It's not the Cell in the X360. It uses a three-core CPU, the Xenon, designed with Microsoft's needs in mind; the Cell is a single core with multiple dedicated vector units, and which isn't explicitly designed for the PS3 (there will be lower-powered variants in TVs, for example).
This discussion is a bit of water under the bridge. Apple isn't going back to PPC short of an obvious revolution in those chips, especially not ones like the Cell or Xenon; Apple would need to reoptimize for them the same way it did with the G5.
Populist is right. What I can't figure out is why so many people on these forums don't get it. Forget the graphics questions and the cost issues. Think about what you need out of a device in the livingroom.
1. Watch live TV (w/ tuner and cable input)
2. Record TV shows for time shift (TiVo)
3. Download video from net (tv shows, movies, trailers, podcasts, etc.)
4. Watch movies (DVD, HD-DVD, Blue Ray)
5. Watch home videos
6. Play video games (like Xbox 360, PS3, etc.)
7. Listen to radio (tuner)
6. Play CDs, iTunes library, Pocasts, internet radio
7. Watch iPhoto slideshows
9. Sync content with iPod or similar portable device
10. Video and/or audio conference, VoIP, IM
11. Access email and internet
12. HiFi audio and AV switcher functionality
Then maximize the iPod screen size and turn the iPod into this.
1. Play music
2. Radio
3. Play video
4. TV receiver
5. Camera (recorder photos and video clips)
6. Play photo slideshows, show photos
7. Phone - how about video phone?
8. text message or IM
9. PDA - Address book, notes, iCal, etc.
10. Web browser
11. GPS, compass, navigation system
12. Audio recorder
13. Data storage (data, video right from a video camera, D-SLR, etc.)
Maybe all these aren't ready to go yet. But they are probably coming.
Apple's mini Mac should be a Mac for the office worker who only runs Office, a browser, email, et. A low cost headless machine.
Apple needs a consumer machine and the iMac fills this bill. The top of the line Mac would still be the media authoring and work platform for everything else.
People are also going to need data storage and the Xserve RAID comes into play here. Same with the Xserve. Look for houses in the future to have all the rooms networked together for sharing data. Or even better is fast wireless networking. Apple is a player here too.
The bottom line is that Apple is close to having all these components. They even have big good LCD displays that could be easily modified for TV. Their new iPod HiFi could be their first move into the livingroom audio market. They could easily produce a 5.1 system.
All they need then is a bunch of stores to market these products - wait, they already have this. This is my vision of what I need and what I would like to see. The question is what is Apple's vision. It seems they are pretty close.
Apple can be the new Sony but games are Apple's biggest weakness.
As for games, if Sony or MS takes a loss on the hardware, why doesn't Apple just have a software translation program to run Sony's or Microsoft's games on the Mac? Instead of Sony and Microsoft losing money, they will make money. Yeah the livingroom Mac would have to be powerful and more expensive, but I and other people want this functionality now and will pay for it.
Apple also needs to keep advancing on the software front. They are doing it but slower than what I would like to see.
Apple needs to blow iWork into a valid office suite with the ability to read and write MS Office docs. I don't know why they haven't released a spreadsheet app in iWork yet. Frustrating. Oh, then give it away free. Bye-bye MS cash cow. In five years when MS pulls Office off the Mac, who will care?
Then they need to do a good accounting package and give the source code away for low cost so programmers can modify it for various business needs. They also need to develop their Mail, iCal, and Address book into a full blown CRM system. Sell it as the pro version. This will help Apple make inroads into the business market.
Also, Apple needs their own Photoshop, Dreamweaver, Flash, etc. Even better, buy out Adobe. Then MS Windows would need Apple/Adobe more then Apple needs MS Office. iPhoto, Aperature, Keynote, iWeb are all good starts for a continued high end pro/low end consumer apps approach.
I know not all of this is ready to go today, but this is the vision Apple needs, and maybe already has. Microsoft and Apple are the two companies that can own the future electronic hub market place. Sony is history.
Apple is developing strongly in most of these catagories but games is their weakness. So is TiVo capabilities. Front Row is not a complete solution yet. Nor is the mini the right platform for the livingroom. Populist is right. Games are a key missing feature. But I don't know if the mini is the platform for what he wants. Close though.
Or...
What about an Apple/Sony partnership facing off against Microsoft? While the two face off on the music front, Sony should conceed this market and Apple should conceed the games market to Sony. Together they can take on MS and win.
Apple could do software and an OS for the PS3 and then have an emulator for PS3 games on the Mac. At least PS2 games. There use to be an PS on the Mac emulator. I imagine the move to intel chips would mess things up about this. But I think a partnership could work if it could be done.
If that is the case then what does the Mini need to be a graphics rocket?
A $300+ dedicated GPU.
Quote:
And I still dont understand the shift from Power PC to Intel processors.
That's very simple. Intel is the one company with the highest odds of advancing CMOS manufacturing to 45 nm, 32 nm, and 20 nm in the quickest time.
Quote:
And why are game consoles now switching to Power PC processors if Intel chips are so awesome?
That's an interesting question. I'm sure there are multiple reasons: media-specialized hardware, ISA and architecture lock-in, protection of other products.
"And why are game consoles now switching to Power PC processors if Intel chips are so awesome?"
The RISC/PPC environment is better suited for 3D gaming, and it's cheaper than building from the ground up what PPCs can already do well enough. The PPC family is very broad, as well. From simplistic designs made for set-top boxes all the way up to big iron. Even Intel's cheapest chips are excessively complex for the intended purpose (just games).
The short answer is that Intel doesn't make chips that are simple and economical enough for a console. Maybe they will someday, but today they don't. The only reason MS used them in the first place was because they threw something together in 18 months. Consoles normally take years, literally, to iron out.
As someone else mentioned, the CPUs in the console are craptacular for desktop tasks. They're PPC-derived but VERY watered down. The Xenon cores executes only two threads at a time, for example. A G5 would blow it away in regular computing tasks, but G5's are also too expensive for consoles.
So the marriage doesn't work without costs getting out of reach of what console gamers are willing to pay. Convergence has been tried many times in the last decade and always failed. People don't want to "compute" from the sofa.
Fair enough people, your replies have answered most of my questions. I especially appreciate the technical specifications that explain why some things work and others dont. I would like to reiterate that I do not have a technical/computer background.
But I still think the Mac Mini could do better, and at a reasonable cost. My Xbox has a Pentium 3, 64 megs of RAM, and 6Ghz memory bandwidth. Surely the next generation of computers can outdo that, so why cant the Mini?
I still think it is possible to have a computer that plays CDs DVDs, can handle High Def, surf, do basic home computing tasks, and have decent graphics. What about a graphics chip that is less expensive than today's top-of-the-line chips, yet still powerful. Or better yet, how about a programmable GPU, that could handle high def, and be programmed to handle what's next?
One more question:
THT said
Quote:
Intel is the one company with the highest odds of advancing CMOS manufacturing to 45 nm, 32 nm, and 20 nm in the quickest time.
If Intel chips are advancing (shrinking) faster than the competition, and Apple has no choice but to switch to Intel processors, what will happen to the Power PC mac towers?
I think the problem is a somewhat idealogical "Who's your audience?" question. The needs/desires of gamers (high-end GPU, HD games, etc) aren't entirely compatible with a basic computing needs.
A "balance" of a modest CPU and GPU is possible but you alienate one group or another (or a little from all) and you have a wonky scenario where the product doesn't quite have a place. With so much fierce competition from both ends, a middle-road solution would have it tough.
The TV does well for entertainment needs. The stuff like Email/work/surfing never caught on. I wouldn't want to do any of that on a 640x480 TV resolution anyway, personally.
"how about a programmable GPU, that could handle high def, and be programmed to handle what's next?"
GPUs are pretty hardware limited. They can be tweaked to improve performance here and there through drivers/software, but you can't get subtantial improvement. It'd be the holy grail if it were possible, though. I'm sure ATI and nVidia would rather do that than release new GPUs ever 6 months.
"what will happen to the Power PC mac towers?"
Apple's entire lineup is switching to Intel this year.
Apple, however, makes a profit - albeit not a large one.
The teardown of the G4 Mac Mini revealed that the cost to Apple was something like $200-$220. That's a pretty healthy profit margin, even after considering retail upcharges, warehousing, etc.
Don't get me wrong. . . I'm not saying that Apple should be compelled to make the Mac Mini any cheaper unless they have a good reason to do so, but they ARE making a good profit on it. (After all, I'm not a fucking populist.)
Sure. But it isn't because of HyperTransport or the on-chip memory controller. It was primarily because the 90 nm P4 was limited to 3.8 GHz for unicores and 3.2/3.4 GHz for multicores due to power concerns. The 65 nm P4 isn't as limited and has some MHz room to go up - still runs really hot though - but Intel has chosen its direction with performance/watt and it looks like they would rather tread water or lose a little than go against the new performance/watt mantra.
HyperTransport may have more bandwidth then a Single/Dual core single processor system can use, but what's the problem with that? The way Intel has things wired is making their processors (which are pretty bad dual-cores, i may add) starved for bandwidth, which is a bad thing. And what does power concerns have to do with bandwidth?
Quote:
Originally posted by THT
If Intel wanted to, they could release 65 nm 4.0 to 4.4 GHz P4 unicores and 3.4 to 3.8 GHz P4 dual-cores. They'll run as hot or hotter than 90 nm Prescott though, but they could have, like Apple, used a liquid cooling system too.
Enter Alienware ALX.
Quote:
Originally posted by THT
Where HyperTransport is truly benificial to AMD is in the aforementationed 4 socket systems. The big iron enterprise servers. No argument that Intel's architecture doesn't compete there.
HyperTransport can have many I/Os on one chip, so each processor can talk to each other on a 1GHz Bi-Directional Bus. But that still says nothing about Intel starviong their chips for the bandwidth they desperately need. We'll see where Intel goes from here, but they cant keep upping the MHz of the FSB.
By your very own list, there is a lot more Intel tech in AMD than the other way around. The reason for that isn't very strange.
We really don't need to remember it since it has very little relevency on system performance. Certainly zero on multiprocessor performance on the Power Mac G5. All Apple uses it for is as a Northbridge to Southbridge bus, and as far as I can tell, no one has been complaining about chipset-to-chipset bandwidth in a very very long time. [/B][/QUOTE]
If Intel chips are advancing (shrinking) faster than the competition, and Apple has no choice but to switch to Intel processors, what will happen to the Power PC mac towers?
They stop being sold when the Mac/Intel towers ship. They may be concurrently sold if there isn't enough software that hasn't been converted to a universal battery yet.
HyperTransport may have more bandwidth then a Single/Dual core single processor system can use, but what's the problem with that? The way Intel has things wired is making their processors (which are pretty bad dual-cores, i may add) starved for bandwidth, which is a bad thing. And what does power concerns have to do with bandwidth?
I believe you are making an incorrect claim that bandwidth is the reason for Intel's performance loss to AMD. I'm telling you bandwidth isn't really the reason. It really is the P4 being limited in clock rate due to power concerns. If Intel sold 4 to 4.4 GHz unicores and 3.4 to 3.8 GHz dual cores in the mass market, instead of the special "factory overclocked" gaming pc market, they wouldn't be losing many benchmarks to AMD at all.
There is one place where AMD has a bandwidth advantage over Intel: it's the usage of HyperTransport on multi-socket systems, especially 4 socket systems. The advantage with HT is that it provides direct processor-to-processor connections for inter-processor commumication while with Intel systems, it has to be shared with the FSB.
In 2 socket systems, it's a small win, but Intel can counter with more on-chip cache. With the Bensley platform, they'll compete effectively by using dual independent buses and FB-DIMM memory. In 4 socket systems, AMD has a big win. Intel can counter with multi-Northbridge systems or even a Northbridge dedicated cache, but they have chosen not to. A big part of them not doing it could be inter-group politics between Itanium and x86, who knows, but Intel can certainly counter with an architecture that doesn't use HyperTransport.
For 1 socket systems, and 2 socket systems after Q1 06, the bandwidth performance is for all intents and purposes a wash. AMD will never have faster memory bandwidth than Intel will in this market. It's just the way the economics work out. I'm sure AMD will have a lead for a short while when AM2 comes out with DDR2-800 support, but Intel can simply move to 1600 MHz FSB data rates in a couple of quarters. 1333 MHz data rates compared to 1600 MHz data rates won't be a huge factor.
Quote:
HyperTransport can have many I/Os on one chip, so each processor can talk to each other on a 1GHz Bi-Directional Bus. But that still says nothing about Intel starviong their chips for the bandwidth they desperately need. We'll see where Intel goes from here, but they cant keep upping the MHz of the FSB.
Sure they can. They can go license some Rambus FlexIO tech, or independently develop their own, and get 1.6, 3.2 and 6.4 GHz data rates using "octo-pumping."
With the current quad-pumped FSB, 1.6 GHz data rate will probable happen in 2007. 2 GHz data rate may also be possible in 2H 07.
Comments
Originally posted by Kickaha
360 is using it. Three of them.
It's not the Cell in the X360. It uses a three-core CPU, the Xenon, designed with Microsoft's needs in mind; the Cell is a single core with multiple dedicated vector units, and which isn't explicitly designed for the PS3 (there will be lower-powered variants in TVs, for example).
This discussion is a bit of water under the bridge. Apple isn't going back to PPC short of an obvious revolution in those chips, especially not ones like the Cell or Xenon; Apple would need to reoptimize for them the same way it did with the G5.
1. Watch live TV (w/ tuner and cable input)
2. Record TV shows for time shift (TiVo)
3. Download video from net (tv shows, movies, trailers, podcasts, etc.)
4. Watch movies (DVD, HD-DVD, Blue Ray)
5. Watch home videos
6. Play video games (like Xbox 360, PS3, etc.)
7. Listen to radio (tuner)
6. Play CDs, iTunes library, Pocasts, internet radio
7. Watch iPhoto slideshows
9. Sync content with iPod or similar portable device
10. Video and/or audio conference, VoIP, IM
11. Access email and internet
12. HiFi audio and AV switcher functionality
Then maximize the iPod screen size and turn the iPod into this.
1. Play music
2. Radio
3. Play video
4. TV receiver
5. Camera (recorder photos and video clips)
6. Play photo slideshows, show photos
7. Phone - how about video phone?
8. text message or IM
9. PDA - Address book, notes, iCal, etc.
10. Web browser
11. GPS, compass, navigation system
12. Audio recorder
13. Data storage (data, video right from a video camera, D-SLR, etc.)
Maybe all these aren't ready to go yet. But they are probably coming.
Apple's mini Mac should be a Mac for the office worker who only runs Office, a browser, email, et. A low cost headless machine.
Apple needs a consumer machine and the iMac fills this bill. The top of the line Mac would still be the media authoring and work platform for everything else.
People are also going to need data storage and the Xserve RAID comes into play here. Same with the Xserve. Look for houses in the future to have all the rooms networked together for sharing data. Or even better is fast wireless networking. Apple is a player here too.
The bottom line is that Apple is close to having all these components. They even have big good LCD displays that could be easily modified for TV. Their new iPod HiFi could be their first move into the livingroom audio market. They could easily produce a 5.1 system.
All they need then is a bunch of stores to market these products - wait, they already have this. This is my vision of what I need and what I would like to see. The question is what is Apple's vision. It seems they are pretty close.
Apple can be the new Sony but games are Apple's biggest weakness.
As for games, if Sony or MS takes a loss on the hardware, why doesn't Apple just have a software translation program to run Sony's or Microsoft's games on the Mac? Instead of Sony and Microsoft losing money, they will make money. Yeah the livingroom Mac would have to be powerful and more expensive, but I and other people want this functionality now and will pay for it.
Apple also needs to keep advancing on the software front. They are doing it but slower than what I would like to see.
Apple needs to blow iWork into a valid office suite with the ability to read and write MS Office docs. I don't know why they haven't released a spreadsheet app in iWork yet. Frustrating. Oh, then give it away free. Bye-bye MS cash cow. In five years when MS pulls Office off the Mac, who will care?
Then they need to do a good accounting package and give the source code away for low cost so programmers can modify it for various business needs. They also need to develop their Mail, iCal, and Address book into a full blown CRM system. Sell it as the pro version. This will help Apple make inroads into the business market.
Also, Apple needs their own Photoshop, Dreamweaver, Flash, etc. Even better, buy out Adobe. Then MS Windows would need Apple/Adobe more then Apple needs MS Office. iPhoto, Aperature, Keynote, iWeb are all good starts for a continued high end pro/low end consumer apps approach.
I know not all of this is ready to go today, but this is the vision Apple needs, and maybe already has. Microsoft and Apple are the two companies that can own the future electronic hub market place. Sony is history.
Apple is developing strongly in most of these catagories but games is their weakness. So is TiVo capabilities. Front Row is not a complete solution yet. Nor is the mini the right platform for the livingroom. Populist is right. Games are a key missing feature. But I don't know if the mini is the platform for what he wants. Close though.
Or...
What about an Apple/Sony partnership facing off against Microsoft? While the two face off on the music front, Sony should conceed this market and Apple should conceed the games market to Sony. Together they can take on MS and win.
Apple could do software and an OS for the PS3 and then have an emulator for PS3 games on the Mac. At least PS2 games. There use to be an PS on the Mac emulator. I imagine the move to intel chips would mess things up about this. But I think a partnership could work if it could be done.
Originally posted by The Populist
If that is the case then what does the Mini need to be a graphics rocket?
A $300+ dedicated GPU.
And I still dont understand the shift from Power PC to Intel processors.
That's very simple. Intel is the one company with the highest odds of advancing CMOS manufacturing to 45 nm, 32 nm, and 20 nm in the quickest time.
And why are game consoles now switching to Power PC processors if Intel chips are so awesome?
That's an interesting question. I'm sure there are multiple reasons: media-specialized hardware, ISA and architecture lock-in, protection of other products.
The RISC/PPC environment is better suited for 3D gaming, and it's cheaper than building from the ground up what PPCs can already do well enough. The PPC family is very broad, as well. From simplistic designs made for set-top boxes all the way up to big iron. Even Intel's cheapest chips are excessively complex for the intended purpose (just games).
The short answer is that Intel doesn't make chips that are simple and economical enough for a console. Maybe they will someday, but today they don't. The only reason MS used them in the first place was because they threw something together in 18 months. Consoles normally take years, literally, to iron out.
As someone else mentioned, the CPUs in the console are craptacular for desktop tasks. They're PPC-derived but VERY watered down. The Xenon cores executes only two threads at a time, for example. A G5 would blow it away in regular computing tasks, but G5's are also too expensive for consoles.
So the marriage doesn't work without costs getting out of reach of what console gamers are willing to pay. Convergence has been tried many times in the last decade and always failed. People don't want to "compute" from the sofa.
But I still think the Mac Mini could do better, and at a reasonable cost. My Xbox has a Pentium 3, 64 megs of RAM, and 6Ghz memory bandwidth. Surely the next generation of computers can outdo that, so why cant the Mini?
I still think it is possible to have a computer that plays CDs DVDs, can handle High Def, surf, do basic home computing tasks, and have decent graphics. What about a graphics chip that is less expensive than today's top-of-the-line chips, yet still powerful. Or better yet, how about a programmable GPU, that could handle high def, and be programmed to handle what's next?
One more question:
THT said
Intel is the one company with the highest odds of advancing CMOS manufacturing to 45 nm, 32 nm, and 20 nm in the quickest time.
If Intel chips are advancing (shrinking) faster than the competition, and Apple has no choice but to switch to Intel processors, what will happen to the Power PC mac towers?
...
A "balance" of a modest CPU and GPU is possible but you alienate one group or another (or a little from all) and you have a wonky scenario where the product doesn't quite have a place. With so much fierce competition from both ends, a middle-road solution would have it tough.
The TV does well for entertainment needs. The stuff like Email/work/surfing never caught on. I wouldn't want to do any of that on a 640x480 TV resolution anyway, personally.
"how about a programmable GPU, that could handle high def, and be programmed to handle what's next?"
GPUs are pretty hardware limited. They can be tweaked to improve performance here and there through drivers/software, but you can't get subtantial improvement. It'd be the holy grail if it were possible, though. I'm sure ATI and nVidia would rather do that than release new GPUs ever 6 months.
"what will happen to the Power PC mac towers?"
Apple's entire lineup is switching to Intel this year.
Originally posted by Commodus
Apple, however, makes a profit - albeit not a large one.
The teardown of the G4 Mac Mini revealed that the cost to Apple was something like $200-$220. That's a pretty healthy profit margin, even after considering retail upcharges, warehousing, etc.
Don't get me wrong. . . I'm not saying that Apple should be compelled to make the Mac Mini any cheaper unless they have a good reason to do so, but they ARE making a good profit on it. (After all, I'm not a fucking populist.)
Originally posted by THT
Sure. But it isn't because of HyperTransport or the on-chip memory controller. It was primarily because the 90 nm P4 was limited to 3.8 GHz for unicores and 3.2/3.4 GHz for multicores due to power concerns. The 65 nm P4 isn't as limited and has some MHz room to go up - still runs really hot though - but Intel has chosen its direction with performance/watt and it looks like they would rather tread water or lose a little than go against the new performance/watt mantra.
HyperTransport may have more bandwidth then a Single/Dual core single processor system can use, but what's the problem with that? The way Intel has things wired is making their processors (which are pretty bad dual-cores, i may add) starved for bandwidth, which is a bad thing. And what does power concerns have to do with bandwidth?
Originally posted by THT
If Intel wanted to, they could release 65 nm 4.0 to 4.4 GHz P4 unicores and 3.4 to 3.8 GHz P4 dual-cores. They'll run as hot or hotter than 90 nm Prescott though, but they could have, like Apple, used a liquid cooling system too.
Enter Alienware ALX.
Originally posted by THT
Where HyperTransport is truly benificial to AMD is in the aforementationed 4 socket systems. The big iron enterprise servers. No argument that Intel's architecture doesn't compete there.
HyperTransport can have many I/Os on one chip, so each processor can talk to each other on a 1GHz Bi-Directional Bus. But that still says nothing about Intel starviong their chips for the bandwidth they desperately need. We'll see where Intel goes from here, but they cant keep upping the MHz of the FSB.
By your very own list, there is a lot more Intel tech in AMD than the other way around. The reason for that isn't very strange.
We really don't need to remember it since it has very little relevency on system performance. Certainly zero on multiprocessor performance on the Power Mac G5. All Apple uses it for is as a Northbridge to Southbridge bus, and as far as I can tell, no one has been complaining about chipset-to-chipset bandwidth in a very very long time. [/B][/QUOTE]
Originally posted by The Populist
If Intel chips are advancing (shrinking) faster than the competition, and Apple has no choice but to switch to Intel processors, what will happen to the Power PC mac towers?
They stop being sold when the Mac/Intel towers ship. They may be concurrently sold if there isn't enough software that hasn't been converted to a universal battery yet.
Originally posted by theapplegenius
HyperTransport may have more bandwidth then a Single/Dual core single processor system can use, but what's the problem with that? The way Intel has things wired is making their processors (which are pretty bad dual-cores, i may add) starved for bandwidth, which is a bad thing. And what does power concerns have to do with bandwidth?
I believe you are making an incorrect claim that bandwidth is the reason for Intel's performance loss to AMD. I'm telling you bandwidth isn't really the reason. It really is the P4 being limited in clock rate due to power concerns. If Intel sold 4 to 4.4 GHz unicores and 3.4 to 3.8 GHz dual cores in the mass market, instead of the special "factory overclocked" gaming pc market, they wouldn't be losing many benchmarks to AMD at all.
There is one place where AMD has a bandwidth advantage over Intel: it's the usage of HyperTransport on multi-socket systems, especially 4 socket systems. The advantage with HT is that it provides direct processor-to-processor connections for inter-processor commumication while with Intel systems, it has to be shared with the FSB.
In 2 socket systems, it's a small win, but Intel can counter with more on-chip cache. With the Bensley platform, they'll compete effectively by using dual independent buses and FB-DIMM memory. In 4 socket systems, AMD has a big win. Intel can counter with multi-Northbridge systems or even a Northbridge dedicated cache, but they have chosen not to. A big part of them not doing it could be inter-group politics between Itanium and x86, who knows, but Intel can certainly counter with an architecture that doesn't use HyperTransport.
For 1 socket systems, and 2 socket systems after Q1 06, the bandwidth performance is for all intents and purposes a wash. AMD will never have faster memory bandwidth than Intel will in this market. It's just the way the economics work out. I'm sure AMD will have a lead for a short while when AM2 comes out with DDR2-800 support, but Intel can simply move to 1600 MHz FSB data rates in a couple of quarters. 1333 MHz data rates compared to 1600 MHz data rates won't be a huge factor.
HyperTransport can have many I/Os on one chip, so each processor can talk to each other on a 1GHz Bi-Directional Bus. But that still says nothing about Intel starviong their chips for the bandwidth they desperately need. We'll see where Intel goes from here, but they cant keep upping the MHz of the FSB.
Sure they can. They can go license some Rambus FlexIO tech, or independently develop their own, and get 1.6, 3.2 and 6.4 GHz data rates using "octo-pumping."
With the current quad-pumped FSB, 1.6 GHz data rate will probable happen in 2007. 2 GHz data rate may also be possible in 2H 07.