You mean that IBM's $3 Billion investment in the DeadFish Plant was all based on PPC sales to Apple, now there's a great business plan. Maybe IBM stole some of MOT's management people when Intel was stealing their engineers.
And I guess this also means that IBM's fancy computerized design process in reality is actually two University of Michigan Computer Science teachers.
Apple has been hard at work designing their own chipset for the 970, a chipset also being made at the Fishkill plant. They have intimate knowledge of the bus and how it works and should be designed. They have real world practical knowledge of it, more so than IBM themselves, arguably! Now imagine a scenario where Motorola would license the G4 core to Apple who would in turn design the 970 bus around the 745X core, have IBM fabricate at the Fishkill plant, and drop L3 cache support. I understand the modularity of the G4 core would allow something like this, albeit with considerable effort on Apple?s part, but the return would be greater. Desktop-wise they would already have a companion chip and the G4 hybrid would satisfy the power requirements of an iMac/PowerBook application.
1) nobody said, that the Apple's deman of G5 is too low. I think that the demand of the G5 will be very high. I schedule to order three G5 in automn and i am not alone.
2) My guess : no G5 in a mac or a powerbook before 90 nm process. The G5 tower 1,6 has a 450 VA supply. Remove an HD and 3 PCI slots add, a 50 watts supply for the screen, and you need a 400 VA for an i mac. Much more than the previous G4 tower.
The current G5 make a lot of heat compared to a G4 1 ghz, the mobo either, lot of heat also are produced, and need a huge cooling.
I say new I mac next year, with a brand new design and a G5 made upon a 90 nm process.
The immediate future of the Imac is the 7457, a better video card (like the 5200 pro). I expect them for Apple expo Paris. It would be a big disapointement if Apple cannot update his i mac line : the sales will drop dramatically.
exec 1:"well this sucks our plant isn't making as much money as we wanted"
exec 2:"hmm...what if we stop improving the 970 which is the one thing making us money from the plant, or maybe just drop it all together?"
exec 1:"sounds like a good business decision to me"
exec 2:"ok done and done, now...lets figure why no money is coming in from the fishkill plant anymore"
Just because you've bought the orthodoxy doesn't make it right, folks.
You're amusing little dialogue would be more accurate this way:
exec 1: "Well this sucks, our plant is making as much money as we want."
exec 2: "Yeah, it's a real drain on our profit margin. Shareholders are gonna get pissed if we keep throwing money down this hole with no return."
exec 1: "well, we can't abandon it. can we?"
exec 2: "No, we're in it for the long haul. But I would say since we're only squirting out a small volume of them 970's, that we cut back on making improvements until the global market improves. After all, our server products based on the 970 aren't in a market space where a couple MHz here or there matter. Our customers buy IBM products for thw whole kit and kaboodle, not just the latest GHz rating. We'll cut back on funding at Fishkill for a while until it's a better business proposition."
exec 1: "That makes sense. What about Apple, though? They are going to be pissed."
exec 2: "Since they aren't buying that many 970's, I think we can just live with their whining for now. What are they gonna do? Go to Motorola?
Both men start laughing. One of them laughs so hard he actually pees a little bit.
Look, I've read the boards here. I know how the orthodox opinion is made. I also know that many of you are Pollyanna-ish about your views w/r/t to what's going on. Which is your right, I guess.
exec 1: "That makes sense. What about Apple, though? They are going to be pissed."
exec 2: "Since they aren't buying that many 970's, I think we can just live with their whining for now. What are they gonna do? Go to Motorola?
I still don't see where you're getting the idea that Apple "isn't buying that many 970s" especially since the damn things aren't even shipping yet. Will it really kill you to wait for a new iMac like the rest of us without whining about it?
IBM isn't worried. IBM isn't going to "ramp down" anything if Apple doesn't release your precious little iMac soon. Get a grip.
I don't see why the G5 can't appear in the iMac until 90nm. Surely, it is possible to produce 1.4GHz-models, and lower, and the FSB multiplier is not fixed to 2:1. A 1.4Ghz G5 with a 466Mhz FSB (effectively 415Mhz, a 3:1 ratio) would require ( I dare say significantly) less power and cooling than the 1.6Ghz tower. Fit it with single-channel PC3200, and you have a very competive system, probably without needing the phattest cooling in the world. Hell, they use a 1Ghz 7455 in the iMac now, which requires some power too. Or a 1.2Ghz one with 333MHz DDR RAM.
In any case, it's a huge jump from the iMac G4, which undoubtably performes like crap for its price.
Apple may have a little impact on the Fishkill plants economy if they choose to use the G5 in as many product lines as possible, but they're nowhere near being big enough to be of any real significance. They're just a brick in a big puzzle.
Sorry, it just seems that simple. IBM is disappointed with 970 demand, and Apple can provide the product that would move more 970's.
...
Oh, yeah! The new G5 iMac! It will probably have to bigger than the dome because I think the G5 is a bit of a powerslut, right? Or at least throw in the mother of all fans or whatever. Just get it done, put it on the market, give IBM more 970 money, and everyone wins.
Am I missing something?
I for one agree with the premise of your idea. I'm not sure about the overcapacity problems, but yes, Apple needs more G5 machines, the sooner the better. Q4 2003 actually. And yes, you did miss some things:
1. All Power Macs should be duals
2. Powerbook G5
3. G5 mini
4. Xserve G5
All told, Apple could be buying 1+ million 130 nm chips per quarter from IBM. Keep in mind that numbers includes the system ASIC that will be in every G5 machine. So, in each dual 2 GHz PowerMac G5 machine, the 3 most important chips in the machine come from IBM's fab.
But the big problem is that Apple simply doesn't have the resources to concurrently design machines anymore. Maybe they do, but I recall they started consolidating the product design teams in the G4 days...
I was thinking the same thing, Id bet a downclocked G5 at 1-1.2 GHZ doesnt use much more juice than a 7457, and the performance should nearly double over the 1gHZ G4 say a P4 2.4. Enough for the consumers too browse their mail.
We'll cut back on funding at Fishkill for a while until it's a better business proposition."
exec 1: "That makes sense. What about Apple, though? They are going to be pissed."
Look, I've read the boards here. I know how the orthodox opinion is made. I also know that many of you are Pollyanna-ish about your views w/r/t to what's going on. Which is your right, I guess.
You are such an idiot that it is funny. I am going to remember your name and forever associate it with the tinfoil hat crowd.
Fishkill is IBM's most advanced FAB. It will produce the Power5. Fishkill is where NVidia gets its GPU's fabbed. If IBM cuts back on funding at fishkill (whatever you mean by this- it is a nonsense satetment!), then they will basically screw their business partners. Moto did that, but IBM has sufficient wits to realize that they should not alienate their customers WHEN IT IS THEIR BUSINESS TO RUN A FAB. Failing to support partners basically guarantees that you will not have a future business. Fishkill isn't going to loose any funding. The place cost billions to set up and you think that a multibillion dollar industry leader with serious CPU plans is going to bail out when the profits aren't as high as desired? How stupid are you? If I could reach my hands out of your monitor and slap some sense into you, I would.
I am not polyannaish about my views of the situataion. I am a hardened pragmatist when it comes to the business world. Fishkill is IBM's ticket to becoming a serious player in the rent-a-fab business. The 970 is a key component in IBM's low end server lineup. Neither the FAB or the chip are going anywhere bad because IBM has invested too much into them and has lots to gain if they both do well. You on the other hand come seem to come from the X-files worldview camp where every press release is really a coverup for some horrorific business catastrophe.
Okay I`m quite sure IBM is going to just say Hey we spent three billion dollars on this fab but its not making money so instead of trying to make improvements faster we will make them slower and lose even more money!
I`m sure Apple and IBM have been over this many times. Your not going to see G5 iMacs. The only reason it went G4 was because of the G5! Sorry guy it just isnt going to happen. Besides the fact at 3GHz in 12 months the iMac is going to scale just as slow as it would with a G4..
But the big problem is that Apple simply doesn't have the resources to concurrently design machines anymore. Maybe they do, but I recall they started consolidating the product design teams in the G4 days...
Another option is that Apple is aiming for another home run motherboard for the iMac. The mobo for the G5 is a serious home run folks, ok, it is more of a grand slam. It has EVERYTHING that you could want (except for the 6 PCI slot folks out there). My guess is that a future iMac mobo will also do Apple well for a few years.
But right now, the iMac doesn't need a G5. It needs a feature refresh and a price drop and it needs it badly. G5's aren't even in the hands of power users- it is a bit early to be demanding that they be in the hands of home users.
More likely is that Apple is waiting for the 0.09 micron processors before putting them into the iMac and portable machines. They have probably been planning this for some time and there is no possible way to re-arrange the production schedules just to take up some of the slack in IBM's capacity. Apple doesn't really care if IBM's plant is below capacity for the first quarter or two of its operation... in fact its better that way since there is no competition for capacity.
IBM isn't going to bail on a 3 billion dollar plant because demand was below expectations in the first quarter or two. It just means they have to work a little more at finding the demand for their capacity, and they might have to lower their rates a little to attract more interest. Shipping a 0.13 (and soon 0.09) 970 and various Apple and nVidia parts is good advertisement as well -- especially if the yields are good and clock rates over the estimates. This will attract more customers. The current economic conditions affect the picture as well, and when this plant first began construction several years ago they may have forcasted a faster recovery and thus more business potential.
This situation should be short lived, IBM knows it and they'll hang in there for a while before doing anything drastic... especially since their POWER and low end server chips are coming out of Fishkill as well.
Ahh, this is what I like about you guys.....you can see trees, but the forest still eludes you.
Simply put: if IBM doesn't make more money out of the new plant (where amongst other things, they make the 970), they may find the business not worthwhile to be in, or may find making better faster 970's on a relatively quick schedule not worth the investment. Improvements cost $$$$, and IBMmay be less interested in getting up to 3 GHz in 12 months or whatever if there's no business around it. For example, maybe they'd decide that the process improvement can wait another year or so, so the faster chips arrive in 2005-6, not 2004.....
Right now, every iMac that goes out the door gives Moto $$$$ and IBM zilch. I'm to understand that it is too early to put 970's or 970-like chips in portables yet, so the only thing remaining to generate orders thus profit thus interest for IBM is the eMac and iMac. Since the eMac is intended to be the most down-scale of the desktops, it would seem that a G5 iMac would be the best place to funnel more money towards sustaining the business at Fishkill.
But whatever. You are all far too smart to see that, since it wasn't on Thinksecret or Spymac.
IBM and Apple do not spend years developing a processor just for IBM to back out when initial G5 orders are not exactly what was projected. Trust me, right now IBM and Apple are looking at where they will be in late 2004 or early 2005. Companies this size don't react to right now but plan and prepare for down the road.
I don't see why the G5 can't appear in the iMac until 90nm. Surely, it is possible to produce 1.4GHz-models, and lower, and the FSB multiplier is not fixed to 2:1. A 1.4Ghz G5 with a 466Mhz FSB (effectively 415Mhz, a 3:1 ratio) would require ( I dare say significantly) less power and cooling than the 1.6Ghz tower. Fit it with single-channel PC3200, and you have a very competive system, probably without needing the phattest cooling in the world.
Yes, this would be a pretty good low end Mac but keep in mind these facts regarding Intel's mainstream chipset (865P "Springdale"):
- 533MHz (QDR) FSB support,
- dual channel DDR333 RAM support,
- Gigabit Ethernet channel
- AGP 8x
- 6 Channel sound
- USB 2
- PCI
Other versions of Springdale (865PE and 865G) support 800MHz FSB, dual DDR400, SATA RAID (also available as an option with 865P)... And then there are the Nvidia NForce2 Athlon mainstream offerings.
Apple needs to keep the 2:1 multiplier even in G5 iMacs below 1.4GHz. This would still allow the controller chip to run slower and be cheaper if the CPUs were <= 1.4 GHz. But if the BigMacs (sorry...) were bumped to the 2.0-2.6 GHz range in Jan and G5 iMacs were intro'ed then too they could start from 1.4-1.8 GHz with a 3:1 mult and get 600 MHz (533 MHz eff) at the top end.
Dual channel DDR333 support would be a must too. As would Gigabit Ethernet capability and a much improved graphics option.
Apple needs to remain competetive on features and price. Intel upped the ante across the board and Apple has matched or exceeded it at the top end. Now they need to do the same at the low end.
Yes, this would be a pretty good low end Mac but keep in mind these facts regarding Intel's mainstream chipset (865P "Springdale"):
- 533MHz (QDR) FSB support,
- dual channel DDR333 RAM support,
- Gigabit Ethernet channel
- AGP 8x
- 6 Channel sound
- USB 2
- PCI
Other versions of Springdale (865PE and 865G) support 800MHz FSB, dual DDR400, SATA RAID (also available as an option with 865P)... And then there are the Nvidia NForce2 Athlon mainstream offerings.
I agree with you, I was just trying to make an argument against those who say the G5 with its northbridge is too power-hungry or too hot for the iMac enclosure.
A 1.4Ghz *Mac (I'd like to see something Cube-like, but affordable ) with the FSB running at a 2:1-ratio, and functionality equivalent to that you mention would make a very competive machine at the lower end, no doubt. Whether it's too much for the iMac enclosure or not is maybe a good question. If it is, I believe Apple needs to do something drastic about it. Or wait for the 0.09 version...
Another option is that Apple is aiming for another home run motherboard for the iMac.
The basic iMac G5 architecture should be exactly the same as the Power Mac G5 architecture sans dual sockets, PCI/PCI-X slots, SATA, and chipset and high end memory (less DIMMs). Probably no optical audio either.
Essentially, hooking the K2 (Key Largo 2?) southbridge directly to the U3 northbridge, instead of through the PCI/PCI-X bridge.
Quote:
But right now, the iMac doesn't need a G5. It needs a feature refresh and a price drop and it needs it badly. G5's aren't even in the hands of power users- it is a bit early to be demanding that they be in the hands of home users.
Yes, either or. I think it needs to drop to a range from $1000 to $1500 with a 1.3 GHz 7457, or stick with the current price and add the 970 architecture.
And no, it's not demanding for the G5 to be avialable to home users. If Apple's PowerMac line up was dual 1.6, dual 1.8 or dual 2 GHz, 1 to 1.4 GHz G5 iMacs are perfectly fine. I still don't understand why Apple doesn't have an all dual Power Mac lineup. But even with the current lineup, there is still or could be a broad performance spectrum.
Lastly, a thing to note is that at 130 nm, 970 CPUs will have clock rates around 1.3 GHz on the low end while 7457 CPUs will have clock rates around 1.3 GHz on the high end. I hope people understand what this potentially means. A 1.3 GHz PPC 970 CPU will be cheaper than a 1.3 GHz 7457. It is to Apple's advantage to use the 970 in the next iMac if the above is true. Better performance, better marketability, and probably better profit margins.
Gee, I wonder if Apple has a contract with IBM to produce chips and upgrades over a period of time... Nah, Apple wouldn't do that, they are stupid and just flither away their stock holders money on engineering a new motherboard knowing it is a one-time board and they don't care if they use it again.
Ahh, some are still missing my point, I see. Well, good. Anyone who can't see that more money for 970's from Apple customers leads to better ROI for Fishkill hence more interest and funding in Fishkill deserves whatever deluded world they live in.
Improvements to the 970 are NOT automatic. They will require more money, money that IBM could take away if it feels like the ROI sucks. Furthermore, they could continue to support profitable, high volume businesses and not as much for less-so. Is the 970 going away? No. Will IBM spend what it takes to move the 970 up and beyond where it is now? Only if demand is strong enough. Is demand strong enough? Well.....more demand would nice, I'm sure. Now, anything Apple can do to increase deamnd for 970's? You all say no, but some of us say yes. Guess who will be right?
Will IBM spend what it takes to move the 970 up and beyond where it is now? Only if demand is strong enough. Is demand strong enough? Well.....more demand would nice, I'm sure. Now, anything Apple can do to increase deamnd for 970's? You all say no, but some of us say yes. Guess who will be right?
39 posts later, and you still haven't enlightened us with the answer to one simple question. Where are you getting this idea that the demand for the 970 isn't as high as it should be? Do you honestly think that Apple isn't working on a way to get the G5 into as many machines as possible, and that you're coming up with some great idea that noone in Cupertino has thought of yet?
Comments
And I guess this also means that IBM's fancy computerized design process in reality is actually two University of Michigan Computer Science teachers.
2) My guess : no G5 in a mac or a powerbook before 90 nm process. The G5 tower 1,6 has a 450 VA supply. Remove an HD and 3 PCI slots add, a 50 watts supply for the screen, and you need a 400 VA for an i mac. Much more than the previous G4 tower.
The current G5 make a lot of heat compared to a G4 1 ghz, the mobo either, lot of heat also are produced, and need a huge cooling.
I say new I mac next year, with a brand new design and a G5 made upon a 90 nm process.
The immediate future of the Imac is the 7457, a better video card (like the 5200 pro). I expect them for Apple expo Paris. It would be a big disapointement if Apple cannot update his i mac line : the sales will drop dramatically.
Originally posted by ast3r3x
exec 1:"well this sucks our plant isn't making as much money as we wanted"
exec 2:"hmm...what if we stop improving the 970 which is the one thing making us money from the plant, or maybe just drop it all together?"
exec 1:"sounds like a good business decision to me"
exec 2:"ok done and done, now...lets figure why no money is coming in from the fishkill plant anymore"
Just because you've bought the orthodoxy doesn't make it right, folks.
You're amusing little dialogue would be more accurate this way:
exec 1: "Well this sucks, our plant is making as much money as we want."
exec 2: "Yeah, it's a real drain on our profit margin. Shareholders are gonna get pissed if we keep throwing money down this hole with no return."
exec 1: "well, we can't abandon it. can we?"
exec 2: "No, we're in it for the long haul. But I would say since we're only squirting out a small volume of them 970's, that we cut back on making improvements until the global market improves. After all, our server products based on the 970 aren't in a market space where a couple MHz here or there matter. Our customers buy IBM products for thw whole kit and kaboodle, not just the latest GHz rating. We'll cut back on funding at Fishkill for a while until it's a better business proposition."
exec 1: "That makes sense. What about Apple, though? They are going to be pissed."
exec 2: "Since they aren't buying that many 970's, I think we can just live with their whining for now. What are they gonna do? Go to Motorola?
Both men start laughing. One of them laughs so hard he actually pees a little bit.
Look, I've read the boards here. I know how the orthodox opinion is made. I also know that many of you are Pollyanna-ish about your views w/r/t to what's going on. Which is your right, I guess.
Originally posted by I, Fred
...I also know that many of you are Pollyanna-ish about your views w/r/t to what's going on. Which is your right, I guess.
...and your view is what?
Oh, and thanks for restoring my rights.
Originally posted by I, Fred
exec 1: "That makes sense. What about Apple, though? They are going to be pissed."
exec 2: "Since they aren't buying that many 970's, I think we can just live with their whining for now. What are they gonna do? Go to Motorola?
I still don't see where you're getting the idea that Apple "isn't buying that many 970s" especially since the damn things aren't even shipping yet. Will it really kill you to wait for a new iMac like the rest of us without whining about it?
IBM isn't worried. IBM isn't going to "ramp down" anything if Apple doesn't release your precious little iMac soon. Get a grip.
In any case, it's a huge jump from the iMac G4, which undoubtably performes like crap for its price.
Apple may have a little impact on the Fishkill plants economy if they choose to use the G5 in as many product lines as possible, but they're nowhere near being big enough to be of any real significance. They're just a brick in a big puzzle.
Originally posted by I, Fred
Sorry, it just seems that simple. IBM is disappointed with 970 demand, and Apple can provide the product that would move more 970's.
...
Oh, yeah! The new G5 iMac! It will probably have to bigger than the dome because I think the G5 is a bit of a powerslut, right? Or at least throw in the mother of all fans or whatever. Just get it done, put it on the market, give IBM more 970 money, and everyone wins.
Am I missing something?
I for one agree with the premise of your idea. I'm not sure about the overcapacity problems, but yes, Apple needs more G5 machines, the sooner the better. Q4 2003 actually. And yes, you did miss some things:
1. All Power Macs should be duals
2. Powerbook G5
3. G5 mini
4. Xserve G5
All told, Apple could be buying 1+ million 130 nm chips per quarter from IBM. Keep in mind that numbers includes the system ASIC that will be in every G5 machine. So, in each dual 2 GHz PowerMac G5 machine, the 3 most important chips in the machine come from IBM's fab.
But the big problem is that Apple simply doesn't have the resources to concurrently design machines anymore. Maybe they do, but I recall they started consolidating the product design teams in the G4 days...
Originally posted by I, Fred
We'll cut back on funding at Fishkill for a while until it's a better business proposition."
exec 1: "That makes sense. What about Apple, though? They are going to be pissed."
Look, I've read the boards here. I know how the orthodox opinion is made. I also know that many of you are Pollyanna-ish about your views w/r/t to what's going on. Which is your right, I guess.
You are such an idiot that it is funny. I am going to remember your name and forever associate it with the tinfoil hat crowd.
Fishkill is IBM's most advanced FAB. It will produce the Power5. Fishkill is where NVidia gets its GPU's fabbed. If IBM cuts back on funding at fishkill (whatever you mean by this- it is a nonsense satetment!), then they will basically screw their business partners. Moto did that, but IBM has sufficient wits to realize that they should not alienate their customers WHEN IT IS THEIR BUSINESS TO RUN A FAB. Failing to support partners basically guarantees that you will not have a future business. Fishkill isn't going to loose any funding. The place cost billions to set up and you think that a multibillion dollar industry leader with serious CPU plans is going to bail out when the profits aren't as high as desired? How stupid are you? If I could reach my hands out of your monitor and slap some sense into you, I would.
I am not polyannaish about my views of the situataion. I am a hardened pragmatist when it comes to the business world. Fishkill is IBM's ticket to becoming a serious player in the rent-a-fab business. The 970 is a key component in IBM's low end server lineup. Neither the FAB or the chip are going anywhere bad because IBM has invested too much into them and has lots to gain if they both do well. You on the other hand come seem to come from the X-files worldview camp where every press release is really a coverup for some horrorific business catastrophe.
I`m sure Apple and IBM have been over this many times. Your not going to see G5 iMacs. The only reason it went G4 was because of the G5! Sorry guy it just isnt going to happen. Besides the fact at 3GHz in 12 months the iMac is going to scale just as slow as it would with a G4..
Originally posted by THT
But the big problem is that Apple simply doesn't have the resources to concurrently design machines anymore. Maybe they do, but I recall they started consolidating the product design teams in the G4 days...
Another option is that Apple is aiming for another home run motherboard for the iMac. The mobo for the G5 is a serious home run folks, ok, it is more of a grand slam. It has EVERYTHING that you could want (except for the 6 PCI slot folks out there). My guess is that a future iMac mobo will also do Apple well for a few years.
But right now, the iMac doesn't need a G5. It needs a feature refresh and a price drop and it needs it badly. G5's aren't even in the hands of power users- it is a bit early to be demanding that they be in the hands of home users.
IBM isn't going to bail on a 3 billion dollar plant because demand was below expectations in the first quarter or two. It just means they have to work a little more at finding the demand for their capacity, and they might have to lower their rates a little to attract more interest. Shipping a 0.13 (and soon 0.09) 970 and various Apple and nVidia parts is good advertisement as well -- especially if the yields are good and clock rates over the estimates. This will attract more customers. The current economic conditions affect the picture as well, and when this plant first began construction several years ago they may have forcasted a faster recovery and thus more business potential.
This situation should be short lived, IBM knows it and they'll hang in there for a while before doing anything drastic... especially since their POWER and low end server chips are coming out of Fishkill as well.
Originally posted by I, Fred
Ahh, this is what I like about you guys.....you can see trees, but the forest still eludes you.
Simply put: if IBM doesn't make more money out of the new plant (where amongst other things, they make the 970), they may find the business not worthwhile to be in, or may find making better faster 970's on a relatively quick schedule not worth the investment. Improvements cost $$$$, and IBMmay be less interested in getting up to 3 GHz in 12 months or whatever if there's no business around it. For example, maybe they'd decide that the process improvement can wait another year or so, so the faster chips arrive in 2005-6, not 2004.....
Right now, every iMac that goes out the door gives Moto $$$$ and IBM zilch. I'm to understand that it is too early to put 970's or 970-like chips in portables yet, so the only thing remaining to generate orders thus profit thus interest for IBM is the eMac and iMac. Since the eMac is intended to be the most down-scale of the desktops, it would seem that a G5 iMac would be the best place to funnel more money towards sustaining the business at Fishkill.
But whatever. You are all far too smart to see that, since it wasn't on Thinksecret or Spymac.
IBM and Apple do not spend years developing a processor just for IBM to back out when initial G5 orders are not exactly what was projected. Trust me, right now IBM and Apple are looking at where they will be in late 2004 or early 2005. Companies this size don't react to right now but plan and prepare for down the road.
Originally posted by Zapchud
I don't see why the G5 can't appear in the iMac until 90nm. Surely, it is possible to produce 1.4GHz-models, and lower, and the FSB multiplier is not fixed to 2:1. A 1.4Ghz G5 with a 466Mhz FSB (effectively 415Mhz, a 3:1 ratio) would require ( I dare say significantly) less power and cooling than the 1.6Ghz tower. Fit it with single-channel PC3200, and you have a very competive system, probably without needing the phattest cooling in the world.
Yes, this would be a pretty good low end Mac but keep in mind these facts regarding Intel's mainstream chipset (865P "Springdale"):
- 533MHz (QDR) FSB support,
- dual channel DDR333 RAM support,
- Gigabit Ethernet channel
- AGP 8x
- 6 Channel sound
- USB 2
- PCI
Other versions of Springdale (865PE and 865G) support 800MHz FSB, dual DDR400, SATA RAID (also available as an option with 865P)... And then there are the Nvidia NForce2 Athlon mainstream offerings.
Apple needs to keep the 2:1 multiplier even in G5 iMacs below 1.4GHz. This would still allow the controller chip to run slower and be cheaper if the CPUs were <= 1.4 GHz. But if the BigMacs (sorry...) were bumped to the 2.0-2.6 GHz range in Jan and G5 iMacs were intro'ed then too they could start from 1.4-1.8 GHz with a 3:1 mult and get 600 MHz (533 MHz eff) at the top end.
Dual channel DDR333 support would be a must too. As would Gigabit Ethernet capability and a much improved graphics option.
Apple needs to remain competetive on features and price. Intel upped the ante across the board and Apple has matched or exceeded it at the top end. Now they need to do the same at the low end.
MM
Originally posted by MartianMatt
Yes, this would be a pretty good low end Mac but keep in mind these facts regarding Intel's mainstream chipset (865P "Springdale"):
- 533MHz (QDR) FSB support,
- dual channel DDR333 RAM support,
- Gigabit Ethernet channel
- AGP 8x
- 6 Channel sound
- USB 2
- PCI
Other versions of Springdale (865PE and 865G) support 800MHz FSB, dual DDR400, SATA RAID (also available as an option with 865P)... And then there are the Nvidia NForce2 Athlon mainstream offerings.
I agree with you, I was just trying to make an argument against those who say the G5 with its northbridge is too power-hungry or too hot for the iMac enclosure.
A 1.4Ghz *Mac (I'd like to see something Cube-like, but affordable
Originally posted by Yevgeny
Another option is that Apple is aiming for another home run motherboard for the iMac.
The basic iMac G5 architecture should be exactly the same as the Power Mac G5 architecture sans dual sockets, PCI/PCI-X slots, SATA, and chipset and high end memory (less DIMMs). Probably no optical audio either.
Essentially, hooking the K2 (Key Largo 2?) southbridge directly to the U3 northbridge, instead of through the PCI/PCI-X bridge.
But right now, the iMac doesn't need a G5. It needs a feature refresh and a price drop and it needs it badly. G5's aren't even in the hands of power users- it is a bit early to be demanding that they be in the hands of home users.
Yes, either or. I think it needs to drop to a range from $1000 to $1500 with a 1.3 GHz 7457, or stick with the current price and add the 970 architecture.
And no, it's not demanding for the G5 to be avialable to home users. If Apple's PowerMac line up was dual 1.6, dual 1.8 or dual 2 GHz, 1 to 1.4 GHz G5 iMacs are perfectly fine. I still don't understand why Apple doesn't have an all dual Power Mac lineup. But even with the current lineup, there is still or could be a broad performance spectrum.
Lastly, a thing to note is that at 130 nm, 970 CPUs will have clock rates around 1.3 GHz on the low end while 7457 CPUs will have clock rates around 1.3 GHz on the high end. I hope people understand what this potentially means. A 1.3 GHz PPC 970 CPU will be cheaper than a 1.3 GHz 7457. It is to Apple's advantage to use the 970 in the next iMac if the above is true. Better performance, better marketability, and probably better profit margins.
Get real, get-a-life.
Improvements to the 970 are NOT automatic. They will require more money, money that IBM could take away if it feels like the ROI sucks. Furthermore, they could continue to support profitable, high volume businesses and not as much for less-so. Is the 970 going away? No. Will IBM spend what it takes to move the 970 up and beyond where it is now? Only if demand is strong enough. Is demand strong enough? Well.....more demand would nice, I'm sure. Now, anything Apple can do to increase deamnd for 970's? You all say no, but some of us say yes. Guess who will be right?
Originally posted by I, Fred
Will IBM spend what it takes to move the 970 up and beyond where it is now? Only if demand is strong enough. Is demand strong enough? Well.....more demand would nice, I'm sure. Now, anything Apple can do to increase deamnd for 970's? You all say no, but some of us say yes. Guess who will be right?
39 posts later, and you still haven't enlightened us with the answer to one simple question. Where are you getting this idea that the demand for the 970 isn't as high as it should be? Do you honestly think that Apple isn't working on a way to get the G5 into as many machines as possible, and that you're coming up with some great idea that noone in Cupertino has thought of yet?