Apple will not use a AMD chip. Sorry won't happen. I'm sure apple and AMD have talked about business plans since they are in the same position against Intel. However, the IBM PPC970 will be the next chip for apple. By 2004 the G4 will even be replaced in the consumer lineup by the PPC7457-RM. (Ok so it's still a G4 but not the same one we have now.)
Fact:
\tBy second quarter 2003 the 7457 will be out. It will have SOI and a .13u process with a top speed of 1.8Ghz. In another 9 months or so we will see the PPC7457-RM which will have a DDRBus and exceed 2Ghz up to around 2.5Ghz. A year form now we will be at around 2.2Ghz have a DDRbus and SOI. Thats all I know for sure. I know nothing more than any of you on the PPC970 time frame. I would suspect that it will come out in MWSF with a 3 month delay. You can order now but don't expect to get them until march thing. From what I have heard the 970 is going to be quite scalable and will start out with a top speed of 1.8Ghz (but you all know that). By the time the PPC7457-RM is at 2.5Ghz within 16months the PPC970 will be over 3GHz. Remember IBM plans to drop the fab to .9U which should boost the speed by at least 700mhz easy just by doing that.
\tThe AMD rumors sound like a bunch of BS to me. Apple has everything going their way next year. They really don't need AMD. If any one wants proof of my time table I can email you Motorola's own Overview of the PPC department and time tables for the G4 and embedded chips.
OK, I agree with programmer that there will absolutely not be any change in architecture for the macintosh platform. While he may come from well understood technical principles, that's all mumbo-jumbo to me. However, IA32 is nearing the end of it's life and the replacement is anything but certain. Will it be IA64 or X86-64 that becomes the defacto standard? Without M$ support either one of these platforms is dead. Apple cannot get left with the loser if the whole point of platform change is to take advantage of mainstream CPU dev/competition in order to ensure an equal footing with windows platforms in the future. Even if you hate the PPC, it's not worth the effort to switch untill this situation has been resolved on the wintelon side. Sorry.
Also, how hot/hungry can CPU's get before they stop getting bigger and faster? Mebbe IBM is a little ahead in concept, but I really think a brace of small cool CPU's is the way of the future (as do some more knowledgeable types than myself). For that matter a G4 could also be such a chip, if it had the FSB to make such an arrangement worthwhile. PPC970 looks set to solve the bandwidth issue in a resounding way, and the future of PPC looks good should MP configs be the way of the future.
Now, all those good reasons why we won't see an AMD based mac aside, I remember, must be almost two years now, that some company claimed to be able to emulate the PPC very effectively using an Athlon chip. The claim was they could deliver 1Ghz g4 performance from a then king of the hill, Athlon 1.2-3, or something like that. Musta been vapour-ware, but the story was on some of the bigger internet tech outlets. Something to think about?
There's no need for another chip 'until there is PPC970'. The G4 will do good enough. Actually, the Motorola roadmap shows that the PPC 7457-RM will be out at the same time as the PPC970 (late 2003) and outperform it without having to port OS and apps to 64bit to take advantage of a new processor.
My guess is that Apple won't leave the G4 in 2003. IBM's processor looks good. Today. But when it finally appears at the end of 2003, it'll have to compare well against what Motorola, Intel and AMD have to offer by THEN.
There's no need for another chip 'until there is PPC970'. The G4 will do good enough. Actually, the Motorola roadmap shows that the PPC 7457-RM will be out at the same time as the PPC970 (late 2003) and outperform it without having to port OS and apps to 64bit to take advantage of a new processor.
My guess is that Apple won't leave the G4 in 2003. IBM's processor looks good. Today. But when it finally appears at the end of 2003, it'll have to compare well against what Motorola, Intel and AMD have to offer by THEN.</strong><hr></blockquote>
That was my point. Why would apple switch to AMD when they don't need to? And remember AMD stock is only at $6.86. Lower than Motorola and not the most secure company to place your future in. IBM on the other hand has a stock price of $84.43 and would be a great backer for the PPC. As I posted earlier Motorola has everything under control for next year. And these G4s are going to be fast. A 7 stage pipe line running at over 2Ghz. Man the P4 has a 20 stage just to do that. If AMD ends of making chips for apple I'll buy all of you a six pack. How about that? lol
I'm curious how you guys are so sure about the 7457 / 7457-RM? There have been no official annoucements, and the one Moto document I have seen about it is relatively old and vague. I do expect a 0.13 G4 w/ on-chip memory controller and RapidIO bus eventually, but we don't really have any solid information about how it will perform or when we'll see it. On the other hand we know quite a bit about the 970 (and some of us might know quite a bit more since IBM is clearly willing to talk about it at least a little) since it is POWER4-derived.
My best guess, considering the lack of 7457-RM info, is that the 970 in 32-bit mode will outperform the 7457-RM by a wide margin at the same clock speed (especially when code is recompiled for it). It has way more internal resources and execution units, and its memory interface will be 20% faster than the 7457-RM's on-chip memory controller -- in one direction, ignoring the same amount of bandwidth in the other direction at the same time. And that assumes full DDR333 throughput which is more than your claims for the 7457-RM, nevermind the 7457. The 970 will also be much more efficient in an SMP configuration due to the full MERSI implementation and the fast bus.
The portable and consumer level Macs will continue to use some G4 variant at least until the 970 hits 0.09 micron. At that point the 970 should be low-power enough for and whether it is economical for the low end will depend on how expensive & hot the bus and chipset are. I'm not holding my breath for anything out of Moto, and I'll believe it when I see it.
There's no need for another chip 'until there is PPC970'. The G4 will do good enough. Actually, the Motorola roadmap shows that the PPC 7457-RM will be out at the same time as the PPC970 (late 2003) and outperform it without having to port OS and apps to 64bit to take advantage of a new processor. <hr></blockquote>
At the risk of being slightly 'off-topic', how can anyone here believe Apple's processor malaise is near a competitive solution? Since the G4 announcement debacle, we've periodically been expecting great leaps in performance -- but always 6 months (or more) in the future. Well, the future never arrives! I'm from Missouri, please "SHOW-ME".
At the risk of being slightly 'off-topic', how can anyone here believe Apple's processor malaise is near a competitive solution? Since the G4 announcement debacle, we've periodically been expecting great leaps in performance -- but always 6 months (or more) in the future. Well, the future never arrives! I'm from Missouri, please "SHOW-ME".</strong><hr></blockquote>
Because those in the know indicate there are much faster boxes running under the desks of Apple employees now. Right now. And have been for some time. Whether it's yields, or some other piece in the puzzle is unknown.
Now that Apple has effectively done an Osbourne on the PowerMacs, either they're:
a) happy that they'll never sell another PowerMac until mid next year.
b) Things are much further along than we know
c) Hope to introduce a whole whack of diversionary iPod-esque appliances come January.
<strong>At the risk of being slightly 'off-topic', how can anyone here believe Apple's processor malaise is near a competitive solution? Since the G4 announcement debacle, we've periodically been expecting great leaps in performance -- but always 6 months (or more) in the future. Well, the future never arrives! I'm from Missouri, please "SHOW-ME".</strong><hr></blockquote>
This time is different -- we have solid info on the arrival time of the next processor. Normally we are completely in the dark.
And while some forward looking types might defer purchases until the expected introduction of the next machine, many people need a machine now and won't wait. Especially since Apple hasn't officially announced a next generation machine.
<strong>Because those in the know indicate there are much faster boxes running under the desks of Apple employees now. Right now. And have been for some time.</strong><hr></blockquote>
Maybe simply all of them are too egoistic to tell upper managment and want to have all the goodies for themselves?
Seriously though, where are the rumored 2.4 ghz moto chips now? Could it be that they come in January and the 970 is actually a low end variant?
Those of you disposed to seriously contemplating AMD rumours are clearly in withdrawal from the days when we only had the bare minimum of forward-looking information coming out of A, I & M. But times have changed with the announcement of the PPC970. Listen to Programmer if you want to be sensible and not get caught up in strange geeky fantasies that will never come true.
So: no-one's listening to Programmer, right?
For the rest of us, who badly need a fix of quixotic geek fantasies, I'll revive an old idea: AMD cores + Transmeta translation layers. That is, perhaps AMD might match their RISC-ish
core to an as yet secret Transmeta PPC -> AMD translation layer, suddenly yielding a new entrant to the PPC market, and providing some much-needed cashflow to these two struggling companies. These AMD/Transmeta boxen might have a unique selling point: by activating either the PPC or x86 translation layer at startup, the one box might be used as either a Mac or a Windoze box, depending on the user's needs at the time. The PPC615 returns!
I hasten to add that I don't believe this scenario for a moment, but it is a fun idea to kick around.
<strong>Now, all those good reasons why we won't see an AMD based mac aside, I remember, must be almost two years now, that some company claimed to be able to emulate the PPC very effectively using an Athlon chip. The claim was they could deliver 1Ghz g4 performance from a then king of the hill, Athlon 1.2-3, or something like that. Musta been vapour-ware, but the story was on some of the bigger internet tech outlets. Something to think about?</strong><hr></blockquote>
The emulation was unbelievably slow and was only a proof of concept. It didn't achieve speeds even remotely close to an effective 1GHz G4.
[QB]I'm curious how you guys are so sure about the 7457 / 7457-RM? There have been no official annoucements, and the one Moto document I have seen about it is relatively old and vague.[QB]
-----
I have some very clear tech/time charts of the G4. Although they are a little less clear about the 7457-RM.
\tI still think that the PPC970 will replace the G4 in the PowerMac lines next year. The 7457 will be for the consumer lineup. This way apple's consumer lineup doesn't fall behind their professional lineup.
At the risk of typing something that everyone might know, wasn't there some kind of an agreement about a year or two ago between Motorola and AMD (Moto to contribute some RISC design expertise and AMD some fab expertise)? If true, could this be at play with regard to all of these AMD rumors?
<strong>At the risk of typing something that everyone might know, wasn't there some kind of an agreement about a year or two ago between Motorola and AMD (Moto to contribute some RISC design expertise and AMD some fab expertise)? If true, could this be at play with regard to all of these AMD rumors?</strong><hr></blockquote>
According to Eskimo, AMD got a good luck at who they were partnering with and backed out. So no, I'm not expecting anything from that partnership.
<strong>The 970 will also be much more efficient in an SMP configuration due to the full MERSI implementation and the fast bus.</strong><hr></blockquote>
I've seen a couple of references for quotes like 'good at SMP up to 8-way' etc. But I haven't found a comment about how much of MERSI the 970 implements. Would a reference be possible? Or at least a confirmation?
[quote]Originally posted by bandalay:
<strong>Because those in the know indicate there are much faster boxes running under the desks of Apple employees now. Right now. </strong><hr></blockquote>
I thought those pointed towards RS6000's running Mac OS X. (or at least Darwin) Making sure things work on a Power3 or Power4 (which would be faster but out of Apple's price line to ship) doesn't necessarily indicate that those 'faster boxes' are ppc970 based. Yet.
<strong>I've seen a couple of references for quotes like 'good at SMP up to 8-way' etc. But I haven't found a comment about how much of MERSI the 970 implements. Would a reference be possible? Or at least a confirmation?
</strong><hr></blockquote>
Sorry, I can confirm MERSI but I can't compromise my sources.
Comments
Fact:
\tBy second quarter 2003 the 7457 will be out. It will have SOI and a .13u process with a top speed of 1.8Ghz. In another 9 months or so we will see the PPC7457-RM which will have a DDRBus and exceed 2Ghz up to around 2.5Ghz. A year form now we will be at around 2.2Ghz have a DDRbus and SOI. Thats all I know for sure. I know nothing more than any of you on the PPC970 time frame. I would suspect that it will come out in MWSF with a 3 month delay. You can order now but don't expect to get them until march thing. From what I have heard the 970 is going to be quite scalable and will start out with a top speed of 1.8Ghz (but you all know that). By the time the PPC7457-RM is at 2.5Ghz within 16months the PPC970 will be over 3GHz. Remember IBM plans to drop the fab to .9U which should boost the speed by at least 700mhz easy just by doing that.
\tThe AMD rumors sound like a bunch of BS to me. Apple has everything going their way next year. They really don't need AMD. If any one wants proof of my time table I can email you Motorola's own Overview of the PPC department and time tables for the G4 and embedded chips.
2Q-03: .13U PPC7457, 1Ghz-1.8Ghz, SOI, 166/200Mhz bus
1Q-04: .9U PPC7457-RM: 1.8Ghz-2Ghz+, SOI, DDRBus ?Mhz
These ships will use less power and produce less heat than the current models.
Also, how hot/hungry can CPU's get before they stop getting bigger and faster? Mebbe IBM is a little ahead in concept, but I really think a brace of small cool CPU's is the way of the future (as do some more knowledgeable types than myself). For that matter a G4 could also be such a chip, if it had the FSB to make such an arrangement worthwhile. PPC970 looks set to solve the bandwidth issue in a resounding way, and the future of PPC looks good should MP configs be the way of the future.
Now, all those good reasons why we won't see an AMD based mac aside, I remember, must be almost two years now, that some company claimed to be able to emulate the PPC very effectively using an Athlon chip. The claim was they could deliver 1Ghz g4 performance from a then king of the hill, Athlon 1.2-3, or something like that. Musta been vapour-ware, but the story was on some of the bigger internet tech outlets. Something to think about?
There's no need for another chip 'until there is PPC970'. The G4 will do good enough. Actually, the Motorola roadmap shows that the PPC 7457-RM will be out at the same time as the PPC970 (late 2003) and outperform it without having to port OS and apps to 64bit to take advantage of a new processor.
My guess is that Apple won't leave the G4 in 2003. IBM's processor looks good. Today. But when it finally appears at the end of 2003, it'll have to compare well against what Motorola, Intel and AMD have to offer by THEN.
<strong>Something's eating all of you...
There's no need for another chip 'until there is PPC970'. The G4 will do good enough. Actually, the Motorola roadmap shows that the PPC 7457-RM will be out at the same time as the PPC970 (late 2003) and outperform it without having to port OS and apps to 64bit to take advantage of a new processor.
My guess is that Apple won't leave the G4 in 2003. IBM's processor looks good. Today. But when it finally appears at the end of 2003, it'll have to compare well against what Motorola, Intel and AMD have to offer by THEN.</strong><hr></blockquote>
That was my point. Why would apple switch to AMD when they don't need to? And remember AMD stock is only at $6.86. Lower than Motorola and not the most secure company to place your future in. IBM on the other hand has a stock price of $84.43 and would be a great backer for the PPC. As I posted earlier Motorola has everything under control for next year. And these G4s are going to be fast. A 7 stage pipe line running at over 2Ghz. Man the P4 has a 20 stage just to do that. If AMD ends of making chips for apple I'll buy all of you a six pack. How about that? lol
My best guess, considering the lack of 7457-RM info, is that the 970 in 32-bit mode will outperform the 7457-RM by a wide margin at the same clock speed (especially when code is recompiled for it). It has way more internal resources and execution units, and its memory interface will be 20% faster than the 7457-RM's on-chip memory controller -- in one direction, ignoring the same amount of bandwidth in the other direction at the same time. And that assumes full DDR333 throughput which is more than your claims for the 7457-RM, nevermind the 7457. The 970 will also be much more efficient in an SMP configuration due to the full MERSI implementation and the fast bus.
The portable and consumer level Macs will continue to use some G4 variant at least until the 970 hits 0.09 micron. At that point the 970 should be low-power enough for and whether it is economical for the low end will depend on how expensive & hot the bus and chipset are. I'm not holding my breath for anything out of Moto, and I'll believe it when I see it.
Originally posted by fryke:
Something's eating all of you...
There's no need for another chip 'until there is PPC970'. The G4 will do good enough. Actually, the Motorola roadmap shows that the PPC 7457-RM will be out at the same time as the PPC970 (late 2003) and outperform it without having to port OS and apps to 64bit to take advantage of a new processor. <hr></blockquote>
At the risk of being slightly 'off-topic', how can anyone here believe Apple's processor malaise is near a competitive solution? Since the G4 announcement debacle, we've periodically been expecting great leaps in performance -- but always 6 months (or more) in the future. Well, the future never arrives! I'm from Missouri, please "SHOW-ME".
<strong>
At the risk of being slightly 'off-topic', how can anyone here believe Apple's processor malaise is near a competitive solution? Since the G4 announcement debacle, we've periodically been expecting great leaps in performance -- but always 6 months (or more) in the future. Well, the future never arrives! I'm from Missouri, please "SHOW-ME".</strong><hr></blockquote>
Because those in the know indicate there are much faster boxes running under the desks of Apple employees now. Right now. And have been for some time. Whether it's yields, or some other piece in the puzzle is unknown.
Now that Apple has effectively done an Osbourne on the PowerMacs, either they're:
a) happy that they'll never sell another PowerMac until mid next year.
b) Things are much further along than we know
c) Hope to introduce a whole whack of diversionary iPod-esque appliances come January.
<strong>At the risk of being slightly 'off-topic', how can anyone here believe Apple's processor malaise is near a competitive solution? Since the G4 announcement debacle, we've periodically been expecting great leaps in performance -- but always 6 months (or more) in the future. Well, the future never arrives! I'm from Missouri, please "SHOW-ME".</strong><hr></blockquote>
This time is different -- we have solid info on the arrival time of the next processor. Normally we are completely in the dark.
And while some forward looking types might defer purchases until the expected introduction of the next machine, many people need a machine now and won't wait. Especially since Apple hasn't officially announced a next generation machine.
<strong>Because those in the know indicate there are much faster boxes running under the desks of Apple employees now. Right now. And have been for some time.</strong><hr></blockquote>
Maybe simply all of them are too egoistic to tell upper managment and want to have all the goodies for themselves?
Seriously though, where are the rumored 2.4 ghz moto chips now? Could it be that they come in January and the 970 is actually a low end variant?
So: no-one's listening to Programmer, right?
For the rest of us, who badly need a fix of quixotic geek fantasies, I'll revive an old idea: AMD cores + Transmeta translation layers. That is, perhaps AMD might match their RISC-ish
core to an as yet secret Transmeta PPC -> AMD translation layer, suddenly yielding a new entrant to the PPC market, and providing some much-needed cashflow to these two struggling companies. These AMD/Transmeta boxen might have a unique selling point: by activating either the PPC or x86 translation layer at startup, the one box might be used as either a Mac or a Windoze box, depending on the user's needs at the time. The PPC615 returns!
I hasten to add that I don't believe this scenario for a moment, but it is a fun idea to kick around.
<strong>Now, all those good reasons why we won't see an AMD based mac aside, I remember, must be almost two years now, that some company claimed to be able to emulate the PPC very effectively using an Athlon chip. The claim was they could deliver 1Ghz g4 performance from a then king of the hill, Athlon 1.2-3, or something like that. Musta been vapour-ware, but the story was on some of the bigger internet tech outlets. Something to think about?</strong><hr></blockquote>
The emulation was unbelievably slow and was only a proof of concept. It didn't achieve speeds even remotely close to an effective 1GHz G4.
[QB]I'm curious how you guys are so sure about the 7457 / 7457-RM? There have been no official annoucements, and the one Moto document I have seen about it is relatively old and vague.[QB]
-----
I have some very clear tech/time charts of the G4. Although they are a little less clear about the 7457-RM.
\tI still think that the PPC970 will replace the G4 in the PowerMac lines next year. The 7457 will be for the consumer lineup. This way apple's consumer lineup doesn't fall behind their professional lineup.
<strong>At the risk of typing something that everyone might know, wasn't there some kind of an agreement about a year or two ago between Motorola and AMD (Moto to contribute some RISC design expertise and AMD some fab expertise)? If true, could this be at play with regard to all of these AMD rumors?</strong><hr></blockquote>
According to Eskimo, AMD got a good luck at who they were partnering with and backed out. So no, I'm not expecting anything from that partnership.
<strong>The 970 will also be much more efficient in an SMP configuration due to the full MERSI implementation and the fast bus.</strong><hr></blockquote>
I've seen a couple of references for quotes like 'good at SMP up to 8-way' etc. But I haven't found a comment about how much of MERSI the 970 implements. Would a reference be possible? Or at least a confirmation?
[quote]Originally posted by bandalay:
<strong>Because those in the know indicate there are much faster boxes running under the desks of Apple employees now. Right now. </strong><hr></blockquote>
I thought those pointed towards RS6000's running Mac OS X. (or at least Darwin) Making sure things work on a Power3 or Power4 (which would be faster but out of Apple's price line to ship) doesn't necessarily indicate that those 'faster boxes' are ppc970 based. Yet.
<strong>I've seen a couple of references for quotes like 'good at SMP up to 8-way' etc. But I haven't found a comment about how much of MERSI the 970 implements. Would a reference be possible? Or at least a confirmation?
</strong><hr></blockquote>
Sorry, I can confirm MERSI but I can't compromise my sources.
Heh, I've always wanted to say that.