I re-read this entire thread, and am still struck by the similarities to the threads to which I linked above. I don't know what to think, but I am not sure that hope alone will get me through this one. Tomorrow will tell, I suppose.
Am I the only one here who thinks we will see absolutely NOTHING tomorrow? Except maybe shipment confirmation of the new iMacs...
I think G5s will be here very soon, and it is entirely possible for Apple to have some low cost G4 towers. They're still offering old school CRT iMacs for dirt cheap aren't they?
I just doesn't seem reasonable that Apple would announce a speedbumped G4 tomorrow and then the G5 in the next 2-3 months. They've been pretty strict about the product rev times over the last few years, and that timeframe seems a little quick. I could imagine it at MWNY, but if they announce a faster G4 tomorrow, I'll be buying it, just because I need one.
The only problem with this is that Apple has been moving away from "scalabel archatecture". They would rather you buy a new G4 or G5 than just add a processor, which Apple does not sell as add ons. And based on Apples past performance on upgrade pricing (such as on RAM), they could not compete in the upgrade card market.</strong><hr></blockquote>
Currently Apple doesn't offer anything in the truely high end (nothing of particular note from last look at least) so they haven't moved away or towards it. They just don't work in it.
They have moved away from upgradeable so to speak in their PowerMacs but they are very different machines compared to what SGI, Sun et al produce and that is different to scaleable.
It's an area that has good sized margins where Apple could definitely compete (although I doubt they want to compete directly with Sun) and should considering some of their markets.
<strong>Am I the only one here who thinks we will see absolutely NOTHING tomorrow? Except maybe shipment confirmation of the new iMacs... </strong><hr></blockquote>
That would suck.
Aren't there excellent education deals on the towers? Doesn't that seem like a good channel to dump inventory into?
Is anyone actually buying the towers at regular price at this point?
Well, back in June or July I think, Apple employees got HUGE discounts on the top of the line TiBook (then 500 MHz). I believe they got them for about $2500 or $2700. That was in July... the TiBooks didn't get updated until October.
the G5 is likely >50 million transistors and on the same process as the Apollo... which virtually guarantees lower yields (and thus higher prices).
Actually the Apollo is said to be made on a 180nm process with SOI (by a Motorola employee in November). I too thought it would be 130nm and SOI, but unless he was mistaken 9and he hasn't bothered fixing it) then it is in fact 180nm. That means only the benefits of SOI would be in effect. That means about 25-30%. That's a confortable increase to 1.0-1.13GHz. The G5 is in fact a 130nm SOI process and should therefore have more chips per wafer than if it was made on 180nm.
the G5 is likely >50 million transistors and on the same process as the Apollo... which virtually guarantees lower yields (and thus higher prices).
Actually the Apollo is said to be made on a 180nm process with SOI (by a Motorola employee in November). I too thought it would be 130nm and SOI, but unless he was mistaken 9and he hasn't bothered fixing it) then it is in fact 180nm. That means only the benefits of SOI would be in effect. That means about 25-30%. That's a confortable increase to 1.0-1.13GHz. The G5 is in fact a 130nm SOI process and should therefore have more chips per wafer than if it was made on 180nm.</strong><hr></blockquote>
Oh, I didn't know that the G5 had in fact been announced and details of its process published. It is fairly normal practice for a design to appear on a non-bleeding edge process and then migrate to the latest once it is established. That's not to say I wouldn't like to see a 130nm SOI G5 right away.
<strong>one more day... </strong><hr></blockquote>
Well it's now 3:45 EST. Looks like Philbot is either a fake or has uncredible sources. Ohwell, this just lends hope that Apple wants to annouce the new PM's with fanfair (read:significant announcement).
That's not to say that Apple will even use a G5 from Motorola's roadmap.(disclaimer: I don't know that they will or won't.)</strong><hr></blockquote>
The data in the roadmap is notoriously "fluid". 2+ GHz became 800MHz - 2GHz, for example. I'm sure the G5 will end up on the stated process, or an even better one, but it wouldn't surprise me to see it appear on a slightly older process initially.
[quote] I am also curious as to your thoughts on the two threads to which I linked above, Mr. Philbot. <hr></blockquote>
you don't mean the illusive Kormac surely?
Apple needs to announce the G5 to keep all the pro users and Tech media happy. No one is buying the G4 towers, but at least if we know what's coming (and for how much (£) we can make our buying decisions.
Come on Steve! you know it makes sense! lovely jubbly!!
[quote] If it is even close to the rumoured performance levels then they aren't going to be able to build them fast enough. This means they can choose to jack up the price until the supply situation gets under control. Even if they don't do that the G5 is likely >50 million transistors and on the same process as the Apollo... which virtually guarantees lower yields (and thus higher prices).
<hr></blockquote>
Just because it has more transistors doesn't mean a thing. If more transistors = more expensive, then CPU prices would have been skyrocketing all throughout the 90s.
Like you said, it depends on how efficiently Moto can fab the G5. But you assume first that Moto will not be able to fab it with good efficiency. AFAIK, moto will work on the fab until they have an efficiency that allows them to sell a CPU at a desired price. Apple could have released G5s 6 months ago, but they would have cost $10000 per CPU.
Remember, that the G4 originally was fabbed with low efficiency, and it went into powermacs costing about the same as the current ones. The G4 has dropped in price quite a bit, now that it's being used in iMacs, and Powermac margins have probably risen.
I doubt the G5 will cost much more than the G4 did initially. Apple will be able to keep Powermac prices about where they are now...alternatively, Apple could do something better:
Keep the G4 Powermacs, albeit with a speed bump. Price them lower than they are now, beginning at, say, $1200 or so.
Add a G5 powermac lineup, and begin pricing at $2000. The G5 powermacs will be the "pro" desktop towers, and the G4 Powermacs will be the "consumer" towers.
I have made up my mind. The G5 will not be released this week but at seybold in february. I am betting on a huge speed increase. Up to 1.8Ghz. The top end will be 200$ cheaper. I doubt there will be any announcement of them this week. I believe Philbot is making this all up just to get a long thread. I have spoken, let it come to pass.
How do I know. Well lets just say its the most obvious choice of when the G5 will be released. And so on. It's the only time that works...
The only problem with this is that Apple has been moving away from "scalabel archatecture".
</strong>
<hr></blockquote>
What do you mean "moving away from?" They've never had one.
They've been moving toward a more modular architecture (as opposed to designing every motherboard from scratch), which is also a step toward an SGI-style architecture.
I doubt they'd adopt SGI's tech, though, for the simple reason that it's not cost effective until the costs get into the tens of thousands of dollars.
[quote]Originally posted by Junkyard Dawg:
<strong>
Just because it has more transistors doesn't mean a thing. If more transistors = more expensive, then CPU prices would have been skyrocketing all throughout the 90s.</strong><hr></blockquote>
You missed Programmer's point. He said that if the G5 was fabbed on the same process as the Apollo, it would cost more. And it would, because for a given process more transistors require more silicon, which means fewer chips per wafer, which raises the cost.
The solution is to fab bigger chips on smaller processes, and that's what Programmer was worrying about: Whether the G5 would start on a smaller process than the Apollo as planned.
Comments
Mandricard
AppleOutsider
[ 01-22-2002: Message edited by: Mandricard ]</p>
I think G5s will be here very soon, and it is entirely possible for Apple to have some low cost G4 towers. They're still offering old school CRT iMacs for dirt cheap aren't they?
<strong>
The only problem with this is that Apple has been moving away from "scalabel archatecture". They would rather you buy a new G4 or G5 than just add a processor, which Apple does not sell as add ons. And based on Apples past performance on upgrade pricing (such as on RAM), they could not compete in the upgrade card market.</strong><hr></blockquote>
Currently Apple doesn't offer anything in the truely high end (nothing of particular note from last look at least) so they haven't moved away or towards it. They just don't work in it.
They have moved away from upgradeable so to speak in their PowerMacs but they are very different machines compared to what SGI, Sun et al produce and that is different to scaleable.
It's an area that has good sized margins where Apple could definitely compete (although I doubt they want to compete directly with Sun) and should considering some of their markets.
[ 01-22-2002: Message edited by: Telomar ]</p>
<strong>Am I the only one here who thinks we will see absolutely NOTHING tomorrow? Except maybe shipment confirmation of the new iMacs... </strong><hr></blockquote>
That would suck.
Aren't there excellent education deals on the towers? Doesn't that seem like a good channel to dump inventory into?
Is anyone actually buying the towers at regular price at this point?
Alex
the G5 is likely >50 million transistors and on the same process as the Apollo... which virtually guarantees lower yields (and thus higher prices).
Actually the Apollo is said to be made on a 180nm process with SOI (by a Motorola employee in November). I too thought it would be 130nm and SOI, but unless he was mistaken 9and he hasn't bothered fixing it) then it is in fact 180nm. That means only the benefits of SOI would be in effect. That means about 25-30%. That's a confortable increase to 1.0-1.13GHz. The G5 is in fact a 130nm SOI process and should therefore have more chips per wafer than if it was made on 180nm.
<strong>Programer said:
the G5 is likely >50 million transistors and on the same process as the Apollo... which virtually guarantees lower yields (and thus higher prices).
Actually the Apollo is said to be made on a 180nm process with SOI (by a Motorola employee in November). I too thought it would be 130nm and SOI, but unless he was mistaken 9and he hasn't bothered fixing it) then it is in fact 180nm. That means only the benefits of SOI would be in effect. That means about 25-30%. That's a confortable increase to 1.0-1.13GHz. The G5 is in fact a 130nm SOI process and should therefore have more chips per wafer than if it was made on 180nm.</strong><hr></blockquote>
Oh, I didn't know that the G5 had in fact been announced and details of its process published. It is fairly normal practice for a design to appear on a non-bleeding edge process and then migrate to the latest once it is established. That's not to say I wouldn't like to see a 130nm SOI G5 right away.
<a href="http://e-www.motorola.com/collateral/PPCRMAP.pdf"; target="_blank">http://e-www.motorola.com/collateral/PPCRMAP.pdf</a>;
That's not to say that Apple will even use a G5 from Motorola's roadmap.(disclaimer: I don't know that they will or won't.)
<strong>one more day...
Well it's now 3:45 EST. Looks like Philbot is either a fake or has uncredible sources. Ohwell, this just lends hope that Apple wants to annouce the new PM's with fanfair (read:significant announcement).
<strong>one more day...
... and philbot sneaks away.
One more day (week?)
<strong>Motorola's roadmap states that the G5 will incorporate the 0.13µ process.
<a href="http://e-www.motorola.com/collateral/PPCRMAP.pdf"; target="_blank">http://e-www.motorola.com/collateral/PPCRMAP.pdf</a>;
That's not to say that Apple will even use a G5 from Motorola's roadmap.(disclaimer: I don't know that they will or won't.)</strong><hr></blockquote>
The data in the roadmap is notoriously "fluid". 2+ GHz became 800MHz - 2GHz, for example. I'm sure the G5 will end up on the stated process, or an even better one, but it wouldn't surprise me to see it appear on a slightly older process initially.
Nothing has made me change my mind about the G5 announcement. Yet.
Don't forget, the new iMac was announced when it wasn't actually available. This was because Apple NEEDED to anounce it.
Now they need to amnnounce the G5 even more!
<strong>seriously though, I did say "next week" which gives me till Friday.
.....Now they need to amnnounce the G5 even more! </strong><hr></blockquote>
I hope you are right, but I still have trouble believing you. Why do you say "they need to announce the G5 even more?" I must be missing something.
I am also curious as to your thoughts on the two threads to which I linked above, Mr. Philbot.
Hope springs eternal,
Mandricard
AppleOutsider
you don't mean the illusive Kormac surely?
Apple needs to announce the G5 to keep all the pro users and Tech media happy. No one is buying the G4 towers, but at least if we know what's coming (and for how much (£) we can make our buying decisions.
Come on Steve! you know it makes sense! lovely jubbly!!
<hr></blockquote>
Just because it has more transistors doesn't mean a thing. If more transistors = more expensive, then CPU prices would have been skyrocketing all throughout the 90s.
Like you said, it depends on how efficiently Moto can fab the G5. But you assume first that Moto will not be able to fab it with good efficiency. AFAIK, moto will work on the fab until they have an efficiency that allows them to sell a CPU at a desired price. Apple could have released G5s 6 months ago, but they would have cost $10000 per CPU.
Remember, that the G4 originally was fabbed with low efficiency, and it went into powermacs costing about the same as the current ones. The G4 has dropped in price quite a bit, now that it's being used in iMacs, and Powermac margins have probably risen.
I doubt the G5 will cost much more than the G4 did initially. Apple will be able to keep Powermac prices about where they are now...alternatively, Apple could do something better:
Keep the G4 Powermacs, albeit with a speed bump. Price them lower than they are now, beginning at, say, $1200 or so.
Add a G5 powermac lineup, and begin pricing at $2000. The G5 powermacs will be the "pro" desktop towers, and the G4 Powermacs will be the "consumer" towers.
How do I know. Well lets just say its the most obvious choice of when the G5 will be released. And so on. It's the only time that works...
<strong>
The only problem with this is that Apple has been moving away from "scalabel archatecture".
</strong>
<hr></blockquote>
What do you mean "moving away from?" They've never had one.
They've been moving toward a more modular architecture (as opposed to designing every motherboard from scratch), which is also a step toward an SGI-style architecture.
I doubt they'd adopt SGI's tech, though, for the simple reason that it's not cost effective until the costs get into the tens of thousands of dollars.
[quote]Originally posted by Junkyard Dawg:
<strong>
Just because it has more transistors doesn't mean a thing. If more transistors = more expensive, then CPU prices would have been skyrocketing all throughout the 90s.</strong><hr></blockquote>
You missed Programmer's point. He said that if the G5 was fabbed on the same process as the Apollo, it would cost more. And it would, because for a given process more transistors require more silicon, which means fewer chips per wafer, which raises the cost.
The solution is to fab bigger chips on smaller processes, and that's what Programmer was worrying about: Whether the G5 would start on a smaller process than the Apollo as planned.