I think there will be some more updates to the G4 processor. Why did they call the Apollo6 the 'MPC7455' instead of 'MPC7460' ? If you continue in steps of 10 this would only have left space for 3 future versions (7470, 7480 and 7490).
I've heard that Apple and Motorola intend to use a strategy similar to the OS strategy after the Copland fiasko. Instead of one big update they will use technology from the next generation and implement it into the current G4 using time intervals of about 6 months.
The next G4 version shall get a DDR interface borrowed from the MPC7500 and can be expected for mid year (SOI 0.13 @ about 1.5 Ghz). Then, in early 2003, the G4 will get a 10 stage pipeline plus 512kb on-die L2-cache and reach 2 Ghz.
This might also explain the rumors about the G5 being 'killed' by Apple. The first MPC using RapidIO should then officially be called 'G5' by Apple.
"I think there will be some more updates to the G4 processor."
Groan. I think most of us are resigned to this.
"Why did they call the Apollo6 the 'MPC7455' instead of 'MPC7460' ? If you continue in steps of 10 this would only have left space for 3 future versions (7470, 7480 and 7490)."
..and if the G4 remains, fundamentally the same processor...then Apple are going to look laughably out of date and ridiculously overprice even with a superior mobo revision.
"I've heard that Apple and Motorola intend to use a strategy similar to the OS strategy after the Copland fiasko. Instead of one big update they will use technology from the next generation and implement it into the current G4 using time intervals of about 6 months."
Using technology from the next revision is a fair idea. Guess it keeps to a bare minimum things that can go 'wrong'. However, the G4 fiasco has left us wanting more than a few things borrowed. Some Apple watchers are happy with a mobo G4 revision. I think that's the least we should have had at San Fran this year and overdue at that!
"The next G4 version shall get a DDR interface borrowed from the MPC7500 and can be expected for mid year (SOI 0.13 @ about 1.5 Ghz)."
That would have been a 'good' (but not excellent ...) machine at this year's San Fran'...but now it's painfully over due...
"Then, in early 2003, the G4 will get a 10 stage pipeline plus 512kb on-die L2-cache and reach 2 Ghz."
Yeah. But even if that were to be delivered at Mac New York this year. It's still catch up. It aint leap frogging the opposition. Apple customers are still paying over the odds for catch up performance.
"This might also explain the rumors about the G5 being 'killed' by Apple. The first MPC using RapidIO should then officially be called 'G5' by Apple."
The first G4 in a Rapid IO framework will be called 'G5'? I think that's fraudulent. It's a con.
Unless...something's done to boost the G4's performance fundamentally.
It's FPU is a joke compared to the Athlon. Take away Altivec and the jumped up G3/G4 is a joke.
I hope your scenario is the worst case. Good post by the way.
<strong>I think there will be some more updates to the G4 processor. Why did they call the Apollo6 the 'MPC7455' instead of 'MPC7460' ? If you continue in steps of 10 this would only have left space for 3 future versions (7470, 7480 and 7490).
I've heard that Apple and Motorola intend to use a strategy similar to the OS strategy after the Copland fiasko. Instead of one big update they will use technology from the next generation and implement it into the current G4 using time intervals of about 6 months.
The next G4 version shall get a DDR interface borrowed from the MPC7500 and can be expected for mid year (SOI 0.13 @ about 1.5 Ghz). Then, in early 2003, the G4 will get a 10 stage pipeline plus 512kb on-die L2-cache and reach 2 Ghz.</strong><hr></blockquote>
This is a reasonable assumption. Look at nVidia and how their strategy allows for 6 month intervals between updates. This also keeps them ahead of the competition.
Regarding the L2 cache, I believe we'll see 512kb cache this summer as the standard G4 7460?? migrates to the .13u HiP7 process.
I'm expecting dual 1.2Ghz systems on the same MOBO this summer at the minimum. Anything beyond that will be more than expected and long overdue. Dual 1.4 Ghz systems with DDR are certaintly possible and would be more than expected.
I think Apple has got to come out with much faster hardware soon if they want pros to buy Shake for OS X. It was announced that Shake 2.5 will be the last for Windows. They will still do Irix and Linux.
DDR mobo with a 2 GHZ 7460 would be about right. MWSF 03?
<strong>I think there will be some more updates to the G4 processor. Why did they call the Apollo6 the 'MPC7455' instead of 'MPC7460' ? If you continue in steps of 10 this would only have left space for 3 future versions (7470, 7480 and 7490).
I've heard that Apple and Motorola intend to use a strategy similar to the OS strategy after the Copland fiasko. Instead of one big update they will use technology from the next generation and implement it into the current G4 using time intervals of about 6 months.
The next G4 version shall get a DDR interface borrowed from the MPC7500 and can be expected for mid year (SOI 0.13 @ about 1.5 Ghz). Then, in early 2003, the G4 will get a 10 stage pipeline plus 512kb on-die L2-cache and reach 2 Ghz.</strong><hr></blockquote>
Yes, the G4 is going to get technology from the G5. It will "trickle down" to the G4, as a way to keep the G4 modern and maintain its clockspeed.
As Apple's consumer chip, the G4 will still need to be under development so that it's clock speed remains competitive. But this does not mean that the G5 no longer exists. On the contrary, it means that the G5 is the chip that will implement Motorola's newest technologies.
As these new technologies are implemented in the G5, they will then be employed in the G4's design. But the the G5 will still have a new and improved architecture.
A key development is that Apple has turned to Motorola for virtually all of their CPUs. Imagine how much money is now flowing into Motorola from their desktop PPC sales! Only a year ago, Apple was buying only G4s for their Powermac lineup from Motorola, now Apple is using Motorola CPU's in all of their Macs except the iBook (which will soon get a G4).
Would Apple give Motorola so much business if they had dropped the ball on the G5? If they had nothing promising for Apple? If they couldn't supply Apple with competitive CPUs? I think not. I believe what is happening is that Motorola are reaping the rewards for their recent commitment to PPC development, and that Apple will soon be buying ALL of their CPU from Motorola, including the G5.
It is an exciting time for Powermac development, and we are going to see the shift to a new PPC architecture and a new high-bandwidth mobo in the near future. Apple is once again going to own the professional high performance hardware market...the new powermacs will be the envy of the Wintel world. Hold onto your butts 'cause it's going to be HUGE!!!
<strong>A key development is that Apple has turned to Motorola for virtually all of their CPUs. Imagine how much money is now flowing into Motorola from their desktop PPC sales! Only a year ago, Apple was buying only G4s for their Powermac lineup from Motorola, now Apple is using Motorola CPU's in all of their Macs except the iBook (which will soon get a G4).</strong><hr></blockquote>
what so many seem to forget is that the embedded market presents much greater sales numbers than the desktop market. compared to many other parts of their business, motorola makes very little money selling chips to apple. there are probably 30 times as many embedded chips in my car than their are in my computer.
True, but Motorola is now selling over twice as many G4 chips to Apple. While this may not be a great number in comparison to moto's embedded chip sales, it's still a significant increase in desktop PPC sales that have a positive impact on Moto's commitment to the G5.
Not only is this good news for Motorola, but perhaps more importantly, it demonstrates to Moto that there still is good money to be made in desktop PPC sales, and that there is the POTENTIAL for a successful G5 to bring in mondo sales for Motorola. Apple's commitment to the G4 shows Motorola what might happen if they deliver on their G5 rhetoric, and surely Motorola must recognize that if Apple's marketshare improves, they will directly benefit. Apple is dangling the carrot in front of Moto--all they need to do is give Apple a bitchin' CPU and then it's wide open for OS X's market share to skyrocket.
<strong>True, but Motorola is now selling over twice as many G4 chips to Apple. While this may not be a great number in comparison to moto's embedded chip sales, it's still a significant increase in desktop PPC sales that have a positive impact on Moto's commitment to the G5.
Not only is this good news for Motorola, but perhaps more importantly, it demonstrates to Moto that there still is good money to be made in desktop PPC sales, and that there is the POTENTIAL for a successful G5 to bring in mondo sales for Motorola. Apple's commitment to the G4 shows Motorola what might happen if they deliver on their G5 rhetoric, and surely Motorola must recognize that if Apple's marketshare improves, they will directly benefit. Apple is dangling the carrot in front of Moto--all they need to do is give Apple a bitchin' CPU and then it's wide open for OS X's market share to skyrocket.
If I was running Moto than I would have to take a long hard look at our relationship with Apple. Not much sales as compared with the other chip sales. Yet Apple demands higher performance chips which pushes costs up, but Apple pays for R&D that can be transferred to other chip products. This is common sense and I write this to point out that there is a balance point. It appears to me from the outside looking in that IBM is a better fit right now. IBM already makes the Power4 which is high end, I'm not saying that Apple will use it, rather that the infrastructure to make high end chips in quantity is already there. IBM ups the capabilities of the infrastructure for themselves and for he other customers that come to them for custom chips. So why not use IBM?? Apple has a relationship with Moto and that is a business relationship. It appears to me that Moto would be better off in concentrating on developing that new communications chip technology that they have, and let IBM take over the Apple next chip development and production. Moto can still supply the G4 for the consumer market of Apple since they already are making them. IBM would take over development of the next chip G5 whatever that is. BookE said it all, "this will make it very easy for PPC development to be transferred between the the companies that are developing and manufacturing PPC chips.
If I was running Moto than I would have to take a long hard look at our relationship with Apple. Not much sales as compared with the other chip sales. Yet Apple demands higher performance chips which pushes costs up, but Apple pays for R&D that can be transferred to other chip products. This is common sense and I write this to point out that there is a balance point.</strong><hr></blockquote>
"Embedded" doesn't mean "wimpy." There are embedded applications (high-end routers, telephone switches) that crush any desktop in terms of raw computational power. The 7455 has won accolades in the embedded-chip press, and the likes of Cisco will buy the fastest chips Mot can crank out. So it's not just Apple pushing them; there's a high-end embedded market, and Motorola's moving into it.
"Embedded" doesn't mean "wimpy." There are embedded applications (high-end routers, telephone switches) that crush any desktop in terms of raw computational power. The 7455 has won accolades in the embedded-chip press, and the likes of Cisco will buy the fastest chips Mot can crank out. So it's not just Apple pushing them; there's a high-end embedded market, and Motorola's moving into it.</strong><hr></blockquote>
Sorry I did not mean to imply embedded==wimpy. Good point. I agree there are very high end embedded chips. My point was that I think that Moto would be better off developing their new technology and still improve the 7xxx series. If Apple wants way beyond that, I guess I hope they do, than Moto would be better served by concentrating on what's on their own plate. If we are looking at two or three FPUs and DDR memory than Moto can handle that. I think that I am implying that Apple might be after something much more, as was eluded to earlier.
<strong>If I was running Moto than I would have to take a long hard look at our relationship with Apple. Not much sales as compared with the other chip sales. Yet Apple demands higher performance chips which pushes costs up, but Apple pays for R&D that can be transferred to other chip products. This is common sense and I write this to point out that there is a balance point. It appears to me from the outside looking in that IBM is a better fit right now. IBM already makes the Power4 which is high end, I'm not saying that Apple will use it, rather that the infrastructure to make high end chips in quantity is already there. IBM ups the capabilities of the infrastructure for themselves and for he other customers that come to them for custom chips. So why not use IBM?? Apple has a relationship with Moto and that is a business relationship. It appears to me that Moto would be better off in concentrating on developing that new communications chip technology that they have, and let IBM take over the Apple next chip development and production. Moto can still supply the G4 for the consumer market of Apple since they already are making them. IBM would take over development of the next chip G5 whatever that is. BookE said it all, "this will make it very easy for PPC development to be transferred between the the companies that are developing and manufacturing PPC chips.
That is what I see at least.</strong><hr></blockquote>
exactly.
[quote]<strong>
True, but Motorola is now selling over twice as many G4 chips to Apple. While this may not be a great number in comparison to moto's embedded chip sales, it's still a significant increase in desktop PPC sales that have a positive impact on Moto's commitment to the G5.
Not only is this good news for Motorola, but perhaps more importantly, it demonstrates to Moto that there still is good money to be made in desktop PPC sales, and that there is the POTENTIAL for a successful G5 to bring in mondo sales for Motorola. Apple's commitment to the G4 shows Motorola what might happen if they deliver on their G5 rhetoric, and surely Motorola must recognize that if Apple's marketshare improves, they will directly benefit. Apple is dangling the carrot in front of Moto--all they need to do is give Apple a bitchin' CPU and then it's wide open for OS X's market share to skyrocket.</strong><hr></blockquote>
well, twice as much of very little is still not much. add to the low sales the fact that apple is a very demanding company to work with and i can see how the relationship could have soured.
let's say you make $100k a year performing a service. i'm one of your customers and i pay you $2k a year, but take up as much time and resources as one of your $30k clients. me paying $4k the next year only makes me somewhat more tolerable, but still not a good or significant enough customer to cause you so much grief.
The G5 will appear as soon as Motorola will be able to fab efficiently in SOI 0,13 .
There is only one company now able to produce efficiently on SOI 0,13 it's IBM. AMD is not supposed to produce SOI 0,13 until the end of the year, and for the moment does not produc 0,13 micron process. Intel fab on 0,13 process but not on SOI (or equivalent).
Mot will continue to improve the G4 line (perhaps using some G5 tweaks) until he can reach o,13 SOI. When Mot will reach this limit, the G5 will appear. The architecture is ready but not the fab process. Contrary to the past where Mot was stuck at 500 mhz, Mot will not wait the new fab process without do any improvement on his chip line. We have the 7455 who add just a new gestion of the L3 cache, we will have soon the 7460 that will have DDR memory controller, but the same amount of L2 cache if still make on a 0,18 micron process. I think that the 512 L2 cache will wait until 0,13 , like the Intel chips.
So i make this statement : 7460 with ddr memory controller for MWNY. G5 in 2003 perhaps in january if Mot is not too late in the developpement of 0,13 SOI.
I am not a chip engineer, but iam ready to bet it's more simple to make an new chip architecture than to develop a new fab process.
I am not a chip engineer, but iam ready to bet it's more simple to make an new chip architecture than to develop a new fab process.</strong><hr></blockquote>
Could you please elaborate.
If this relates to the IBM stuff, BookE was mentioned. And IBM has manufactured the G4 for Apple when supplies from Moto were low.
If this relates to the IBM stuff, BookE was mentioned. And IBM has manufactured the G4 for Apple when supplies from Moto were low.</strong><hr></blockquote>
I'll will say that the amount of R&D necessary for developping new fab process is more important than the amount of R&D needed for chip design. Motorola alone is not anymore rich enough to develop new fab process : that's why with Philips and other he created a new research center in France. Many companies are able to design chips, like nvidia, but very few companies are able to develop new fab process. You certainly know the Moore's law about the number of transistor included in a chip, there is an another law, fab process are more and more expansive in an exponantial way either.
I'll will say that the amount of R&D necessary for developping new fab process is more important than the amount of R&D needed for chip design. Motorola alone is not anymore rich enough to develop new fab process : that's why with Philips and other he created a new research center in France. Many companies are able to design chips, like nvidia, but very few companies are able to develop new fab process. You certainly know the Moore's law about the number of transistor included in a chip, there is an another law, fab process are more and more expansive in an exponantial way either.</strong><hr></blockquote>
Making Other Companies Chips is something that IBM does, they are just a hired gun, or manufacturer if you will. This is nothing new. Am I missing something? If you wanted to develop the PowerDoc G5 chip, and you had the money, IBM would make it for you, in quantity. It would help allot and probably be cheaper if the design were BookE compliant. This is where Apple could be. They maybe have designed a chip based on BookE, or have asked IBM to develop a chip for them based on BookE. As far as capacity is concerned IBM may have more than enough capacity. And IBM has leading technology that is shipping. Copper, SOI, Low k dielectric. I don't think that IBM would have to add capacity for Apple, is that what you are saying? I mean to be conservative IBM would need at the outside 4 Million chips to more than cover Apples orders, that may sound like allot, but I venture that 4mil is a drop in the bucket.
Making Other Companies Chips is something that IBM does, they are just a hired gun, or manufacturer if you will. This is nothing new. Am I missing something? If you wanted to develop the PowerDoc G5 chip, and you had the money, IBM would make it for you, in quantity. It would help allot and probably be cheaper if the design were BookE compliant. This is where Apple could be. They maybe have designed a chip based on BookE, or have asked IBM to develop a chip for them based on BookE. As far as capacity is concerned IBM may have more than enough capacity. And IBM has leading technology that is shipping. Copper, SOI, Low k dielectric. I don't think that IBM would have to add capacity for Apple, is that what you are saying? I mean to be conservative IBM would need at the outside 4 Million chips to more than cover Apples orders, that may sound like allot, but I venture that 4mil is a drop in the bucket.</strong><hr></blockquote>
I agree with what you say , but i was speakind of Mot and not IBM. Mot own the altivec licenses. Without Altivec the PPC chips are very crappy and will loose all the benchmarks against the X86 chips. Perhaps Mot want to keep one of this biggest customer : Apple and wont let him going elsewhere. It's possible that Mot does not wanted to license altivec to anyone else to protect him.
Without altivec , i am ready to bet that all the PPC Apple chips will be fabbed by IBM.
I agree with what you say , but i was speakind of Mot and not IBM. Mot own the altivec licenses. Without Altivec the PPC chips are very crappy and will loose all the benchmarks against the X86 chips. Perhaps Mot want to keep one of this biggest customer : Apple and wont let him going elsewhere. It's possible that Mot does not wanted to license altivec to anyone else to protect him.
Without altivec , i am ready to bet that all the PPC Apple chips will be fabbed by IBM.</strong><hr></blockquote>
Ah Ha...Altivec, I see, I agree. It does seem that Apple would be interested in Altivec, but what if instead of buying Altivec Apple licenses it from Moto. That way Moto can keep making chips for Cisco, and Apple can then still have it included in the chips made for them. I think that this would be good for both. And would be much cheaper than trying to buy the entire PPC license from Moto, which I always thought was unrealistic. And I think that Apple had some hand, probably small, in developing the initial version of altivec. If i remember correctly Apple had more to do with the latest version of Altivec. There was a time, maybe a year ago maybe a little longer that Moto announced a slightly improved version of Altivec. I think that Apple had more than a passive role in bring that to market. So co-licensing Altivec could be done, and would still serve both companies. I think that Apple has the talent on board to develop their own next version of altivec.
[quote]Work on the G5 (MPC7500) is progressing nicely at Motorola and I can only assume Apple had a role in the development as there are many features included that are condusive to a top-notch desktop processor capable of bring the PowerMac into the 21st centure as a real performer in the compute intensive field.
A recent development box has allowed us to open and examine some details that appeal to us as engineers and hardware designers. One of the machines is built in a G4 case with a smaller motherboard. the interesting part is that the motherboard is a cheaper 4 layer variety while the CPU daughter card is made on a more expensive 8 layer board. the reason for this is very simple and I will get to that in a moment. The CPU card itself is similar in shape and pin connector to the one used in modern G4s but is a bit larger in surface area. The reason for this is the fact that the G5 CPU has a built-in memory controller that interacts with high-speed DDR-SDRAM. This RAM runs at 133MHz DDR but the G5 documentation we have (preliminary) states that 166MHz DDR (333MHz effective) is supported. To run at these fast speeds you must utilize 2 dimensional PCB space efficiently. Apple has done this well. On the CPU daughter card you will find 3 slots for DDR-SDRAM DIMMS. They are distanced from the actual CPU core to allow for the heatsink placement, but in this configuration (since on the card the CPU is hard soldered to the PCB) you can design DIMM slots that can operate at 133, 166, and quite possibly 200MHz at DDR speeds. In fact, I believe the impedence should be low enough to allow for 4 DIMM slots. No L3 RAM is located on the card although it is fully supported under the documentation.
RapidIO is used to connect the CPU to a centralized controller that controls peripherals such as gigabit ethernet, Firewire, USB, ATA interface, audio input/output, AGP/PCI bridge, etc. The RapidIO interface operates in a 16 bit full duplex mode running at 500MHz. This bus is seperate from the memory bus and therefore asyncronous. It does not share the bus either like with the G4 and MPX. PCI-X seems to be implimented fully (it's backwards compatible with PCI 32/64) and AGP remains operating at 4X speeds.
A whole new paradigm in PowerMac design will be introduced soon and it will be a modern example in smart computer design, taking advantage of new trechnologies and foward thinking design. These are late beta machines and the finishing touches are being applied to cut out the rough edges for consumer consumption. <hr></blockquote>
Comments
I thought the G5 was going to be released in a '32 bit' version first...then 64 bit some time later?
So that's 'two' G5s...but...
But it looks as if there's going to be at least one more round of G4s before that.
So it's Apollo 2nd round of G4s at 1 gig to 1.6?
Then 7500 (G5?) at 32 bit flavour first...
...then 64 bit flavour 2nd.
So...Apollo bump at Newyork this year.
7500 introduced at San Fran 2003.
64 bit flavour at New York 2003?
Just guesses.
What further complicates these issues is the rumoured IBM chip and talk of Moto's 'G5' being supposedly 'rejected'.
A recent 'news' item at the register indicated IBM is to ready some PowerPCs that are multicore with Rapid IO...and G3s at up to 2 gig with SIMD unit?
Are Moto's and IBM's paths converging after the Somerset rift?
I'm dying to see how the processor thing plays out over the next year.
The AMD/Intel slugfest is indirectly piling on the pressue for the AIM alliance to do 'SOMETHING'!!!
Lemon Bon Bon <img src="graemlins/hmmm.gif" border="0" alt="[Hmmm]" />
I've heard that Apple and Motorola intend to use a strategy similar to the OS strategy after the Copland fiasko. Instead of one big update they will use technology from the next generation and implement it into the current G4 using time intervals of about 6 months.
The next G4 version shall get a DDR interface borrowed from the MPC7500 and can be expected for mid year (SOI 0.13 @ about 1.5 Ghz). Then, in early 2003, the G4 will get a 10 stage pipeline plus 512kb on-die L2-cache and reach 2 Ghz.
This might also explain the rumors about the G5 being 'killed' by Apple. The first MPC using RapidIO should then officially be called 'G5' by Apple.
Groan. I think most of us are resigned to this.
"Why did they call the Apollo6 the 'MPC7455' instead of 'MPC7460' ? If you continue in steps of 10 this would only have left space for 3 future versions (7470, 7480 and 7490)."
..and if the G4 remains, fundamentally the same processor...then Apple are going to look laughably out of date and ridiculously overprice even with a superior mobo revision.
"I've heard that Apple and Motorola intend to use a strategy similar to the OS strategy after the Copland fiasko. Instead of one big update they will use technology from the next generation and implement it into the current G4 using time intervals of about 6 months."
Using technology from the next revision is a fair idea. Guess it keeps to a bare minimum things that can go 'wrong'. However, the G4 fiasco has left us wanting more than a few things borrowed. Some Apple watchers are happy with a mobo G4 revision. I think that's the least we should have had at San Fran this year and overdue at that!
"The next G4 version shall get a DDR interface borrowed from the MPC7500 and can be expected for mid year (SOI 0.13 @ about 1.5 Ghz)."
That would have been a 'good' (but not excellent ...) machine at this year's San Fran'...but now it's painfully over due...
"Then, in early 2003, the G4 will get a 10 stage pipeline plus 512kb on-die L2-cache and reach 2 Ghz."
Yeah. But even if that were to be delivered at Mac New York this year. It's still catch up. It aint leap frogging the opposition. Apple customers are still paying over the odds for catch up performance.
"This might also explain the rumors about the G5 being 'killed' by Apple. The first MPC using RapidIO should then officially be called 'G5' by Apple."
The first G4 in a Rapid IO framework will be called 'G5'? I think that's fraudulent. It's a con.
Unless...something's done to boost the G4's performance fundamentally.
It's FPU is a joke compared to the Athlon. Take away Altivec and the jumped up G3/G4 is a joke.
I hope your scenario is the worst case. Good post by the way.
Lemon Bon Bon
<strong>I think there will be some more updates to the G4 processor. Why did they call the Apollo6 the 'MPC7455' instead of 'MPC7460' ? If you continue in steps of 10 this would only have left space for 3 future versions (7470, 7480 and 7490).
I've heard that Apple and Motorola intend to use a strategy similar to the OS strategy after the Copland fiasko. Instead of one big update they will use technology from the next generation and implement it into the current G4 using time intervals of about 6 months.
The next G4 version shall get a DDR interface borrowed from the MPC7500 and can be expected for mid year (SOI 0.13 @ about 1.5 Ghz). Then, in early 2003, the G4 will get a 10 stage pipeline plus 512kb on-die L2-cache and reach 2 Ghz.</strong><hr></blockquote>
This is a reasonable assumption. Look at nVidia and how their strategy allows for 6 month intervals between updates. This also keeps them ahead of the competition.
Regarding the L2 cache, I believe we'll see 512kb cache this summer as the standard G4 7460?? migrates to the .13u HiP7 process.
I'm expecting dual 1.2Ghz systems on the same MOBO this summer at the minimum. Anything beyond that will be more than expected and long overdue. Dual 1.4 Ghz systems with DDR are certaintly possible and would be more than expected.
Steve
DDR mobo with a 2 GHZ 7460 would be about right. MWSF 03?
[ 05-03-2002: Message edited by: CodeWarrior ]</p>
<strong>I think there will be some more updates to the G4 processor. Why did they call the Apollo6 the 'MPC7455' instead of 'MPC7460' ? If you continue in steps of 10 this would only have left space for 3 future versions (7470, 7480 and 7490).
I've heard that Apple and Motorola intend to use a strategy similar to the OS strategy after the Copland fiasko. Instead of one big update they will use technology from the next generation and implement it into the current G4 using time intervals of about 6 months.
The next G4 version shall get a DDR interface borrowed from the MPC7500 and can be expected for mid year (SOI 0.13 @ about 1.5 Ghz). Then, in early 2003, the G4 will get a 10 stage pipeline plus 512kb on-die L2-cache and reach 2 Ghz.</strong><hr></blockquote>
Yes, the G4 is going to get technology from the G5. It will "trickle down" to the G4, as a way to keep the G4 modern and maintain its clockspeed.
As Apple's consumer chip, the G4 will still need to be under development so that it's clock speed remains competitive. But this does not mean that the G5 no longer exists. On the contrary, it means that the G5 is the chip that will implement Motorola's newest technologies.
As these new technologies are implemented in the G5, they will then be employed in the G4's design. But the the G5 will still have a new and improved architecture.
A key development is that Apple has turned to Motorola for virtually all of their CPUs. Imagine how much money is now flowing into Motorola from their desktop PPC sales! Only a year ago, Apple was buying only G4s for their Powermac lineup from Motorola, now Apple is using Motorola CPU's in all of their Macs except the iBook (which will soon get a G4).
Would Apple give Motorola so much business if they had dropped the ball on the G5? If they had nothing promising for Apple? If they couldn't supply Apple with competitive CPUs? I think not. I believe what is happening is that Motorola are reaping the rewards for their recent commitment to PPC development, and that Apple will soon be buying ALL of their CPU from Motorola, including the G5.
It is an exciting time for Powermac development, and we are going to see the shift to a new PPC architecture and a new high-bandwidth mobo in the near future. Apple is once again going to own the professional high performance hardware market...the new powermacs will be the envy of the Wintel world. Hold onto your butts 'cause it's going to be HUGE!!!
<strong>A key development is that Apple has turned to Motorola for virtually all of their CPUs. Imagine how much money is now flowing into Motorola from their desktop PPC sales! Only a year ago, Apple was buying only G4s for their Powermac lineup from Motorola, now Apple is using Motorola CPU's in all of their Macs except the iBook (which will soon get a G4).</strong><hr></blockquote>
what so many seem to forget is that the embedded market presents much greater sales numbers than the desktop market. compared to many other parts of their business, motorola makes very little money selling chips to apple. there are probably 30 times as many embedded chips in my car than their are in my computer.
Not only is this good news for Motorola, but perhaps more importantly, it demonstrates to Moto that there still is good money to be made in desktop PPC sales, and that there is the POTENTIAL for a successful G5 to bring in mondo sales for Motorola. Apple's commitment to the G4 shows Motorola what might happen if they deliver on their G5 rhetoric, and surely Motorola must recognize that if Apple's marketshare improves, they will directly benefit. Apple is dangling the carrot in front of Moto--all they need to do is give Apple a bitchin' CPU and then it's wide open for OS X's market share to skyrocket.
[ 05-04-2002: Message edited by: Junkyard Dawg ]</p>
<strong>True, but Motorola is now selling over twice as many G4 chips to Apple. While this may not be a great number in comparison to moto's embedded chip sales, it's still a significant increase in desktop PPC sales that have a positive impact on Moto's commitment to the G5.
Not only is this good news for Motorola, but perhaps more importantly, it demonstrates to Moto that there still is good money to be made in desktop PPC sales, and that there is the POTENTIAL for a successful G5 to bring in mondo sales for Motorola. Apple's commitment to the G4 shows Motorola what might happen if they deliver on their G5 rhetoric, and surely Motorola must recognize that if Apple's marketshare improves, they will directly benefit. Apple is dangling the carrot in front of Moto--all they need to do is give Apple a bitchin' CPU and then it's wide open for OS X's market share to skyrocket.
[ 05-04-2002: Message edited by: Junkyard Dawg ]</strong><hr></blockquote>
If I was running Moto than I would have to take a long hard look at our relationship with Apple. Not much sales as compared with the other chip sales. Yet Apple demands higher performance chips which pushes costs up, but Apple pays for R&D that can be transferred to other chip products. This is common sense and I write this to point out that there is a balance point. It appears to me from the outside looking in that IBM is a better fit right now. IBM already makes the Power4 which is high end, I'm not saying that Apple will use it, rather that the infrastructure to make high end chips in quantity is already there. IBM ups the capabilities of the infrastructure for themselves and for he other customers that come to them for custom chips. So why not use IBM?? Apple has a relationship with Moto and that is a business relationship. It appears to me that Moto would be better off in concentrating on developing that new communications chip technology that they have, and let IBM take over the Apple next chip development and production. Moto can still supply the G4 for the consumer market of Apple since they already are making them. IBM would take over development of the next chip G5 whatever that is. BookE said it all, "this will make it very easy for PPC development to be transferred between the the companies that are developing and manufacturing PPC chips.
That is what I see at least.
<strong>
If I was running Moto than I would have to take a long hard look at our relationship with Apple. Not much sales as compared with the other chip sales. Yet Apple demands higher performance chips which pushes costs up, but Apple pays for R&D that can be transferred to other chip products. This is common sense and I write this to point out that there is a balance point.</strong><hr></blockquote>
"Embedded" doesn't mean "wimpy." There are embedded applications (high-end routers, telephone switches) that crush any desktop in terms of raw computational power. The 7455 has won accolades in the embedded-chip press, and the likes of Cisco will buy the fastest chips Mot can crank out. So it's not just Apple pushing them; there's a high-end embedded market, and Motorola's moving into it.
<strong>
"Embedded" doesn't mean "wimpy." There are embedded applications (high-end routers, telephone switches) that crush any desktop in terms of raw computational power. The 7455 has won accolades in the embedded-chip press, and the likes of Cisco will buy the fastest chips Mot can crank out. So it's not just Apple pushing them; there's a high-end embedded market, and Motorola's moving into it.</strong><hr></blockquote>
Sorry I did not mean to imply embedded==wimpy. Good point. I agree there are very high end embedded chips. My point was that I think that Moto would be better off developing their new technology and still improve the 7xxx series. If Apple wants way beyond that, I guess I hope they do, than Moto would be better served by concentrating on what's on their own plate. If we are looking at two or three FPUs and DDR memory than Moto can handle that. I think that I am implying that Apple might be after something much more, as was eluded to earlier.
<strong>If I was running Moto than I would have to take a long hard look at our relationship with Apple. Not much sales as compared with the other chip sales. Yet Apple demands higher performance chips which pushes costs up, but Apple pays for R&D that can be transferred to other chip products. This is common sense and I write this to point out that there is a balance point. It appears to me from the outside looking in that IBM is a better fit right now. IBM already makes the Power4 which is high end, I'm not saying that Apple will use it, rather that the infrastructure to make high end chips in quantity is already there. IBM ups the capabilities of the infrastructure for themselves and for he other customers that come to them for custom chips. So why not use IBM?? Apple has a relationship with Moto and that is a business relationship. It appears to me that Moto would be better off in concentrating on developing that new communications chip technology that they have, and let IBM take over the Apple next chip development and production. Moto can still supply the G4 for the consumer market of Apple since they already are making them. IBM would take over development of the next chip G5 whatever that is. BookE said it all, "this will make it very easy for PPC development to be transferred between the the companies that are developing and manufacturing PPC chips.
That is what I see at least.</strong><hr></blockquote>
exactly.
[quote]<strong>
True, but Motorola is now selling over twice as many G4 chips to Apple. While this may not be a great number in comparison to moto's embedded chip sales, it's still a significant increase in desktop PPC sales that have a positive impact on Moto's commitment to the G5.
Not only is this good news for Motorola, but perhaps more importantly, it demonstrates to Moto that there still is good money to be made in desktop PPC sales, and that there is the POTENTIAL for a successful G5 to bring in mondo sales for Motorola. Apple's commitment to the G4 shows Motorola what might happen if they deliver on their G5 rhetoric, and surely Motorola must recognize that if Apple's marketshare improves, they will directly benefit. Apple is dangling the carrot in front of Moto--all they need to do is give Apple a bitchin' CPU and then it's wide open for OS X's market share to skyrocket.</strong><hr></blockquote>
well, twice as much of very little is still not much. add to the low sales the fact that apple is a very demanding company to work with and i can see how the relationship could have soured.
let's say you make $100k a year performing a service. i'm one of your customers and i pay you $2k a year, but take up as much time and resources as one of your $30k clients. me paying $4k the next year only makes me somewhat more tolerable, but still not a good or significant enough customer to cause you so much grief.
There is only one company now able to produce efficiently on SOI 0,13 it's IBM. AMD is not supposed to produce SOI 0,13 until the end of the year, and for the moment does not produc 0,13 micron process. Intel fab on 0,13 process but not on SOI (or equivalent).
Mot will continue to improve the G4 line (perhaps using some G5 tweaks) until he can reach o,13 SOI. When Mot will reach this limit, the G5 will appear. The architecture is ready but not the fab process. Contrary to the past where Mot was stuck at 500 mhz, Mot will not wait the new fab process without do any improvement on his chip line. We have the 7455 who add just a new gestion of the L3 cache, we will have soon the 7460 that will have DDR memory controller, but the same amount of L2 cache if still make on a 0,18 micron process. I think that the 512 L2 cache will wait until 0,13 , like the Intel chips.
So i make this statement : 7460 with ddr memory controller for MWNY. G5 in 2003 perhaps in january if Mot is not too late in the developpement of 0,13 SOI.
I am not a chip engineer, but iam ready to bet it's more simple to make an new chip architecture than to develop a new fab process.
<strong>
I am not a chip engineer, but iam ready to bet it's more simple to make an new chip architecture than to develop a new fab process.</strong><hr></blockquote>
Could you please elaborate.
If this relates to the IBM stuff, BookE was mentioned. And IBM has manufactured the G4 for Apple when supplies from Moto were low.
<strong>
Could you please elaborate.
If this relates to the IBM stuff, BookE was mentioned. And IBM has manufactured the G4 for Apple when supplies from Moto were low.</strong><hr></blockquote>
I'll will say that the amount of R&D necessary for developping new fab process is more important than the amount of R&D needed for chip design. Motorola alone is not anymore rich enough to develop new fab process : that's why with Philips and other he created a new research center in France. Many companies are able to design chips, like nvidia, but very few companies are able to develop new fab process. You certainly know the Moore's law about the number of transistor included in a chip, there is an another law, fab process are more and more expansive in an exponantial way either.
Continued stable sales and profits for Apple Computer. The company
[ 05-05-2002: Message edited by: fellow722 ]</p>
<strong>
I'll will say that the amount of R&D necessary for developping new fab process is more important than the amount of R&D needed for chip design. Motorola alone is not anymore rich enough to develop new fab process : that's why with Philips and other he created a new research center in France. Many companies are able to design chips, like nvidia, but very few companies are able to develop new fab process. You certainly know the Moore's law about the number of transistor included in a chip, there is an another law, fab process are more and more expansive in an exponantial way either.</strong><hr></blockquote>
Making Other Companies Chips is something that IBM does, they are just a hired gun, or manufacturer if you will. This is nothing new. Am I missing something? If you wanted to develop the PowerDoc G5 chip, and you had the money, IBM would make it for you, in quantity. It would help allot and probably be cheaper if the design were BookE compliant. This is where Apple could be. They maybe have designed a chip based on BookE, or have asked IBM to develop a chip for them based on BookE. As far as capacity is concerned IBM may have more than enough capacity. And IBM has leading technology that is shipping. Copper, SOI, Low k dielectric. I don't think that IBM would have to add capacity for Apple, is that what you are saying? I mean to be conservative IBM would need at the outside 4 Million chips to more than cover Apples orders, that may sound like allot, but I venture that 4mil is a drop in the bucket.
<strong>
Making Other Companies Chips is something that IBM does, they are just a hired gun, or manufacturer if you will. This is nothing new. Am I missing something? If you wanted to develop the PowerDoc G5 chip, and you had the money, IBM would make it for you, in quantity. It would help allot and probably be cheaper if the design were BookE compliant. This is where Apple could be. They maybe have designed a chip based on BookE, or have asked IBM to develop a chip for them based on BookE. As far as capacity is concerned IBM may have more than enough capacity. And IBM has leading technology that is shipping. Copper, SOI, Low k dielectric. I don't think that IBM would have to add capacity for Apple, is that what you are saying? I mean to be conservative IBM would need at the outside 4 Million chips to more than cover Apples orders, that may sound like allot, but I venture that 4mil is a drop in the bucket.</strong><hr></blockquote>
I agree with what you say , but i was speakind of Mot and not IBM. Mot own the altivec licenses. Without Altivec the PPC chips are very crappy and will loose all the benchmarks against the X86 chips. Perhaps Mot want to keep one of this biggest customer : Apple and wont let him going elsewhere. It's possible that Mot does not wanted to license altivec to anyone else to protect him.
Without altivec , i am ready to bet that all the PPC Apple chips will be fabbed by IBM.
<strong>
I agree with what you say , but i was speakind of Mot and not IBM. Mot own the altivec licenses. Without Altivec the PPC chips are very crappy and will loose all the benchmarks against the X86 chips. Perhaps Mot want to keep one of this biggest customer : Apple and wont let him going elsewhere. It's possible that Mot does not wanted to license altivec to anyone else to protect him.
Without altivec , i am ready to bet that all the PPC Apple chips will be fabbed by IBM.</strong><hr></blockquote>
Ah Ha...Altivec, I see, I agree. It does seem that Apple would be interested in Altivec, but what if instead of buying Altivec Apple licenses it from Moto. That way Moto can keep making chips for Cisco, and Apple can then still have it included in the chips made for them. I think that this would be good for both. And would be much cheaper than trying to buy the entire PPC license from Moto, which I always thought was unrealistic. And I think that Apple had some hand, probably small, in developing the initial version of altivec. If i remember correctly Apple had more to do with the latest version of Altivec. There was a time, maybe a year ago maybe a little longer that Moto announced a slightly improved version of Altivec. I think that Apple had more than a passive role in bring that to market. So co-licensing Altivec could be done, and would still serve both companies. I think that Apple has the talent on board to develop their own next version of altivec.
A recent development box has allowed us to open and examine some details that appeal to us as engineers and hardware designers. One of the machines is built in a G4 case with a smaller motherboard. the interesting part is that the motherboard is a cheaper 4 layer variety while the CPU daughter card is made on a more expensive 8 layer board. the reason for this is very simple and I will get to that in a moment. The CPU card itself is similar in shape and pin connector to the one used in modern G4s but is a bit larger in surface area. The reason for this is the fact that the G5 CPU has a built-in memory controller that interacts with high-speed DDR-SDRAM. This RAM runs at 133MHz DDR but the G5 documentation we have (preliminary) states that 166MHz DDR (333MHz effective) is supported. To run at these fast speeds you must utilize 2 dimensional PCB space efficiently. Apple has done this well. On the CPU daughter card you will find 3 slots for DDR-SDRAM DIMMS. They are distanced from the actual CPU core to allow for the heatsink placement, but in this configuration (since on the card the CPU is hard soldered to the PCB) you can design DIMM slots that can operate at 133, 166, and quite possibly 200MHz at DDR speeds. In fact, I believe the impedence should be low enough to allow for 4 DIMM slots. No L3 RAM is located on the card although it is fully supported under the documentation.
RapidIO is used to connect the CPU to a centralized controller that controls peripherals such as gigabit ethernet, Firewire, USB, ATA interface, audio input/output, AGP/PCI bridge, etc. The RapidIO interface operates in a 16 bit full duplex mode running at 500MHz. This bus is seperate from the memory bus and therefore asyncronous. It does not share the bus either like with the G4 and MPX. PCI-X seems to be implimented fully (it's backwards compatible with PCI 32/64) and AGP remains operating at 4X speeds.
A whole new paradigm in PowerMac design will be introduced soon and it will be a modern example in smart computer design, taking advantage of new trechnologies and foward thinking design. These are late beta machines and the finishing touches are being applied to cut out the rough edges for consumer consumption. <hr></blockquote>
Hmmm...Maybe we should discuss this again...