what about a 64bit chip for tiger early next year? even for mere marketing purposes? is the e700 something to look to?
It is not known if the e600 and e700 will go into production simultaneously. From the pdf posted above in this thread, it looks like the e600 will show up first and the e700 will follow. I don't believe that we will see the e700 next year.
I think it's pretty sure that the next PB revision (Q1/ 2005) will sport a G5.
G5 iMac and G4 PB doesn't match.
Maybe the e700 (G5-M )
Power Mac G5 = IBM PPC970MP (Q1/2005), with DDR2 & PCIe *drivel*
Powerbook G5 = IBM PPC970FX or Freescale e700 (Q1/2005)
iMac G5 = IBM PPC970FX (Q4/ 2004)
iBook G4 Extreme = Freescale e600 (Q4/ 2004)
eMac G4 Extreme = Freescale e600 (Q4/ 2004)
Cube2 G4 Extreme = Freescale e600 (Q4/ 2004)
But i have the feeling that Freescale chips don't fit into Apples product line. Last but not least because of RIO.
Just my two ¢
Seeing as the 2.5 GHZ PowerMac uses water cooling, I highly doubt the Power Books will be getting a G5 - and is a G5 really all that?
What would you rather have in a laptop:
A 1.6 or 1.8GHZ G5 with a slowed down bus
or
A dual-core 2GHZ G4 with improved architecture and an on-die memory controller and in-die firewire and USB?
I'd take the G4 in a second, especially because it would likely dissappate less and be a more effective mobile chip. The IBM G5 is really ideal for desktops, but if FreeScale can deliver then Apple could have a great, very competitive portable chip.
I'd take the G4 in a second, especially because it would likely dissappate less and be a more effective mobile chip. The IBM G5 is really ideal for desktops, but if FreeScale can deliver then Apple could have a great, very competitive portable chip.
Sorry, but Dual-Core G4 in a PB is not going to happen. I know the MOTO PDF, Dual-Core 2GHz with 25W...
IBM and Intel. They all have to deal with the same problems with the 90nm process and i think Freescale, too. Why was AMD delaying the process, too?
Although...
Dual-Core G4 2GHz 25W
PPC970FX 2GHz 24.5W... idle...
You get the idea
Look at the power dissipation of the MPC7447. Bring it down to 90nm, add RIO and a memory-controller on it... 25W at 2GHzis not going to happen, Dual-Core less than ever.
You know MOSR?
Don't get me wrong, i hope you're right, but i don't think so.
Sorry, but Dual-Core G4 in a PB is not going to happen. I know the MOTO PDF, Dual-Core 2GHz with 25W...
IBM and Intel. They all have to deal with the same problems with the 90nm process and i think Freescale, too. Why was AMD delaying the process, too?
Although...
Dual-Core G4 2GHz 25W
PPC970FX 2GHz 24.5W... idle...
You get the idea
Look at the power dissipation of the MPC7447. Bring it down to 90nm, add RIO and a memory-controller on it... 25W at 2GHzis not going to happen, Dual-Core less than ever.
You know MOSR?
Don't get me wrong, i hope you're right, but i don't think so.
Well, they're going to have to go with Moto for the PowerBook until late next year. My sources at IBM are firm that they cannot deliver that kind of volume before the Fall of next year - and they have a SoC solution they're working on with Apple.
As for FreeScale getting to 90nm, I think they'll have the easiest time. They've spent a lot of time in design isolating the high heat density areas.
I don't know exactly what FreeScale is working on, but I do know the SoC project with IBM's future is dependent upon what FreeScale delivers in the next 6 months or so. Apple is very optimistic about the future of notebook solutions.
As for the dissapation with the G4, it's hard to tell. We'll know more in October. However, I think a single-core 2GHZ should run about equal or better in dissipation against the 970FX and I'd rather have the SoC option from Moto.
As for the dissapation with the G4, it's hard to tell. We'll know more in October. However, I think a single-core 2GHZ should run about equal or better in dissipation against the 970FX and I'd rather have the SoC option from Moto.
No question about that. That's for sure, but Dual-Core
Heat density is most certainly a concern. That's what hurt Intel, and it's what put liquid cooling in the latest PowerMacs.
As for signal crosstalk, I'd assume that given Motorola's less-dense designs (because of the emphasis on eliminating "hot spots" on the chip) and their use of a more sophisticated process, they'll have this in hand. They and AMD both were wise to move to 90nm slowly. They've had lots of time to work all these issues out without any public faceplants.
Heat density is most certainly a concern. That's what hurt Intel, and it's what put liquid cooling in the latest PowerMacs.
As for signal crosstalk, I'd assume that given Motorola's less-dense designs (because of the emphasis on eliminating "hot spots" on the chip) and their use of a more sophisticated process, they'll have this in hand. They and AMD both were wise to move to 90nm slowly. They've had lots of time to work all these issues out without any public faceplants.
Heat density is the result of crosstalking, IMO.
And liquid cooling was needed because of the die-shrink. Power density on the new PPC970FX 2.5GHz is the same as it was on the old PPC970 2GHz, but the die-size is the half
As you seemed to note in your next post already crosstalk, or at least current leakage in general, causes increased heat. One's just the cause of the other but heat density has been going up for all chips for a while.
On an aside the power is the same but the die area is half so the power density went up beyond what air cooling could manage.
Sorry, but Dual-Core G4 in a PB is not going to happen. I know the MOTO PDF, Dual-Core 2GHz with 25W...
Well dual core is very likely to happen in the G4 assuming that Freescale can stay around long enough to produce it. As to the PowerBook; a multiprocessor could be used to advantage in such a machine, but I really believe that the future is 64 bits. Now the question is is that 970 based 64 bits or something else. The 970FX at the moment doesn have what it takes.
Quote:
IBM and Intel. They all have to deal with the same problems with the 90nm process and i think Freescale, too. Why was AMD delaying the process, too?
There is no clear evidence that IBM and Intel are dealing with the same problems.
Quote:
Although...
Dual-Core G4 2GHz 25W
PPC970FX 2GHz 24.5W... idle...
You get the idea
Look at the power dissipation of the MPC7447. Bring it down to 90nm, add RIO and a memory-controller on it... 25W at 2GHzis not going to happen, Dual-Core less than ever.
WE are being far to negative here, there is a great deal of low power technology being developed for 90nm. The fact that IBM's first chips leave a bit to be desired power wise means nothing.
As to a high integration chip don't forget that even if it runs a bit hotter due to the integration the overall result will be a rather huge power drop. Level transistions and buffering that takes place between chips all result in wasted energy. We are at the point right now where Apple could have a custom chip built with a memroy controller and a bunch of low power serial I/O for everything else. Maybe shipping such a chip today is a bit of a stretch but I can't see it being more that a year or two away. Things such as PCI-Express, SATA, USB, and so on, are enablers for high integrations low pin count devices.
Quote:
You know MOSR?
Don't get me wrong, i hope you're right, but i don't think so.
What we will get will probally depend on what Apple is willing to throw at the problem $$$$$$$$$$ wise. Since the PowerBook / Portable market is very hot right now and is likely to be in the future, I could see where investment in SoC or high integration would be something Apple would want to pursue. For the PowerBook though I would tend to agree that this machine needs to move to a 64 bit platform soon.
What we will get will probally depend on what Apple is willing to throw at the problem $$$$$$$$$$ wise. Since the PowerBook / Portable market is very hot right now and is likely to be in the future, I could see where investment in SoC or high integration would be something Apple would want to pursue. For the PowerBook though I would tend to agree that this machine needs to move to a 64 bit platform soon.
That would be the FreeScale e700 if they can deliver, right? The biggest issue is the timing. If they can deliver the e600 early next year then that will give the PowerBook some added life - though a 64bit portable solution needs to emerge about a year later. If the e700 is on course for this then that is good.
As to the IBM/Apple SoC collaboration, I need to ask if it's 64 bit. If it remains on schedule, it should go into production by Sept/Oct of next year. I feel stupid for not asking if it was 32 or 64 now.
Well dual core is very likely to happen in the G4 assuming that Freescale can stay around long enough to produce it. As to the PowerBook; a multiprocessor could be used to advantage in such a machine, but I really believe that the future is 64 bits. Now the question is is that 970 based 64 bits or something else. The 970FX at the moment doesn have what it takes.
Dual-Core G4 is out of question, but 25W
Quote:
There is no clear evidence that IBM and Intel are dealing with the same problems.
Hmm, i don't think so. They all uses the same lithography techniques, AFAIK. And they all have problems with signal integrity, crosstalking.
What we will get will probally depend on what Apple is willing to throw at the problem $$$$$$$$$$ wise. Since the PowerBook / Portable market is very hot right now and is likely to be in the future, I could see where investment in SoC or high integration would be something Apple would want to pursue. For the PowerBook though I would tend to agree that this machine needs to move to a 64 bit platform soon.
Yeah, maybe the e700.
I'm hopping that Apple will break the 2GHz barrier as early as next year and 3GHz on the Power Mac.
Interesting... My roommate ordered a 17 inch powerbook a couple of weeks ago. It was supposed to ship yesterday and she just got an email stating that it would be delayed until September 17th. She called to find out why and both people she talked to said "We don't know."
Look at the power dissipation of the MPC7447. Bring it down to 90nm, add RIO and a memory-controller on it... 25W at 2GHzis not going to happen, Dual-Core less than ever.
You know MOSR?
Don't get me wrong, i hope you're right, but i don't think so.
The 7447 low power @ 1.1V dissipates just 9.3W @ 1.2G at the current process.
So 2G @ 20W is doable - but when?.
Also, on the PC side you have the Centrino for the laptops and P4's for the desktop so I think it will be fine if the G4 remains in the laptops - especially if Freescale can get those dual core G4's FH of 2005!.
A 2GHz G4 will be comparable to the Dothan well close at least.
Comments
Originally posted by Peter North
what about a 64bit chip for tiger early next year? even for mere marketing purposes? is the e700 something to look to?
It is not known if the e600 and e700 will go into production simultaneously. From the pdf posted above in this thread, it looks like the e600 will show up first and the e700 will follow. I don't believe that we will see the e700 next year.
Originally posted by PB
Good catch Freddy. I think it is new news
Obviously, we talk about the e600 here. So, we can expect the e600 in the Powerbooks for the next spring, or rather beginning of summer 2005, I think.
I think it's pretty sure that the next PB revision (Q1/ 2005) will sport a G5.
G5 iMac and G4 PB doesn't match.
Maybe the e700 (G5-M
Power Mac G5 = IBM PPC970MP (Q1/2005), with DDR2 & PCIe *drivel*
Powerbook G5 = IBM PPC970FX or Freescale e700 (Q1/2005)
iMac G5 = IBM PPC970FX (Q4/ 2004)
iBook G4 Extreme = Freescale e600 (Q4/ 2004)
eMac G4 Extreme = Freescale e600 (Q4/ 2004)
Cube2 G4 Extreme = Freescale e600 (Q4/ 2004)
But i have the feeling that Freescale chips don't fit into Apples product line. Last but not least because of RIO.
Just my two ¢
Originally posted by Fat Freddy
iMac G5 = IBM PPC970FX (Q4/ 2004)
Just my two ¢
Apple's Q4, or the real Q4?
They are releasing this in September, or at least that is what they have made the entire world believe.
Originally posted by NovaVengeance
Apple's Q4, or the real Q4?
They are releasing this in September, or at least that is what they have made the entire world believe.
I mean real Q4 (Oct. - Dec.).
Apple says September, but shipping in October
But you're right
Originally posted by Fat Freddy
I think it's pretty sure that the next PB revision (Q1/ 2005) will sport a G5.
G5 iMac and G4 PB doesn't match.
Maybe the e700 (G5-M
Power Mac G5 = IBM PPC970MP (Q1/2005), with DDR2 & PCIe *drivel*
Powerbook G5 = IBM PPC970FX or Freescale e700 (Q1/2005)
iMac G5 = IBM PPC970FX (Q4/ 2004)
iBook G4 Extreme = Freescale e600 (Q4/ 2004)
eMac G4 Extreme = Freescale e600 (Q4/ 2004)
Cube2 G4 Extreme = Freescale e600 (Q4/ 2004)
But i have the feeling that Freescale chips don't fit into Apples product line. Last but not least because of RIO.
Just my two ¢
Seeing as the 2.5 GHZ PowerMac uses water cooling, I highly doubt the Power Books will be getting a G5 - and is a G5 really all that?
What would you rather have in a laptop:
A 1.6 or 1.8GHZ G5 with a slowed down bus
or
A dual-core 2GHZ G4 with improved architecture and an on-die memory controller and in-die firewire and USB?
I'd take the G4 in a second, especially because it would likely dissappate less and be a more effective mobile chip. The IBM G5 is really ideal for desktops, but if FreeScale can deliver then Apple could have a great, very competitive portable chip.
Originally posted by Mr. MacPhisto
I'd take the G4 in a second, especially because it would likely dissappate less and be a more effective mobile chip. The IBM G5 is really ideal for desktops, but if FreeScale can deliver then Apple could have a great, very competitive portable chip.
Sorry, but Dual-Core G4 in a PB is not going to happen. I know the MOTO PDF, Dual-Core 2GHz with 25W...
IBM and Intel. They all have to deal with the same problems with the 90nm process and i think Freescale, too. Why was AMD delaying the process, too?
Although...
Dual-Core G4 2GHz 25W
PPC970FX 2GHz 24.5W... idle...
You get the idea
Look at the power dissipation of the MPC7447. Bring it down to 90nm, add RIO and a memory-controller on it... 25W at 2GHzis not going to happen, Dual-Core less than ever.
You know MOSR?
Don't get me wrong, i hope you're right, but i don't think so.
Originally posted by Fat Freddy
Sorry, but Dual-Core G4 in a PB is not going to happen. I know the MOTO PDF, Dual-Core 2GHz with 25W...
IBM and Intel. They all have to deal with the same problems with the 90nm process and i think Freescale, too. Why was AMD delaying the process, too?
Although...
Dual-Core G4 2GHz 25W
PPC970FX 2GHz 24.5W... idle...
You get the idea
Look at the power dissipation of the MPC7447. Bring it down to 90nm, add RIO and a memory-controller on it... 25W at 2GHzis not going to happen, Dual-Core less than ever.
You know MOSR?
Don't get me wrong, i hope you're right, but i don't think so.
Well, they're going to have to go with Moto for the PowerBook until late next year. My sources at IBM are firm that they cannot deliver that kind of volume before the Fall of next year - and they have a SoC solution they're working on with Apple.
As for FreeScale getting to 90nm, I think they'll have the easiest time. They've spent a lot of time in design isolating the high heat density areas.
I don't know exactly what FreeScale is working on, but I do know the SoC project with IBM's future is dependent upon what FreeScale delivers in the next 6 months or so. Apple is very optimistic about the future of notebook solutions.
As for the dissapation with the G4, it's hard to tell. We'll know more in October. However, I think a single-core 2GHZ should run about equal or better in dissipation against the 970FX and I'd rather have the SoC option from Moto.
Originally posted by Mr. MacPhisto
As for FreeScale getting to 90nm, I think they'll have the easiest time. They've spent a lot of time in design isolating the high heat density areas.
Not heat density, crosstalking is the keyword: http://arstechnica.com/news/posts/20040426-3697.html
As for the dissapation with the G4, it's hard to tell. We'll know more in October. However, I think a single-core 2GHZ should run about equal or better in dissipation against the 970FX and I'd rather have the SoC option from Moto.
No question about that. That's for sure, but Dual-Core
That's we're talking about, isn't it?
Originally posted by Fat Freddy
Not heat density, crosstalking is the keyword: http://arstechnica.com/news/posts/20040426-3697.html
Heat density is most certainly a concern. That's what hurt Intel, and it's what put liquid cooling in the latest PowerMacs.
As for signal crosstalk, I'd assume that given Motorola's less-dense designs (because of the emphasis on eliminating "hot spots" on the chip) and their use of a more sophisticated process, they'll have this in hand. They and AMD both were wise to move to 90nm slowly. They've had lots of time to work all these issues out without any public faceplants.
Originally posted by Amorph
Heat density is most certainly a concern. That's what hurt Intel, and it's what put liquid cooling in the latest PowerMacs.
As for signal crosstalk, I'd assume that given Motorola's less-dense designs (because of the emphasis on eliminating "hot spots" on the chip) and their use of a more sophisticated process, they'll have this in hand. They and AMD both were wise to move to 90nm slowly. They've had lots of time to work all these issues out without any public faceplants.
Heat density is the result of crosstalking, IMO.
And liquid cooling was needed because of the die-shrink. Power density on the new PPC970FX 2.5GHz is the same as it was on the old PPC970 2GHz, but the die-size is the half
Originally posted by Fat Freddy
Not heat density, crosstalking is the keyword
As you seemed to note in your next post already crosstalk, or at least current leakage in general, causes increased heat. One's just the cause of the other but heat density has been going up for all chips for a while.
On an aside the power is the same but the die area is half so the power density went up beyond what air cooling could manage.
Originally posted by Telomar
On an aside the power is the same but the die area is half so the power density went up beyond what air cooling could manage.
Oh sorry!
Of course, the power density is twice as much as before. Power dissipation is the same as it was on the PPC970 at 2GHz.
My fault.
Originally posted by Fat Freddy
Sorry, but Dual-Core G4 in a PB is not going to happen. I know the MOTO PDF, Dual-Core 2GHz with 25W...
Well dual core is very likely to happen in the G4 assuming that Freescale can stay around long enough to produce it. As to the PowerBook; a multiprocessor could be used to advantage in such a machine, but I really believe that the future is 64 bits. Now the question is is that 970 based 64 bits or something else. The 970FX at the moment doesn have what it takes.
IBM and Intel. They all have to deal with the same problems with the 90nm process and i think Freescale, too. Why was AMD delaying the process, too?
There is no clear evidence that IBM and Intel are dealing with the same problems.
Although...
Dual-Core G4 2GHz 25W
PPC970FX 2GHz 24.5W... idle...
You get the idea
Look at the power dissipation of the MPC7447. Bring it down to 90nm, add RIO and a memory-controller on it... 25W at 2GHzis not going to happen, Dual-Core less than ever.
WE are being far to negative here, there is a great deal of low power technology being developed for 90nm. The fact that IBM's first chips leave a bit to be desired power wise means nothing.
As to a high integration chip don't forget that even if it runs a bit hotter due to the integration the overall result will be a rather huge power drop. Level transistions and buffering that takes place between chips all result in wasted energy. We are at the point right now where Apple could have a custom chip built with a memroy controller and a bunch of low power serial I/O for everything else. Maybe shipping such a chip today is a bit of a stretch but I can't see it being more that a year or two away. Things such as PCI-Express, SATA, USB, and so on, are enablers for high integrations low pin count devices.
You know MOSR?
Don't get me wrong, i hope you're right, but i don't think so.
What we will get will probally depend on what Apple is willing to throw at the problem $$$$$$$$$$ wise. Since the PowerBook / Portable market is very hot right now and is likely to be in the future, I could see where investment in SoC or high integration would be something Apple would want to pursue. For the PowerBook though I would tend to agree that this machine needs to move to a 64 bit platform soon.
Originally posted by wizard69
What we will get will probally depend on what Apple is willing to throw at the problem $$$$$$$$$$ wise. Since the PowerBook / Portable market is very hot right now and is likely to be in the future, I could see where investment in SoC or high integration would be something Apple would want to pursue. For the PowerBook though I would tend to agree that this machine needs to move to a 64 bit platform soon.
That would be the FreeScale e700 if they can deliver, right? The biggest issue is the timing. If they can deliver the e600 early next year then that will give the PowerBook some added life - though a 64bit portable solution needs to emerge about a year later. If the e700 is on course for this then that is good.
As to the IBM/Apple SoC collaboration, I need to ask if it's 64 bit. If it remains on schedule, it should go into production by Sept/Oct of next year. I feel stupid for not asking if it was 32 or 64 now.
Originally posted by wizard69
Well dual core is very likely to happen in the G4 assuming that Freescale can stay around long enough to produce it. As to the PowerBook; a multiprocessor could be used to advantage in such a machine, but I really believe that the future is 64 bits. Now the question is is that 970 based 64 bits or something else. The 970FX at the moment doesn have what it takes.
Dual-Core G4 is out of question, but 25W
There is no clear evidence that IBM and Intel are dealing with the same problems.
Hmm, i don't think so. They all uses the same lithography techniques, AFAIK. And they all have problems with signal integrity, crosstalking.
http://arstechnica.com/news/posts/20040426-3697.html
What we will get will probally depend on what Apple is willing to throw at the problem $$$$$$$$$$ wise. Since the PowerBook / Portable market is very hot right now and is likely to be in the future, I could see where investment in SoC or high integration would be something Apple would want to pursue. For the PowerBook though I would tend to agree that this machine needs to move to a 64 bit platform soon.
Yeah, maybe the e700.
I'm hopping that Apple will break the 2GHz barrier as early as next year and 3GHz on the Power Mac.
Hmmm.
Interesting info, Arty50.
The 2G sounds MUCH better - but MWSF (JAN 05) does not. Here's hoping.....
Originally posted by Fat Freddy
Look at the power dissipation of the MPC7447. Bring it down to 90nm, add RIO and a memory-controller on it... 25W at 2GHzis not going to happen, Dual-Core less than ever.
You know MOSR?
Don't get me wrong, i hope you're right, but i don't think so.
The 7447 low power @ 1.1V dissipates just 9.3W @ 1.2G at the current process.
So 2G @ 20W is doable - but when?.
Also, on the PC side you have the Centrino for the laptops and P4's for the desktop so I think it will be fine if the G4 remains in the laptops - especially if Freescale can get those dual core G4's FH of 2005!.
A 2GHz G4 will be comparable to the Dothan well close at least.