Comment Re:Too much hype, too long to deliver (Score 1) 668
As the clock speed goes up, and as the other processors find their limitations and drop out of the race, the Itanium will look better and better. There is, however, a large investment in time and software that must be made before it becomes truly useful. It is unlikely that MS is going to support more than one architecture simultaneously for the desktop or server as it tried to do for x86/alpha.
That was said about every other rival architecture to x86 - and every single one of them has either been marginalized or killed off.
The brute fact of the matter is that a chip's ISA just isn't that important to its performance. When the limitations of an x86 chip's microarchitecture become too great, its designers are free to create a new one and slap on a (now cheap) x86 decoder. If they need a few more instructions, they add them.
Because we tend to classify microprocessor hardware by its ISA, it's easy to lapse (perhaps even subconsciously) into believing that chips of a common ISA must necessarily share some deep hardware heredity. But they don't have to, and history teaches that microprocessor hardware innovations are largely orthagonal to ISA. Most of the improvements to IA-64 chips will probably benefit regular CPUs as well (for an interesting parallel, consider Linus's comments re: Monolithic vs. Micro kernels). And with the absurd amount of money pushing x86 development, such improvements won't take long to trickle down either.
That was said about every other rival architecture to x86 - and every single one of them has either been marginalized or killed off.
The brute fact of the matter is that a chip's ISA just isn't that important to its performance. When the limitations of an x86 chip's microarchitecture become too great, its designers are free to create a new one and slap on a (now cheap) x86 decoder. If they need a few more instructions, they add them.
Because we tend to classify microprocessor hardware by its ISA, it's easy to lapse (perhaps even subconsciously) into believing that chips of a common ISA must necessarily share some deep hardware heredity. But they don't have to, and history teaches that microprocessor hardware innovations are largely orthagonal to ISA. Most of the improvements to IA-64 chips will probably benefit regular CPUs as well (for an interesting parallel, consider Linus's comments re: Monolithic vs. Micro kernels). And with the absurd amount of money pushing x86 development, such improvements won't take long to trickle down either.
