I agree with a most of what you say(especially about IBM's interest in Apple's business), but you leave the impression that IBM enigeers are not too swift the the cabasa. This I disagree with.
[quote] So now that the 970 is set to debut in rougly a 1.2-2ghz range the whole philosophy of rabidly clocked G3/PPC seems to have flown out the window. Mebbe doing more work with less clocks is a good idea if you want to keep temps low and cram a lot of chips into one box, ya think? <hr></blockquote>
We will probably never know the reasons IBM came later to the SIMD table, but I'd be willing to bet they knew all along the benefits of SIMD and that MHz alone is not indicative of system speed(what speed was the Power 4 running at?).
I seriously doubt that IBM has put a SIMD on the 970 JUST for Apple. After all IBM plans on using this chip itself. I think that we can take a good guess as to why IBM didnt put a SIMD on a chip earlier as well, they didnt have the market for it outside of Apple, and thier own uses for the Power chip were directed at Servers, where the benefits of a SIMD might not be as cost effective, at least verses a dual core chip...thus they have the Power 4. However now the 970 is targeted at low end servers and DESKTOPS, which will take advantage of the SIMD. Also IBM has been investing in Linex, and if they come out with the OS mods to optemise for their SIMD, then they can make that a big advantage in selling their 970 blade servers as renderfarms. I think that there is a lot more to it than Apples desires, if anything IBM just chose to make their SIMD Altivec compatable so that they could market the chip to Apple as well as use it themselves.
As to MP systems using G3 chips, didt BE make some of these?
I agree with a most of what you say(especially about IBM's interest in Apple's business), but you leave the impression that IBM enigeers are not too swift the the cabasa. This I disagree with.
</strong><hr></blockquote>
It's not that they're stupid, it's that IBM was insisting on pure RISC as a corporate philosophy. After all, IBM invented RISC in the first place. They took one look at a piece of silicon that added 162 instructions to the ISA (pure RISC has about 100 total) and balked. Their philosophy was that that's the sort of thing you put in a companion chip. This isn't just a CPU philosophy, it's a whole architecture philosophy: IBM has for years used the CPU primarily to do basic arithmetic and tasking, and to delegate the rest of the work to ASICs. RISC was a distillation of that philosophy applied to the CPU.
It really did take them several years (and a few G3 + SIMD designs for customers like Nintendo) to realize that SIMD-on-chip really is the way things are going, rather than being a niche idea for Motorola's telephony customers. From there, it was a short hop to realize that VMX/AltiVec/Velocity Engine simply blows the doors off of every alternative.
I'm sure some of their people had to swallow a little pride to deviate that dramatically from pure RISC (more dramatically than, say, the G3's having five pipeline stages instead of the pure 4) but they have, and now they're on board.
I have to agree with Matsu, however, that even given Motorola's performance recently, the real problem was that Apple was stuck with a single vendor once they went past the G3. It's very much in Apple's interest to have Mot as a partner, even if it means waiting a little while for them to get a radical Dilbertectomy and become profitable again.
[quote]Oh god, more crap. Excuse me, this is nothing personal against you, just the whole IBm as savior crowd. Be weary, IBM has been disinterested in the past and the CAN afford to be disinterested again. 970 may repair past injuries, but IBM has been as much of a drag on PPC development as moto.<hr></blockquote>
No offense taken. I understand that the very monolithic nature of IBM is double-sided and that the move to Moto was an attempt to pique creative interest in the possibilities of the PPC.
[quote]Apple NEEDS IBM more than IBM needs Apple. Yep. Absolutely true. You have to think then that if IBM doesn't need Apple, and the SIMD-less route is truly superior, why the heck are they building an altivec chip "just for Apple." Translation, THEY AREN'T! They're builing because after 4 years they could no longer ignore the unmistakeable advantages of the design. IBM dropped the ball 4 years ago, they were wrong. They would need roughly 2.5-3Ghz G3's to compete with the fastest G4's and they aren't anywhere near delivering chips of that speed, and NEVER were.<hr></blockquote>
It's great to hear that IBM can learn from its mistakes. Now, if only Motorola can do so as well, it would do nothing but further the development of the PPC.
This next quote makes no sense to me:
[quote]They're builing because after 4 years they could no longer ignore the unmistakeable advantages of the design.<hr></blockquote>
The 'advantages' of this design have left us behind the Intel/AMD crowd for far too long, and at premium prices to boot. Of course IBM is going to profit off the advances made by Motorola, but notice that they only do so when it's become clear that Motorola itself cannot or is not interested in advancing the G4 or the Mac platform.
The existence of a viable alternative to an opposing monolithic software company could be worth a heck of a lot to someone who creates hardware solutions. It's true that IBM can't get in on creating the full box for Apple, but they have much less opportunity in the Windows World. IBM can place itself as a dominant provider for a great platform that has proven resilient beyond all reasonable expectation. The Mac platform is vibrant, healthy, and is even enjoying a modest fire of resurgence that IBM would do well to feed with a faster processor with improved specs that could compete against the the Wintel world.
This, of course, has nothing to do with revamped G3s. I don't claim to know anything about the viability of new G3 chips -- I'm just skeptical that a 'year of the laptop' can progress very far with G4s and likewise skeptical that 970s will be in laptops anytime soon.
Dealer channels suggest that the current sum total number of 17" Lapzilla in the USA is exactly ONE.
This is the ONLY 17" machine... shown by SJ during the Keynote (opened but not actively demo'd)... seen on MiniMe's lap in the Yao Ming commercial (with screen digitally painted in... not actively demo'd).
This single 17 inch PB is couriered around the country to Apple Stores and PR shoots. Year of the laptop.
Another source suggested there is in fact an entire warehouse of new Aluminum 15.4 inch PBs sitting under plastic... cooling its heels until the previously announced model ships before it can be announced itself and note generate bad news buzz about the wait for the Last great-but-late product.
<evil genius>
Not sure if i agree with the 2nd rumour of a warehouse of 15.4AlPBs, but it would sure make a more interesting target of a heist movie. The SOHO Apple store got burglarized of all its Ti867s at one point, IIRC... imaging taking a whole warehouse full of unannounced widescreen GigAlBooks....
Back to the G3: If MojaveMP really means some sort of SMP support, then that makes a certain amount of sense for IBM. The G3 has historically not had SMP support, because it's been targeted as a budget processor, but adding that support is not a tremendous endeavour. If Mojave abandons the 60x bus in favor of RapidIO, that makes MP work in a different way.
I honestly don't expect Apple to adopt these G3s in current Mac models. The G4 will shortly move to a point where it can power Apple's consumer lines, and should Apple decide to go with SMP configurations on the consumer end (I'd love a DP iMac!) the G4 already has robust support.
Don't forget the embedded market. The G3's a sweet little chip for that market, SIMD or no.
Remember Tik, that however far behind the G4 is, the G3 is up to 3X further behind. Moto's problem is not design, it's fab. If the FSB were even 266DDR and we had plentiful G4's running about 2Ghz or so, I'd guess that chip would be very competitive within the confines of Apple's tightter HW-SF integration.
[quote]The G4 will shortly move to a point where it can power Apple's consumer lines, and should Apple decide to go with SMP configurations on the consumer end (I'd love a DP iMac!) the G4 already has robust support.<hr></blockquote>
Unless Moto gives up making them altogether.
This *could* just be IBMs stab at a G4 replacement for a Moto backout.
So, in reality, it's not a G3....but it could put dual processors in a laptop! And THAT'S what I'm excited about.
Unless Moto gives up making them altogether.</strong><hr></blockquote>
Not likely.
Mot SPS just turned a much-needed profit, if memory serves, and the G4 is a cash cow for them. It has a good, long future ahead of it.
[quote]<strong>This *could* just be IBMs stab at a G4 replacement for a Moto backout.</strong><hr></blockquote>
It could also be a competitor to the G4. We don't really know.
[quote]<strong>So, in reality, it's not a G3....but it could put dual processors in a laptop! And THAT'S what I'm excited about.</strong><hr></blockquote>
The next die shrink for the G4 could make an MP PowerBook feasible as well. The only question would be whether it would be better than a single 970.
I've said this before but you'll find Apple calling this the year of the laptop was more about talking to Wall Street than magical upcoming releases.
Laptops are a massive growth sector and if you pronounce this is Apple's "Year of the Laptop" it a) hopefully gets attention away from the fact the rest of your products are pretty subpar and b) aggressively focused on gaining market share in an area where you can compete.
[quote]Imagine current clock speeds (lower actually) and NO altivec, that's what an all IBM mac line-up would have looked like. Incompetence or indifference makes no difference. Whatever the reasons, Moto has clearly contributed much more than IBM has over the last 4 years.
I would worry intensely about an all IBM line-up. They grew disinterested once, and may yet do so again a few years from now.<hr></blockquote>
A lot of people love to say this but it relies on the assumption that IBM had a reason to bother with the desktop market or to assist.
I think you'll find Apple contracted themselves to Motorola and left IBM in the cold not the other way round.
They did get it wrong on the subject of SIMD but the 970 does have Altivec and they stated on Sahara's release the follow up would have a SIMD unit as well.
Motorola can still offer something to the mobile market but I rather expect the clock speed increases will drop once he Powermacs go. All depends on the contracts Apple has in place.
If IBM were to revamp the 7xx line it could easily make a nicer offering than what Motorola. The 7xx series of chips that IBM makes is an excellent chip for its market it's really just a question of how how will evolve it and if they'll go after the mobile computer market or not.
I'm not saying that IBM is incompetent, I'm saying they can afford to change their minds. Apple hasn't nearly as many options.
What you say about incentives is true. But remember that IBM still sells 604 based computers, that isn't exactly a commitment to advancing the PPC for their own use. If they lose interest, Apple is not going to hold their attention any better than Jobs could eke favors out of moto.
Maybe it's just me, but I haven't heard a word of the "year of the notebook" slogan since the keynote. It seems to be all but moot. If it has any basis, though, it means we'll see 7457-powered PowerBooks AND iBooks by year end and 970 (if it even happens) in January.
Sorry to be realistic. Hope the dreams surrounding our reality move in a little closer...
<strong>I'm not saying that IBM is incompetent, I'm saying they can afford to change their minds. Apple hasn't nearly as many options.
What you say about incentives is true. But remember that IBM still sells 604 based computers, that isn't exactly a commitment to advancing the PPC for their own use. If they lose interest, Apple is not going to hold their attention any better than Jobs could eke favors out of moto.</strong><hr></blockquote>
You sort of make it sound like IBM could just stop development overnight if they were going to, which they can't. Apple will have contracts in place and a lot of lead time if IBM does decide to move from that segment. I expect they will probably stay there, Apple will buy chips from them, Motorola will try and get itself in order and everybody will be happy.
All that "lead time" did them a world of good with Moto didn't it. uh-un, ther's no such iron clad guarantee, contracts or no. I can decide to stop developing chips and I can decide to stop developing chips, if you follow. Scale back, 2 engineers instead of 20, stop updating facilities, miss targets, get sloppy etc etc... What so we have a contract? I got chips, here's a nice 1.42Ghz FSB crippled .18u G4 for ya, see I got chips.
Contracts don't mean shit. By your logic Apple knew quite well of Moto's problems, but that didn't help them to have a replacement in the cards in any sort of timely fashion now did it? You've got it backwards mon amis, starting production takes time, stopping it takes nothing.
I tell you now, and I hope for the good of the mac platform that it doesn't happen, but a one CPU maker game has as much danger in a post 970 world as it does in the pre970 G4 world. Apple IS small enough to screw over, easily. By the time the lawyers sorted it out, Micheal Dell would be dancing on the ashes.
<strong>All that "lead time" did them a world of good with Moto didn't it. uh-un, ther's no such iron clad guarantee, contracts or no. I can decide to stop developing chips and I can decide to stop developing chips, if you follow. Scale back, 2 engineers instead of 20, stop updating facilities, miss targets, get sloppy etc etc... What so we have a contract? I got chips, here's a nice 1.42Ghz FSB crippled .18u G4 for ya, see I got chips.
Contracts don't mean shit. By your logic Apple knew quite well of Moto's problems, but that didn't help them to have a replacement in the cards in any sort of timely fashion now did it? You've got it backwards mon amis, starting production takes time, stopping it takes nothing.</strong><hr></blockquote>
You're a little confused on the Motorola front. For a start the initial faults with the G4 were production. That's a different case to stopping development. If you took away the 1 year of stall that the G4 had it has actually progressed damn well except at these late stages and yet again that is production problems not development issues.
None of the events you've described Motorola has done on purpose. For Motorola the G4 is a cash cow and in troubled times that's where you turn back to and that's exactly what Motorola has done only they aren't doing a very good job. You seem to think Motorola has done all this just to spite Apple and that's irrational. You're mistaking Motorola's incompetence for spite and there's no evidence to support that.
Lets talk bus speeds for a moment though. You are again assuming Apple wasn't told what limitations there would be and just agreed to it. It's very likely Apple would have been involved in the G4s development and known the limitations. Again if you take out the production problems Motorola has pretty much scaled the G4 exactly as they stated they would and Apple would have had that in advance.
Most of the faults with the G4 fall into 2 categories. Incompetence of Motorola and poor foresight from Apple. Apple bet on an architecture that really wasn't that well suited to the desktop market it turns out.
Finally I'm not sure how old you are but I'll tell you virtually every contract I've entered into has specified penalties for failure to meet deadlines and those that didn't are based on faith that something can be delivered by a certain timeline. Apple would have seen roadmaps and entered into the contract with a faith that that's what would be followed. Failure to meet that roadmap could be costly depending on the contract itself and the conigencies that were allowed for. You'd really need to see the individual contract.
There are methods around contracts but I'll give you an example of how hard what you think can occur is. HP wants the Alpha dead. They're actually trying to kill that architecture and wouldn't care if they didn't sell one more. Yet because of contracts with major buyers they've had to continue development.
Are you getting the idea? It isn't as easy as you think to get out of contracts without enduring some massive consequences.
Hopefully, by going with the 9xx family, Apple will have a stable road of dependable processors leading them into the future. IBM has a well planned future roadmap, and hopefuly, this will mean that Apple won't be stuck using the same processor for 3+ years like the way that the G4 is happening.
Comments
I agree with a most of what you say(especially about IBM's interest in Apple's business), but you leave the impression that IBM enigeers are not too swift the the cabasa. This I disagree with.
[quote] So now that the 970 is set to debut in rougly a 1.2-2ghz range the whole philosophy of rabidly clocked G3/PPC seems to have flown out the window. Mebbe doing more work with less clocks is a good idea if you want to keep temps low and cram a lot of chips into one box, ya think? <hr></blockquote>
We will probably never know the reasons IBM came later to the SIMD table, but I'd be willing to bet they knew all along the benefits of SIMD and that MHz alone is not indicative of system speed(what speed was the Power 4 running at?).
[ 02-24-2003: Message edited by: rickag ]</p>
As to MP systems using G3 chips, didt BE make some of these?
[ 02-24-2003: Message edited by: @homenow ]</p>
<strong>Matsu
I agree with a most of what you say(especially about IBM's interest in Apple's business), but you leave the impression that IBM enigeers are not too swift the the cabasa. This I disagree with.
</strong><hr></blockquote>
It's not that they're stupid, it's that IBM was insisting on pure RISC as a corporate philosophy. After all, IBM invented RISC in the first place. They took one look at a piece of silicon that added 162 instructions to the ISA (pure RISC has about 100 total) and balked. Their philosophy was that that's the sort of thing you put in a companion chip. This isn't just a CPU philosophy, it's a whole architecture philosophy: IBM has for years used the CPU primarily to do basic arithmetic and tasking, and to delegate the rest of the work to ASICs. RISC was a distillation of that philosophy applied to the CPU.
It really did take them several years (and a few G3 + SIMD designs for customers like Nintendo) to realize that SIMD-on-chip really is the way things are going, rather than being a niche idea for Motorola's telephony customers. From there, it was a short hop to realize that VMX/AltiVec/Velocity Engine simply blows the doors off of every alternative.
I'm sure some of their people had to swallow a little pride to deviate that dramatically from pure RISC (more dramatically than, say, the G3's having five pipeline stages instead of the pure 4) but they have, and now they're on board.
I have to agree with Matsu, however, that even given Motorola's performance recently, the real problem was that Apple was stuck with a single vendor once they went past the G3. It's very much in Apple's interest to have Mot as a partner, even if it means waiting a little while for them to get a radical Dilbertectomy and become profitable again.
[ 02-24-2003: Message edited by: Amorph ]</p>
<strong>No <img src="graemlins/bugeye.gif" border="0" alt="[Skeptical]" /> </strong><hr></blockquote>
Could you elaborate for the continually
No offense taken. I understand that the very monolithic nature of IBM is double-sided and that the move to Moto was an attempt to pique creative interest in the possibilities of the PPC.
[quote]Apple NEEDS IBM more than IBM needs Apple. Yep. Absolutely true. You have to think then that if IBM doesn't need Apple, and the SIMD-less route is truly superior, why the heck are they building an altivec chip "just for Apple." Translation, THEY AREN'T! They're builing because after 4 years they could no longer ignore the unmistakeable advantages of the design. IBM dropped the ball 4 years ago, they were wrong. They would need roughly 2.5-3Ghz G3's to compete with the fastest G4's and they aren't anywhere near delivering chips of that speed, and NEVER were.<hr></blockquote>
It's great to hear that IBM can learn from its mistakes. Now, if only Motorola can do so as well, it would do nothing but further the development of the PPC.
This next quote makes no sense to me:
[quote]They're builing because after 4 years they could no longer ignore the unmistakeable advantages of the design.<hr></blockquote>
The 'advantages' of this design have left us behind the Intel/AMD crowd for far too long, and at premium prices to boot. Of course IBM is going to profit off the advances made by Motorola, but notice that they only do so when it's become clear that Motorola itself cannot or is not interested in advancing the G4 or the Mac platform.
The existence of a viable alternative to an opposing monolithic software company could be worth a heck of a lot to someone who creates hardware solutions. It's true that IBM can't get in on creating the full box for Apple, but they have much less opportunity in the Windows World. IBM can place itself as a dominant provider for a great platform that has proven resilient beyond all reasonable expectation. The Mac platform is vibrant, healthy, and is even enjoying a modest fire of resurgence that IBM would do well to feed with a faster processor with improved specs that could compete against the the Wintel world.
This, of course, has nothing to do with revamped G3s. I don't claim to know anything about the viability of new G3 chips -- I'm just skeptical that a 'year of the laptop' can progress very far with G4s and likewise skeptical that 970s will be in laptops anytime soon.
<strong> "Year of the Laptop" </strong>
note singular, not plural
Dealer channels suggest that the current sum total number of 17" Lapzilla in the USA is exactly ONE.
This is the ONLY 17" machine... shown by SJ during the Keynote (opened but not actively demo'd)... seen on MiniMe's lap in the Yao Ming commercial (with screen digitally painted in... not actively demo'd).
This single 17 inch PB is couriered around the country to Apple Stores and PR shoots. Year of the laptop.
Another source suggested there is in fact an entire warehouse of new Aluminum 15.4 inch PBs sitting under plastic... cooling its heels until the previously announced model ships before it can be announced itself and note generate bad news buzz about the wait for the Last great-but-late product.
<evil genius>
Not sure if i agree with the 2nd rumour of a warehouse of 15.4AlPBs, but it would sure make a more interesting target of a heist movie. The SOHO Apple store got burglarized of all its Ti867s at one point, IIRC... imaging taking a whole warehouse full of unannounced widescreen GigAlBooks....
mua ha ha
</evil genius>
</devils_advocate>
no dual g3... unsupported function.
more than one Lapzilla is a better goal.
[ 02-24-2003: Message edited by: curiousuburb ]</p>
I honestly don't expect Apple to adopt these G3s in current Mac models. The G4 will shortly move to a point where it can power Apple's consumer lines, and should Apple decide to go with SMP configurations on the consumer end (I'd love a DP iMac!) the G4 already has robust support.
Don't forget the embedded market. The G3's a sweet little chip for that market, SIMD or no.
Unless Moto gives up making them altogether.
This *could* just be IBMs stab at a G4 replacement for a Moto backout.
So, in reality, it's not a G3....but it could put dual processors in a laptop! And THAT'S what I'm excited about.
<strong>
Unless Moto gives up making them altogether.</strong><hr></blockquote>
Not likely.
Mot SPS just turned a much-needed profit, if memory serves, and the G4 is a cash cow for them. It has a good, long future ahead of it.
[quote]<strong>This *could* just be IBMs stab at a G4 replacement for a Moto backout.</strong><hr></blockquote>
It could also be a competitor to the G4. We don't really know.
[quote]<strong>So, in reality, it's not a G3....but it could put dual processors in a laptop! And THAT'S what I'm excited about.</strong><hr></blockquote>
The next die shrink for the G4 could make an MP PowerBook feasible as well. The only question would be whether it would be better than a single 970.
I've said this before but you'll find Apple calling this the year of the laptop was more about talking to Wall Street than magical upcoming releases.
Laptops are a massive growth sector and if you pronounce this is Apple's "Year of the Laptop" it a) hopefully gets attention away from the fact the rest of your products are pretty subpar and b) aggressively focused on gaining market share in an area where you can compete.
[quote]Imagine current clock speeds (lower actually) and NO altivec, that's what an all IBM mac line-up would have looked like. Incompetence or indifference makes no difference. Whatever the reasons, Moto has clearly contributed much more than IBM has over the last 4 years.
I would worry intensely about an all IBM line-up. They grew disinterested once, and may yet do so again a few years from now.<hr></blockquote>
A lot of people love to say this but it relies on the assumption that IBM had a reason to bother with the desktop market or to assist.
I think you'll find Apple contracted themselves to Motorola and left IBM in the cold not the other way round.
They did get it wrong on the subject of SIMD but the 970 does have Altivec and they stated on Sahara's release the follow up would have a SIMD unit as well.
Motorola can still offer something to the mobile market but I rather expect the clock speed increases will drop once he Powermacs go. All depends on the contracts Apple has in place.
If IBM were to revamp the 7xx line it could easily make a nicer offering than what Motorola. The 7xx series of chips that IBM makes is an excellent chip for its market it's really just a question of how how will evolve it and if they'll go after the mobile computer market or not.
What you say about incentives is true. But remember that IBM still sells 604 based computers, that isn't exactly a commitment to advancing the PPC for their own use. If they lose interest, Apple is not going to hold their attention any better than Jobs could eke favors out of moto.
Any ideas (besides short-sighted idiocy on the part of both)?
Sorry to be realistic. Hope the dreams surrounding our reality move in a little closer...
I couldn't help it...I needed something faster than this iBook 600mhz.
<strong>I'm not saying that IBM is incompetent, I'm saying they can afford to change their minds. Apple hasn't nearly as many options.
What you say about incentives is true. But remember that IBM still sells 604 based computers, that isn't exactly a commitment to advancing the PPC for their own use. If they lose interest, Apple is not going to hold their attention any better than Jobs could eke favors out of moto.</strong><hr></blockquote>
You sort of make it sound like IBM could just stop development overnight if they were going to, which they can't. Apple will have contracts in place and a lot of lead time if IBM does decide to move from that segment. I expect they will probably stay there, Apple will buy chips from them, Motorola will try and get itself in order and everybody will be happy.
Contracts don't mean shit. By your logic Apple knew quite well of Moto's problems, but that didn't help them to have a replacement in the cards in any sort of timely fashion now did it? You've got it backwards mon amis, starting production takes time, stopping it takes nothing.
I tell you now, and I hope for the good of the mac platform that it doesn't happen, but a one CPU maker game has as much danger in a post 970 world as it does in the pre970 G4 world. Apple IS small enough to screw over, easily. By the time the lawyers sorted it out, Micheal Dell would be dancing on the ashes.
Yet another reason why market share matters.
[ 02-25-2003: Message edited by: Matsu ]</p>
<strong>All that "lead time" did them a world of good with Moto didn't it. uh-un, ther's no such iron clad guarantee, contracts or no. I can decide to stop developing chips and I can decide to stop developing chips, if you follow. Scale back, 2 engineers instead of 20, stop updating facilities, miss targets, get sloppy etc etc... What so we have a contract? I got chips, here's a nice 1.42Ghz FSB crippled .18u G4 for ya, see I got chips.
Contracts don't mean shit. By your logic Apple knew quite well of Moto's problems, but that didn't help them to have a replacement in the cards in any sort of timely fashion now did it? You've got it backwards mon amis, starting production takes time, stopping it takes nothing.</strong><hr></blockquote>
You're a little confused on the Motorola front. For a start the initial faults with the G4 were production. That's a different case to stopping development. If you took away the 1 year of stall that the G4 had it has actually progressed damn well except at these late stages and yet again that is production problems not development issues.
None of the events you've described Motorola has done on purpose. For Motorola the G4 is a cash cow and in troubled times that's where you turn back to and that's exactly what Motorola has done only they aren't doing a very good job. You seem to think Motorola has done all this just to spite Apple and that's irrational. You're mistaking Motorola's incompetence for spite and there's no evidence to support that.
Lets talk bus speeds for a moment though. You are again assuming Apple wasn't told what limitations there would be and just agreed to it. It's very likely Apple would have been involved in the G4s development and known the limitations. Again if you take out the production problems Motorola has pretty much scaled the G4 exactly as they stated they would and Apple would have had that in advance.
Most of the faults with the G4 fall into 2 categories. Incompetence of Motorola and poor foresight from Apple. Apple bet on an architecture that really wasn't that well suited to the desktop market it turns out.
Finally I'm not sure how old you are but I'll tell you virtually every contract I've entered into has specified penalties for failure to meet deadlines and those that didn't are based on faith that something can be delivered by a certain timeline. Apple would have seen roadmaps and entered into the contract with a faith that that's what would be followed. Failure to meet that roadmap could be costly depending on the contract itself and the conigencies that were allowed for. You'd really need to see the individual contract.
There are methods around contracts but I'll give you an example of how hard what you think can occur is. HP wants the Alpha dead. They're actually trying to kill that architecture and wouldn't care if they didn't sell one more. Yet because of contracts with major buyers they've had to continue development.
Are you getting the idea? It isn't as easy as you think to get out of contracts without enduring some massive consequences.