice as long to complete a task with a CPU that uses half the power, you really are not saving any battery life over a faster and more power hungry CPU. At idle the Crusoe may save power but as soon as a task is thrown at it the chip is brought to its knees. Although a valiant effort and certainly revolutionary in its design, the Crusoe still left a lot to be desired.
Intel's response to the Transmeta Crusoe came in the form of new Pentium III-M chips. Intel took advantage of a Pentium III-M die shrink (from .15 micron to .13) to produce some lower speed Pentium III-M chips that had significantly less power requirements. Intel's new chip were the low volt Pentium III-M and the ultra low volt Pentium III-M.
With the current low volt Pentium III-M chips currently available in clock speeds up to 1.0GHz and ultra low volt Pentium III-M scaling up to 933MHz, the chips proved to be a good deal faster than even the fastest clocked Crusoe processors. Plus the power consumption of these chips was reduced to 1.10 volts and 0.975 volts respectively, down from the 1.35 volts required to power a full Pentium III-M chip. The low volt and ultra low volt Pentium III-M chips have steadily gained support since their release and more and more thin and light and ultraportable solutions are finding one of these two chips at the helm.
While the low power Pentium III-M chips did provide a nice alternative to the Transmeta Crusoe line of processors, they still did not perform up to the speed of their desktop counterparts. The chips are based on desktop processor technology from about two years ago and remains tied down by an aging mobile chipset. Desktop parts have been increasing in speed steadily for months now while the low power Pentium III-M chips had their hands tied by the aging processor and chipset technology.
Enter the Pentium M processor and Centrino technology. The Pentium M processor takes aim at both the Transmeta Crusoe as well as the ultra low volt and low volt Pentium III-M chips, seeking to find its way into thin and light and ultraportable solutions. We have already seen that the performance of the Pentium M represents a giant leap forward in low power, high performance mobile processing. Now it is time to look at how manufacturers are utilizing the new technology to produce high performance mobile systems in form factors not previously possible.