Admiral Ackbar
CAGiversary!
- Feedback
- 25 (100%)
Xbox Next "Xenon" Processor Specs.
- Thread starter Admiral Ackbar
- Start date
captainfrizo
CAGiversary!
- Feedback
- 6 (100%)
I don't understand most of those specs, but those I do understand deserve one word: "wow."
magilacudy
CAGiversary!
- Feedback
- 16 (100%)
Eh, within the next 2-3 years when they launch Xbox Next that will be considered a midrange CPU. I'm interested in seeing what the video card will be capable of.
I hope the specs continue to be insane though, that's what Xbox is all about.
I hope the specs continue to be insane though, that's what Xbox is all about.
Those are simply not the numbers for any machine shipping by 4Q 2005 for $300 or less. To put it another way, no game console is going to have CPU specs that exceed anything IBM will offer for systems costing orders of magnitude more in the same time frame. Another MacWorld has just gone by without any new performance grades in the Mac G5 family being announced. If IBM cannot dependably deliver processors faster than 2.5 GHz for high-end desktop systems in this year what would make any sane person think the greatest combined clock speed increase and cost decrese in the history of the microprocessor industry is going to spring up in time for Microsoft's consumer level product?
janglypangly
CAGiversary!
- Feedback
- 24 (100%)
Well if the XBOX NEXT does have those spec's, it'll cost around 1,000 :X!!
Pylis
CAGiversary!
Microsoft is actually starting to turn a profit on the XBox, not because of the hardware (they're losing money on that) but because of the games. I really think that the XBox should be absolutely kicking Sony's ass, but they unfortunately don't have the diverse third-party support that the PS2 enjoys. I wouldn't be surprised if Microsoft shoots for a low system cost right out the door. It'll cost a fortune early on, but if they can expand their core audience, third-party support will grow on its own, and in the long run will actually be profitable.
King Bahamut
CAGiversary!
How are they starting to turn a profit?? Last I saw they lost over a billion dollars and still losing.
Pylis
CAGiversary!
I guess I need to clarify. The XBox is still ultimately in debt by like 700 million dollars or something (not a billion) but according to recent figures, Microsoft is slowly starting to make money, not go deeper in debt.
Edit: The amount of money being made is minimal, and it's hard to say how long this will last. Currently Microsoft is more concerned about selling consoles and expanding its user base than it is with making money.
Edit: The amount of money being made is minimal, and it's hard to say how long this will last. Currently Microsoft is more concerned about selling consoles and expanding its user base than it is with making money.
Thunderscope
CAGiversary!
a Dual 1.8 ghz G5 (3.6ghz) goes for $2000 alone so even if Microsoft used this processor in a stripped down XboxNextit would cost $1200 now and around $800 to make next year if they are lucky.
[quote name='epobirs']If IBM cannot dependably deliver processors faster than 2.5 GHz for high-end desktop systems in this year what would make any sane person think the greatest combined clock speed increase and cost decrese in the history of the microprocessor industry is going to spring up in time for Microsoft's consumer level product?[/quote]
Economies of scale. They're going to mass produce a lot more CPU's for Xbox than they will for Apple by an order of magnitude, and all identical.
I will admit those specs seem a bit high, but we're talking 18 months from now. Would today's processors seem a bit high in December 02? G4's then were at most 1.25ghz. I wouldn't be surprised if it ends up lower than that though.
Economies of scale. They're going to mass produce a lot more CPU's for Xbox than they will for Apple by an order of magnitude, and all identical.
I will admit those specs seem a bit high, but we're talking 18 months from now. Would today's processors seem a bit high in December 02? G4's then were at most 1.25ghz. I wouldn't be surprised if it ends up lower than that though.
[quote name='dafoomie'][quote name='epobirs']If IBM cannot dependably deliver processors faster than 2.5 GHz for high-end desktop systems in this year what would make any sane person think the greatest combined clock speed increase and cost decrese in the history of the microprocessor industry is going to spring up in time for Microsoft's consumer level product?[/quote]
Economies of scale. They're going to mass produce a lot more CPU's for Xbox than they will for Apple by an order of magnitude, and all identical.
I will admit those specs seem a bit high, but we're talking 18 months from now. Would today's processors seem a bit high in December 02? G4's then were at most 1.25ghz. I wouldn't be surprised if it ends up lower than that though.[/quote]
No, we aren't talking 18 months from now. If this machine is to be in stores in time for the Xmas season the mass production must start several months earlier. Developers will need final hardware earlier still.
Clockspeeds from the most speed obseesed chip company, Intel, have stopped advancing. They're struggling to get 3.6 GHz chips into the extreme high end markets. It wasn't that long ago that their roadmaps called for 4 GHz by the first quarter of this year. IBM is vastly less driven to push clock speed. The 2.5 GHz Mac G5 systems are in short supply because IBM doesn't actually make a 2.5 GHz chip. They produce 2.0 GHz chips and some of them test as exceptionally stable and are sold running at 2.5 GHz. This suits Apple's purposes but IBM cannot do this dependably enough that they would actually try to sell PPC970 chips at that speed. 3.5 GHz in the volumes Microsoft would need are not going to appear in 2005. If IBM felt they could do so reliably you'd see it reflected in their own very high margin POWER5 systems as well as Apple's products.
As for price, it gets very tiresome when people parrot the phrase 'economies of scale' as if it solves everything. It only applies under specific circumstances such as increased purchases of raw materials. IBM is already a massive foundry operation. Microsoft's business is not going to get them a better price on wafer. The other helpful circumstance is increased manufacturing efficiencies gained from things learned over a longer period of manufacturing than a shorter period. Again, this is quite unlikely as IBM is already as good as they're going to get with current process technology. The price of G5 chips will decrease but not because of Microsoft's business. It will be the same for every G5 derivative customer and most immediately visible in Apple's products since they'll have several product refreshes before Microsoft makes any definitive Xbox 2 announcement.
Economies of scale. They're going to mass produce a lot more CPU's for Xbox than they will for Apple by an order of magnitude, and all identical.
I will admit those specs seem a bit high, but we're talking 18 months from now. Would today's processors seem a bit high in December 02? G4's then were at most 1.25ghz. I wouldn't be surprised if it ends up lower than that though.[/quote]
No, we aren't talking 18 months from now. If this machine is to be in stores in time for the Xmas season the mass production must start several months earlier. Developers will need final hardware earlier still.
Clockspeeds from the most speed obseesed chip company, Intel, have stopped advancing. They're struggling to get 3.6 GHz chips into the extreme high end markets. It wasn't that long ago that their roadmaps called for 4 GHz by the first quarter of this year. IBM is vastly less driven to push clock speed. The 2.5 GHz Mac G5 systems are in short supply because IBM doesn't actually make a 2.5 GHz chip. They produce 2.0 GHz chips and some of them test as exceptionally stable and are sold running at 2.5 GHz. This suits Apple's purposes but IBM cannot do this dependably enough that they would actually try to sell PPC970 chips at that speed. 3.5 GHz in the volumes Microsoft would need are not going to appear in 2005. If IBM felt they could do so reliably you'd see it reflected in their own very high margin POWER5 systems as well as Apple's products.
As for price, it gets very tiresome when people parrot the phrase 'economies of scale' as if it solves everything. It only applies under specific circumstances such as increased purchases of raw materials. IBM is already a massive foundry operation. Microsoft's business is not going to get them a better price on wafer. The other helpful circumstance is increased manufacturing efficiencies gained from things learned over a longer period of manufacturing than a shorter period. Again, this is quite unlikely as IBM is already as good as they're going to get with current process technology. The price of G5 chips will decrease but not because of Microsoft's business. It will be the same for every G5 derivative customer and most immediately visible in Apple's products since they'll have several product refreshes before Microsoft makes any definitive Xbox 2 announcement.
[quote name='dafoomie']
I will admit those specs seem a bit high, but we're talking 18 months from now. Would today's processors seem a bit high in December 02? G4's then were at most 1.25ghz. I wouldn't be surprised if it ends up lower than that though.[/quote]
Not a valid comparison. The supplier for the G4 chips was Motorola, a company that had only stayed in the business of making CPUs for the personal computer market at Apple's behest and hadn't made a profit on Apple's business in several years. They were understandably in no hurry to produce faster G4s that were of no interest to their primary markets in embedded systems. As well, there several technical impediments to making the G4 faster. Apple had no choice but to stay with Motorola because IBM's PPC970, which would have much better growth potential, was not going to be ready for quite a long while.
If you examine benchmarks for the slowest G5 against the fastest G4 it becomes clear that clock speed is only a small part of the story. The G5 is simply a much stronger implementation of the PPC ISA.
This is where the blatant fantasizing in the claimed Xbox 2 specs come in. The writing it wanted to create something that would appear devastatingly powerful yet almost believable so he took the speed of the fastest currently shipping CPU, 3.4 GHz, and added a bonus 100 MHz. This pattern in fantasy console specs has been going on for years.
I will admit those specs seem a bit high, but we're talking 18 months from now. Would today's processors seem a bit high in December 02? G4's then were at most 1.25ghz. I wouldn't be surprised if it ends up lower than that though.[/quote]
Not a valid comparison. The supplier for the G4 chips was Motorola, a company that had only stayed in the business of making CPUs for the personal computer market at Apple's behest and hadn't made a profit on Apple's business in several years. They were understandably in no hurry to produce faster G4s that were of no interest to their primary markets in embedded systems. As well, there several technical impediments to making the G4 faster. Apple had no choice but to stay with Motorola because IBM's PPC970, which would have much better growth potential, was not going to be ready for quite a long while.
If you examine benchmarks for the slowest G5 against the fastest G4 it becomes clear that clock speed is only a small part of the story. The G5 is simply a much stronger implementation of the PPC ISA.
This is where the blatant fantasizing in the claimed Xbox 2 specs come in. The writing it wanted to create something that would appear devastatingly powerful yet almost believable so he took the speed of the fastest currently shipping CPU, 3.4 GHz, and added a bonus 100 MHz. This pattern in fantasy console specs has been going on for years.
(Let me preface this by saying I thought the "leak" was bs too.)
Xbox was released on November 8 01. Xbox production started on October 10 01. The chips were probably produced much earlier, but they didn't need them in hand until 1 month before release. But to be fair, I'll say 12-14 months until they need to start mass producing them. Developer models and prototypes are not mass production.
The Xbox CPU has been rumored to be (by people and places much more credible than some made up whitepaper) a variant of the 970FX. The 970FX went from 0.13 micron to 90nm, and uses silicon on insulator technology. It has two cores, where the whitepaper says 3, so thats immediate bs. But, having this run at 2.5-3ghz is not farfetched at all. They aren't producing any higher than 2.5ghz because of manufacturing difficulties related to the new 90nm process, but I think a year is more than enough time to get that worked out.
It doesn't solve everything. But by making and buying these on as large a scale as they are, it would make it cheap enough to be possible. IBM has much room for improvement in manufacturing
Again, they are having some problems with the new 90nm process. They're getting much lower yields than expected (more processors on each wafer are defective).
I agree that the "whitepaper" is nothing but fantasy. But I think you'll see 3-3.5GHZ in high end desktops by the time Xbox 2 launches, but not in a console.
Xbox was released on November 8 01. Xbox production started on October 10 01. The chips were probably produced much earlier, but they didn't need them in hand until 1 month before release. But to be fair, I'll say 12-14 months until they need to start mass producing them. Developer models and prototypes are not mass production.
The Xbox CPU has been rumored to be (by people and places much more credible than some made up whitepaper) a variant of the 970FX. The 970FX went from 0.13 micron to 90nm, and uses silicon on insulator technology. It has two cores, where the whitepaper says 3, so thats immediate bs. But, having this run at 2.5-3ghz is not farfetched at all. They aren't producing any higher than 2.5ghz because of manufacturing difficulties related to the new 90nm process, but I think a year is more than enough time to get that worked out.
It doesn't solve everything. But by making and buying these on as large a scale as they are, it would make it cheap enough to be possible. IBM has much room for improvement in manufacturing
Again, they are having some problems with the new 90nm process. They're getting much lower yields than expected (more processors on each wafer are defective).
The price of the chips for Microsoft will decrease, because they're Microsoft, and because of the quantity.
I agree that the "whitepaper" is nothing but fantasy. But I think you'll see 3-3.5GHZ in high end desktops by the time Xbox 2 launches, but not in a console.
Pylis
CAGiversary!
I agree. While it's possible that these chips could be in the Xenon, at the rate that chips are currently improving it seems very unlikely.
Something that always seems to bother me though... look at the ultra-cool shit that the X-Box can currently accomplish. Sure, it might not run Doom III well, but Doom III is a lousy, unoptimized game engine, or at least it sure looks that way to me. I simply don't think it's feasable to have a 3.5 ghz processor because, outside of maybe launching resolutions through the roof, I don't see what the hell it's going to be used for based on the current technology being used in games. A system twice as fast as the current system would probably suffice, but that's my opinion.
Something that always seems to bother me though... look at the ultra-cool shit that the X-Box can currently accomplish. Sure, it might not run Doom III well, but Doom III is a lousy, unoptimized game engine, or at least it sure looks that way to me. I simply don't think it's feasable to have a 3.5 ghz processor because, outside of maybe launching resolutions through the roof, I don't see what the hell it's going to be used for based on the current technology being used in games. A system twice as fast as the current system would probably suffice, but that's my opinion.
I can buy a 3.4 GHz system today and 3.6 GHz in a few weeks. It will not be a PowerPC system, though. What part of 'IBM doesn't care about pushing clockspeeds and nobody can force them to care' do you not understand? The desktop CPU business is not a driving force in their operations. It takes several hundred Apple G5 system sales to match the profit on a single POWER5 system sale. Likewise, they aren't all that excited about embeeded market. In fact they recently sold off their embedded PPC (PowerQICC) operation. They'll still perform as the foundry for those products but they otherwise don't find it worth their time. They simply aren't going to follow Intel down the rabbit hole of putting a zillion stages in the CPU pipeline and thus clock speed will increase at a placid pace. This is not a formula that should prompt expectation of majorly faster and cheaper PPC970 derivatives within the next year.
Yields always suck on new process technology. Sony's premature commitment to .18 for the PS2 chips was a disaster that added tens of millions of dollars in cost to early PS2 production. For a low- or even negative-margin product like a console it's foolish to make a planned technology that doesn't yet exist a critical part of your product, as happened when Sony designed the PS2 chipset with proof of concept production at .25 micron. This lead to a high defect rate before the .18 production was worked out just in time for US shipments. ,90 nm will be old news by the time any next gen consoles start production. Any efficiencie to be gain will long since have been achieved with higher margin products that more readily absorb low yields.
Microsoft does not have any magic control over IBM. They will not sell them chips for a loss as Motorola long did for Apple. This is why IBM is a titan while Moto is a train wreck. none of IBM's production capacity is sitting idle. They don't have any need to make unfavorable deals. IBM and Microsoft got a divorce about thirteen years ago. When the ex shows up at the office she's just another customer, especially when two other bimbos are also looking for favors.
That kind of clock speed is simply not needed to make a PPC based console a good performer. The high clock rates of Intel chips are tied to their extremely deep pipelines. There are other ways to exact more performance from a given ISA as AMD has demonstrated and benchmark comparisons betwen a 2.0 GHz G5 and a 3.0 GHz P4 reinforce. The next Xbox will not be in competition with top of the line PCs and will not need to match their numbers any more than its predecessor did. By the time the first Xbox shipped P-III/733 MHz systems were too low-end to be found at most retailers any longer. The XGPU had a slight lead over the then PC champ GeForce 3 but this period was measure in mere months before the GeForce 4 came out. Basing console design on PC industry numbers is a trap for he unwary. The markets driving them are so different as too make any crossreference worthless.
Yields always suck on new process technology. Sony's premature commitment to .18 for the PS2 chips was a disaster that added tens of millions of dollars in cost to early PS2 production. For a low- or even negative-margin product like a console it's foolish to make a planned technology that doesn't yet exist a critical part of your product, as happened when Sony designed the PS2 chipset with proof of concept production at .25 micron. This lead to a high defect rate before the .18 production was worked out just in time for US shipments. ,90 nm will be old news by the time any next gen consoles start production. Any efficiencie to be gain will long since have been achieved with higher margin products that more readily absorb low yields.
Microsoft does not have any magic control over IBM. They will not sell them chips for a loss as Motorola long did for Apple. This is why IBM is a titan while Moto is a train wreck. none of IBM's production capacity is sitting idle. They don't have any need to make unfavorable deals. IBM and Microsoft got a divorce about thirteen years ago. When the ex shows up at the office she's just another customer, especially when two other bimbos are also looking for favors.
That kind of clock speed is simply not needed to make a PPC based console a good performer. The high clock rates of Intel chips are tied to their extremely deep pipelines. There are other ways to exact more performance from a given ISA as AMD has demonstrated and benchmark comparisons betwen a 2.0 GHz G5 and a 3.0 GHz P4 reinforce. The next Xbox will not be in competition with top of the line PCs and will not need to match their numbers any more than its predecessor did. By the time the first Xbox shipped P-III/733 MHz systems were too low-end to be found at most retailers any longer. The XGPU had a slight lead over the then PC champ GeForce 3 but this period was measure in mere months before the GeForce 4 came out. Basing console design on PC industry numbers is a trap for he unwary. The markets driving them are so different as too make any crossreference worthless.
What part of they will be at 3+GHZ by next year because they planned to be this summer do you not understand? What part of their improvements in the 970FX and upcoming lines that are geared towards getting better clock speeds do you not understand? Yes, they aren't stupidly obsessed about clock speed like Intel is. But it doesn't mean they won't go from 2.5ghz to 3ghz in over a year. Will Xbox have a 3ghz processor? No.
True to a degree. They sold off their embedded PPC 4xx line, they don't care as much now about the general embedded market. But they're focusing in on one major part of that market. Video game consoles. They're going to make CPUs for all 3 major consoles.
No one says they are. 3ghz is not a major increase, again, they planned that for this summer (Jobs let that cat out of the bag) but they didn't.
The chips that Microsoft buys will be cheaper for them, not because Microsoft and IBM had a relationship in the past, but because Microsoft has lots and lots of money, and will be buying lots and lots of chips. This is why IBM likes the Video Game Console market, they will be making lots and lots of the same chip for a sustained period of time. When you do that in comparison to another embedded device that will be produced on a smaller scale, of course its going to be cheaper for Microsoft, or Sony, or Nintendo.
It is not unreasonable to say that IBM will have a 3ghz 970FX by next summer. It is not a huge increase and they were planning on having that already. Will Xbox 2 have it? No, of course not. But 2ghz would not exactly be out of this world either.
[quote name='Pylis']I agree. While it's possible that these chips could be in the Xenon, at the rate that chips are currently improving it seems very unlikely.
Something that always seems to bother me though... look at the ultra-cool shit that the X-Box can currently accomplish. Sure, it might not run Doom III well, but Doom III is a lousy, unoptimized game engine, or at least it sure looks that way to me. I simply don't think it's feasable to have a 3.5 ghz processor because, outside of maybe launching resolutions through the roof, I don't see what the hell it's going to be used for based on the current technology being used in games. A system twice as fast as the current system would probably suffice, but that's my opinion.[/quote]
There is absolutely no end to the things developers would given the horsepower and the tools to let them get things done quickly enough to keep the budget within reason. For example, compare the real-time animation quality of the PS2 FF X and X-2 to the PS1 FF entries. The difference goes way beyond polygon counts. The figures don't just look better, they also have much better articulation, and their clothing is separate objects with their own behavior properties. Even so, you need only watch the FMV sequences to see just how much better this can become. We still have a ways to go before real-time animation matches that of CGI movies produced in the 90's.
That is just one aspect. There is also the matter of keeping a lot of NPCs running around with their own agendas. Consider the city in Jak II. It would have been hard to do on earlier hardware yet is still mostly made up of wandering zombie clones of less than a dozen people. Important characters immediately stand out from the crowd just by virtue of not coming from the same stock of clones. Having so many distinctive characters running around creates demand for greater memory and processing power to keep up their behavorial routines while keeping the AV interesting and glitch free.
The list is endless.
Something that always seems to bother me though... look at the ultra-cool shit that the X-Box can currently accomplish. Sure, it might not run Doom III well, but Doom III is a lousy, unoptimized game engine, or at least it sure looks that way to me. I simply don't think it's feasable to have a 3.5 ghz processor because, outside of maybe launching resolutions through the roof, I don't see what the hell it's going to be used for based on the current technology being used in games. A system twice as fast as the current system would probably suffice, but that's my opinion.[/quote]
There is absolutely no end to the things developers would given the horsepower and the tools to let them get things done quickly enough to keep the budget within reason. For example, compare the real-time animation quality of the PS2 FF X and X-2 to the PS1 FF entries. The difference goes way beyond polygon counts. The figures don't just look better, they also have much better articulation, and their clothing is separate objects with their own behavior properties. Even so, you need only watch the FMV sequences to see just how much better this can become. We still have a ways to go before real-time animation matches that of CGI movies produced in the 90's.
That is just one aspect. There is also the matter of keeping a lot of NPCs running around with their own agendas. Consider the city in Jak II. It would have been hard to do on earlier hardware yet is still mostly made up of wandering zombie clones of less than a dozen people. Important characters immediately stand out from the crowd just by virtue of not coming from the same stock of clones. Having so many distinctive characters running around creates demand for greater memory and processing power to keep up their behavorial routines while keeping the AV interesting and glitch free.
The list is endless.
[quote name='Thunderscope']a Dual 1.8 ghz G5 (3.6ghz) goes for $2000 alone so even if Microsoft used this processor in a stripped down XboxNextit would cost $1200 now and around $800 to make next year if they are lucky.[/quote]
You do realize that two processors of speed X does not equal one processor of speed 2X? Multiple processors are useful for case where there are several diverse tasks to perform or one task that is easily divided up into sub tasks. In case where that cannot be done multiple processors do not cumulatively equal a much faster processor of the same architecture.
You do realize that two processors of speed X does not equal one processor of speed 2X? Multiple processors are useful for case where there are several diverse tasks to perform or one task that is easily divided up into sub tasks. In case where that cannot be done multiple processors do not cumulatively equal a much faster processor of the same architecture.
wesley982002
CAG Veteran
A xenon 3.2ghz runs around 700 right now, so in 18 months that will be down to probably about a hundred and some change. If you do a look at the intel roadmap you will see that the 4 ghz xenon is supposed to be out q1 of 2005, so that gives plenty of time for price cuts. It could still be close to 200 though, so theres no telling on a 300 dollar price or that processor being used.
wesley982002
CAG Veteran
Oh my bad, I thought it was saying a xenon processor
[quote name='wesley982002']A xenon 3.2ghz runs around 700 right now, so in 18 months that will be down to probably about a hundred and some change. If you do a look at the intel roadmap you will see that the 4 ghz xenon is supposed to be out q1 of 2005, so that gives plenty of time for price cuts. It could still be close to 200 though, so theres no telling on a 300 dollar price or that processor being used.[/quote]
I take it you mean a Xeon. This is Intel's line for high-end workstations and servers, primarily distinguished by large caches and SMP support. (SMP used to be universally supported but Intel deided to make it more of a premium with the P4. The system I'm writing this on is a dual Celeron 533 MHz BP6 mobo.)
The 3.2 GHz Xeon will never, ever be down to a hundred dollars. (Which would be extremely expensive for a console in any case.) It will instead simply by discontinued as have all its predecessor which no longer had a market against consumer oriented chips.
I wouldn't hold my breath on that 4 GHz P4. Intell recently canceled the next two generations of the P4 (NetBurst) architecture to to severe problems with power and heat. The recently introduced Prescott P4 chips will continue but they would never be suitable for console applications in their current form. If a new Intel based console were starting design today it would instead look to the Pentium M series that currnetly drive the laptop market and will be the basis of future desktop CPUs. These have a desing much closer to the P-III and AMD chips which among other hings means much lower clocks speeds. The Pentium M line just recently hit 2 GHZ.
As I mentioned previously, P4 speeds have absolutely no bearing on expected clock speed increase to the IBM PPC products. The Pentium 4 architecture was specifically designed for high clock rates and only outpaces competing architectures where it has a substantial clock speed advantage. This is why a CPU that AMD gives a performance rating equal to a 3.0 GHz P4 runs at a much lower clock rate but is head to head in most performance measures. For example, note the clock speed of this CPU: http://shop4.outpost.com/product/3559924
AMD recently moves to a 64-bit design with integrated memory controller that greatly up their performance without any major boost to clock speed. See here how their ass-kicking 3400+ runs at only 2.2 GHz.
PowerPC similarly is not designed to emphasize clock rate. IBM is in no hurry to reach the speeds that are currently casting Intel's roadmap into disarray.
I take it you mean a Xeon. This is Intel's line for high-end workstations and servers, primarily distinguished by large caches and SMP support. (SMP used to be universally supported but Intel deided to make it more of a premium with the P4. The system I'm writing this on is a dual Celeron 533 MHz BP6 mobo.)
The 3.2 GHz Xeon will never, ever be down to a hundred dollars. (Which would be extremely expensive for a console in any case.) It will instead simply by discontinued as have all its predecessor which no longer had a market against consumer oriented chips.
I wouldn't hold my breath on that 4 GHz P4. Intell recently canceled the next two generations of the P4 (NetBurst) architecture to to severe problems with power and heat. The recently introduced Prescott P4 chips will continue but they would never be suitable for console applications in their current form. If a new Intel based console were starting design today it would instead look to the Pentium M series that currnetly drive the laptop market and will be the basis of future desktop CPUs. These have a desing much closer to the P-III and AMD chips which among other hings means much lower clocks speeds. The Pentium M line just recently hit 2 GHZ.
As I mentioned previously, P4 speeds have absolutely no bearing on expected clock speed increase to the IBM PPC products. The Pentium 4 architecture was specifically designed for high clock rates and only outpaces competing architectures where it has a substantial clock speed advantage. This is why a CPU that AMD gives a performance rating equal to a 3.0 GHz P4 runs at a much lower clock rate but is head to head in most performance measures. For example, note the clock speed of this CPU: http://shop4.outpost.com/product/3559924
AMD recently moves to a 64-bit design with integrated memory controller that greatly up their performance without any major boost to clock speed. See here how their ass-kicking 3400+ runs at only 2.2 GHz.
PowerPC similarly is not designed to emphasize clock rate. IBM is in no hurry to reach the speeds that are currently casting Intel's roadmap into disarray.
bread's done