I'm with you! Those were the days , let them Rest in Peace .
It's a different "game" now ... with IBM providing the CPUs.
I wouldn't "rule out" dual cores in the 980. It all comes down to which config will provide the maximum performance in the "cheapest package.
Who knows ... HyperThreaded with huge caches, or dual core with marginal cache increase. I can tell you that they're already messing with the 90 nano processes, and it makes sense to test the different ideas.
What will the 980 be called by Apple? G6? And when is it supposed to be coming?
Steve said that IBM and they were already at work on "the next generation" CPU. I would think that a next generation would get its own name, for marketing reasons if nothing else.
in the G5 introduction video on the Apple website, they chap from IBM states that they already have prototypes of the next gen chip for Apple. the only question now is: "is it 970+ or 980?"
You can check the web for commentary by Apple, in another thread somewhere in AI right now, can't remember which tho, stating that Apple-Moto still has a LONG life ahead and that pretty much guarantees you won't see a 970 based PB any time soon, not even in six months.
Apple is STILL using G3's and will be using G4's still (in its consumer products) 18+ months from now.
Moto will "fix" the G4.
Hang on there Matsu! I have a question for you or anyone in the know. What will have to be done to fix the G4? Everyone talks about the slow FSB that strangles the G4 which would otherwise be alot more faster. Who is to be blamed for this then? Motorola for making the processor or Apple for making the motherboard. What should be done to enhance the architecture? I'd appreciate if Matsu or someone knowledgable set me straight on this. I'm looking to making the switch from a PC and buying me a 15.4" Al-Powerbook before years end and the way I see it that means I'll get a G4- not a G5-processor. Programmer, I need your wisdom!
I'd just like to point out that a PPC 980 might very well still be marketed as G5 or G5+, after all the G4 also has undergone some major changes without actually changing name.
However I don't expect the 980 to come any sooner than after 12 months. And I don't think that the mentioned 3GHz chip is a 980 either.
I agree with G-News. The 3GHz chip they speak of is most likely a 90nm version of the 970. As to what it will be called, who knows. IBM uses letter suffix for the lower end line like 750CX, 750CXe, 750FX (and this one included a bump in cache), or 750GX. This even goes for the PowerPC 405 and 440 lines. But the POWER line uses the "+" nomenclature for processors going to the next process level. So a 970 on 90nm would probably be called either the 970FX or the 970+. If it also includes extra cache, maybe 970GX? I think the 750CX name came about when the 750 first incorporated built in L2 cache. "C" for cache, "X" for extra. The 970 already has has built in cache. The FX line came out when they went to the next process level (130nm) and raised the cache. "F" is for "???" WHo knows how they really name things.
Anyways thats not important. What is important is if the 90nm process shrink will include more L2 or not. I say yes. And it'll probably scale to just over 3GHz, probably to 4GHz even before the 980, or whatever the 9X0 series based on the Power5 will be called, comes out.
Hang on there Matsu! I have a question for you or anyone in the know. What will have to be done to fix the G4? Everyone talks about the slow FSB that strangles the G4 which would otherwise be alot more faster. Who is to be blamed for this then? Motorola for making the processor or Apple for making the motherboard. What should be done to enhance the architecture?
The biggest problem is the slow connection to the rest of the system. The G4 is like a Ferrari with a lawnmower engine. Motorola is to blame for this. That's pretty clear if you look at the PowerMac G5 with a 1 Ghz bus and Dual-channel DDR400. With Motorola out of the way, Apple was able to build one kick-ass system.
I'd just like to point out that a PPC 980 might very well still be marketed as G5 or G5+, after all the G4 also has undergone some major changes without actually changing name.
The G4 hasn't had many new features, though. The changes were mostly extensions of the pipeline and other modifications to increase the clockrate. A processor with SMT, dual-cores and/or other goodies should have a new name. It might be different if we don't see either of those, but I doubt that.
Hang on there Matsu! I have a question for you or anyone in the know. What will have to be done to fix the G4? Everyone talks about the slow FSB that strangles the G4 which would otherwise be alot more faster. Who is to be blamed for this then? Motorola for making the processor or Apple for making the motherboard. What should be done to enhance the architecture? I'd appreciate if Matsu or someone knowledgable set me straight on this. I'm looking to making the switch from a PC and buying me a 15.4" Al-Powerbook before years end and the way I see it that means I'll get a G4- not a G5-processor. Programmer, I need your wisdom!
Hang on there Matsu! I have a question for you or anyone in the know. What will have to be done to fix the G4?
Not really so much. If all they did was move the memory controller on board and use RapidIO for a system bus, they'd remove the bus bottleneck and the CPU would suddenly run a lot faster - especially at AltiVec tasks.
If they added another integer unit and another FPU as well, they'd have a real performer. Hyperthreading really isn't an option, because the G4 doesn't do a lot of OOE, but a dual-core G4 would be nice too.
The two things Mot has to do are: 1) get away from MaxBus; 2) get their stuff out of their own fabs. Everything else is gravy.
Quote:
Everyone talks about the slow FSB that strangles the G4 which would otherwise be alot more faster. Who is to be blamed for this then? Motorola for making the processor or Apple for making the motherboard.
Motorola, because the CPU dictates the bus used to connect it to the motherboard. Some people have blamed Apple for not convincing Mot to adopt another technology, but that doesn't really persuade me.
Not really so much. If all they did was move the memory controller on board and use RapidIO for a system bus, they'd remove the bus bottleneck and the CPU would suddenly run a lot faster - especially at AltiVec tasks.
If they added another integer unit and another FPU as well, they'd have a real performer. Hyperthreading really isn't an option, because the G4 doesn't do a lot of OOE, but a dual-core G4 would be nice too.
The two things Mot has to do are: 1) get away from MaxBus; 2) get their stuff out of their own fabs. Everything else is gravy.
Motorola, because the CPU dictates the bus used to connect it to the motherboard. Some people have blamed Apple for not convincing Mot to adopt another technology, but that doesn't really persuade me.
Amorph,
Why would Apple want to bother with the G4 (improved or not) if the comparative price of the G4 and 970 being bantered about are even remotely close to correct? (It is said that IBM is providing 970s at a price which is less than what Apple has been paying for G4s.)
Why would Apple want to bother with the G4 (improved or not) if the comparative price of the G4 and 970 being bantered about are even remotely close to correct? (It is said that IBM is providing 970s at a price which is less than what Apple has been paying for G4s.)
The 7457 is supposed to cost half of what the 7455 does and is pin compatible with the 7455.
That means it costs less than the current G4 and the G5, and no change in architecture is needed. That's a huge savings overall. If Moto can deliver.
Why would Apple want to bother with the G4 (improved or not) if the comparative price of the G4 and 970 being bantered about are even remotely close to correct? (It is said that IBM is providing 970s at a price which is less than what Apple has been paying for G4s.)
Good question.
First of all, we don't know that the 970 is cheaper than the XPC7455 in the towers. That's a rumor.
We certainly don't know if it's cheaper than the more mainstream MPC7455s, which the 7457 will be replacing at half the cost. So my bet is that the 7457 (and the 7447) will be significantly smaller, cooler and cheaper than the 970 - certainly much cooler. A low voltage '47 could even go into the iBook, where the 970 won't show up for a good long time.
If Mot comes through with a souped-up, RapidIO-connected PowerPC (maybe a G4, maybe not) in early 2004 like they're now threatening to, the 970 might stay out of Apple's laptop line for a good while. The 970 has excellent power characteristics considering what it is, but I have no doubt that Mot can deliver a CPU with exceptional performance given its small size, clean design and low power consumption. If there's anything they're good at, it's that.
Comments
Originally posted by Hobbes
Those were the days.
Different game now with IBM.
I'm with you! Those were the days , let them Rest in Peace .
It's a different "game" now ... with IBM providing the CPUs.
I wouldn't "rule out" dual cores in the 980. It all comes down to which config will provide the maximum performance in the "cheapest package.
Who knows ... HyperThreaded with huge caches, or dual core with marginal cache increase. I can tell you that they're already messing with the 90 nano processes, and it makes sense to test the different ideas.
Originally posted by Placebo
What will the 980 be called by Apple? G6? And when is it supposed to be coming?
Steve said that IBM and they were already at work on "the next generation" CPU. I would think that a next generation would get its own name, for marketing reasons if nothing else.
As far as I know, there is no roadmap that mentions such a thing and it's just some rumor-monger's invention.
Please correct if I'm wrong, but is there anything that would substantiate IBM's plans to release a PPC980?
Originally posted by Tomb of the Unknown
Just out of curiousity, where has IBM released any info about a "980"?
As far as I know, there is no roadmap that mentions such a thing and it's just some rumor-monger's invention.
Please correct if I'm wrong, but is there anything that would substantiate IBM's plans to release a PPC980?
As was the PPC970 as far as Apple using it
g::masta
Originally posted by Matsu
You can check the web for commentary by Apple, in another thread somewhere in AI right now, can't remember which tho, stating that Apple-Moto still has a LONG life ahead and that pretty much guarantees you won't see a 970 based PB any time soon, not even in six months.
Apple is STILL using G3's and will be using G4's still (in its consumer products) 18+ months from now.
Moto will "fix" the G4.
Hang on there Matsu! I have a question for you or anyone in the know. What will have to be done to fix the G4? Everyone talks about the slow FSB that strangles the G4 which would otherwise be alot more faster. Who is to be blamed for this then? Motorola for making the processor or Apple for making the motherboard. What should be done to enhance the architecture? I'd appreciate if Matsu or someone knowledgable set me straight on this. I'm looking to making the switch from a PC and buying me a 15.4" Al-Powerbook before years end and the way I see it that means I'll get a G4- not a G5-processor. Programmer, I need your wisdom!
However I don't expect the 980 to come any sooner than after 12 months. And I don't think that the mentioned 3GHz chip is a 980 either.
Originally posted by hmurchison
As was the PPC970 as far as Apple using it
Yes, but in that case we had a good deal of info from IBM about the chip. There's nothing but smoke where the "980" is concerned.
Anyways thats not important. What is important is if the 90nm process shrink will include more L2 or not. I say yes. And it'll probably scale to just over 3GHz, probably to 4GHz even before the 980, or whatever the 9X0 series based on the Power5 will be called, comes out.
Originally posted by KANE
Hang on there Matsu! I have a question for you or anyone in the know. What will have to be done to fix the G4? Everyone talks about the slow FSB that strangles the G4 which would otherwise be alot more faster. Who is to be blamed for this then? Motorola for making the processor or Apple for making the motherboard. What should be done to enhance the architecture?
The biggest problem is the slow connection to the rest of the system. The G4 is like a Ferrari with a lawnmower engine. Motorola is to blame for this. That's pretty clear if you look at the PowerMac G5 with a 1 Ghz bus and Dual-channel DDR400. With Motorola out of the way, Apple was able to build one kick-ass system.
Originally posted by G-News
I'd just like to point out that a PPC 980 might very well still be marketed as G5 or G5+, after all the G4 also has undergone some major changes without actually changing name.
The G4 hasn't had many new features, though. The changes were mostly extensions of the pipeline and other modifications to increase the clockrate. A processor with SMT, dual-cores and/or other goodies should have a new name. It might be different if we don't see either of those, but I doubt that.
Originally posted by wfzelle
The G4 is like a Ferrari with a lawnmower engine.
I'd say the G4 is like a Ferrari with a one litre gas tank.
Originally posted by KANE
Hang on there Matsu! I have a question for you or anyone in the know. What will have to be done to fix the G4? Everyone talks about the slow FSB that strangles the G4 which would otherwise be alot more faster. Who is to be blamed for this then? Motorola for making the processor or Apple for making the motherboard. What should be done to enhance the architecture? I'd appreciate if Matsu or someone knowledgable set me straight on this. I'm looking to making the switch from a PC and buying me a 15.4" Al-Powerbook before years end and the way I see it that means I'll get a G4- not a G5-processor. Programmer, I need your wisdom!
Please read this (page 32):
http://e-www.motorola.com/collateral...ROPE_H1101.pdf
Then take this into consideration:
http://www.siliconstrategies.com/story/OEG20030623S0089
Originally posted by bunge
I'd say the G4 is like a Ferrari with a one litre gas tank.
No, the best analogy is the G4 was like a Ferrari with a 2 gear transmission.
Originally posted by KANE
Hang on there Matsu! I have a question for you or anyone in the know. What will have to be done to fix the G4?
Not really so much. If all they did was move the memory controller on board and use RapidIO for a system bus, they'd remove the bus bottleneck and the CPU would suddenly run a lot faster - especially at AltiVec tasks.
If they added another integer unit and another FPU as well, they'd have a real performer. Hyperthreading really isn't an option, because the G4 doesn't do a lot of OOE, but a dual-core G4 would be nice too.
The two things Mot has to do are: 1) get away from MaxBus; 2) get their stuff out of their own fabs. Everything else is gravy.
Everyone talks about the slow FSB that strangles the G4 which would otherwise be alot more faster. Who is to be blamed for this then? Motorola for making the processor or Apple for making the motherboard.
Motorola, because the CPU dictates the bus used to connect it to the motherboard. Some people have blamed Apple for not convincing Mot to adopt another technology, but that doesn't really persuade me.
Originally posted by Amorph
Not really so much. If all they did was move the memory controller on board and use RapidIO for a system bus, they'd remove the bus bottleneck and the CPU would suddenly run a lot faster - especially at AltiVec tasks.
If they added another integer unit and another FPU as well, they'd have a real performer. Hyperthreading really isn't an option, because the G4 doesn't do a lot of OOE, but a dual-core G4 would be nice too.
The two things Mot has to do are: 1) get away from MaxBus; 2) get their stuff out of their own fabs. Everything else is gravy.
Motorola, because the CPU dictates the bus used to connect it to the motherboard. Some people have blamed Apple for not convincing Mot to adopt another technology, but that doesn't really persuade me.
Amorph,
Why would Apple want to bother with the G4 (improved or not) if the comparative price of the G4 and 970 being bantered about are even remotely close to correct? (It is said that IBM is providing 970s at a price which is less than what Apple has been paying for G4s.)
Originally posted by Outsider
No, the best analogy is the G4 was like a Ferrari with a 2 gear transmission.
Both gears in reverse (lo and hi)???
Originally posted by RBR
Amorph,
Why would Apple want to bother with the G4 (improved or not) if the comparative price of the G4 and 970 being bantered about are even remotely close to correct? (It is said that IBM is providing 970s at a price which is less than what Apple has been paying for G4s.)
The 7457 is supposed to cost half of what the 7455 does and is pin compatible with the 7455.
That means it costs less than the current G4 and the G5, and no change in architecture is needed. That's a huge savings overall. If Moto can deliver.
Originally posted by RBR
Why would Apple want to bother with the G4 (improved or not) if the comparative price of the G4 and 970 being bantered about are even remotely close to correct? (It is said that IBM is providing 970s at a price which is less than what Apple has been paying for G4s.)
Good question.
First of all, we don't know that the 970 is cheaper than the XPC7455 in the towers. That's a rumor.
We certainly don't know if it's cheaper than the more mainstream MPC7455s, which the 7457 will be replacing at half the cost. So my bet is that the 7457 (and the 7447) will be significantly smaller, cooler and cheaper than the 970 - certainly much cooler. A low voltage '47 could even go into the iBook, where the 970 won't show up for a good long time.
If Mot comes through with a souped-up, RapidIO-connected PowerPC (maybe a G4, maybe not) in early 2004 like they're now threatening to, the 970 might stay out of Apple's laptop line for a good while. The 970 has excellent power characteristics considering what it is, but I have no doubt that Mot can deliver a CPU with exceptional performance given its small size, clean design and low power consumption. If there's anything they're good at, it's that.