
www.Usenet.com
| <-- __Chronological__ --> | <-- __Thread__ --> |
In article <[EMAIL PROTECTED]>, [EMAIL PROTECTED] (zalzon) wrote: > Read an article down at The Inquirer which mentions that Prescott will > have "64 bit extentions". It seems fairly unlikely to me that this is correct, in the sense of a 64-bit address space and corresponding instruction set, like an AMD Opteron. It could be that there's a drastic overstatement of some different feature. For example, upgrading PAE to allow the chip to handle really large amounts of physical memory, but still with a limit of 4Gb of address space per process. However, AFAIK, the story is based on interpretation of die photographs, and while such a picture - with a scale - can tell you a bit about pinouts and cache sizes, I strongly suspect that this story is based on over-interpretation. > I know its a 32 bit chip, but does this mean it could work on Win64 by > emulating a 64 bit processor? Well, yes, in an absolute sense. But probably not for practical purposes. I've used an Itanium emulator on 32-bit x86. The net 64-bit performance was about 1% of the raw performance of the 32-bit processor. An emulator could be a whole lot better than that and still useless for doing real work. If Prescott had a real 64-bit mode, Intel might want to keep it quiet as long as possible, but since the chip's due for introduction in a couple of months and third parties are working on motherboards, one would feel it would have leaked out by now, somehow. There'd be more information about it than the Inquirer story. They might be trying to keep it quiet until operating systems and software were ready, but one Inquirer story that I believe is that Microsoft would be extremely reluctant to support two incompatible 64-bit x86 instruction sets. And Intel probably have too much pride to sell something that's AMD- compatible. > We are straddling the 32 to 64 bit transition time frame here. Is it > worth waiting for 64 bit chips to establish themselves on the market > before upgrading? It depends, entirely, what you're using your computer for. Some hint about that would make it more possible to give sensible answers. > Now if you listen to Intel, they say they are not sure whether 64 bit > is ready for the desktop yet. That may be because they don't have a > 64 bit desktop chip yet and are trying to play down the whole affair. Well, since they don't have a 64-bit desktop-targeted chip ready with all the software to take over from the 32-bit chips they make a living out of, they'd be foolish to talk up the idea. Please remember that statements like that from any company are largely marketing-driven. There are applications in MCAD and ECAD, for example, where 64-bit desktops are genuinely useful now, and have been for a few years. They just aren't using Windows, or Intel processors, very much. This is starting to change, but fairly slowly, because people who are used to 64-bit UNIX desktops often resist being switched to Windows, for a variety of reasons. > But if that is the case, why are they introducing 32 bit chips like > centrino and prescott? If they don't have a 64-bit chip ready, they still need to keep on introducing new chips, or their marketing department will become under-employed. It's a whole lot quicker and simpler to modify and upgrade an existing design to something better, but not drastically different, than it is to do a complete redesign. > I don't know what to do now. I need to upgrade but I don't want to > upgrade to something which lasts only a short while before yet ANOTHER > upgrade is needed. Well, an Opteron or Athlon 64 will run all the existing x86 32-bit software, and its own 64-bit code. That seems to be the most versatile PC processor available at present. Will it succeed? Well, there are much bigger bets than your machine upgrade budget on it. Both ways. --- John Dallman [EMAIL PROTECTED] "Any sufficiently advanced technology is indistinguishable from a well-rigged demo"
| <-- __Chronological__ --> | <-- __Thread__ --> |