Graphic cards to do 3d rendering? As a CPU co processor?</strong><hr></blockquote>
The CPU will stay the "central" processor -- it has control and it tells the GPU what to do. The GPU is just much faster (at certain things). I had to laugh at the Pixar guy back when he made his statement two years ago... even then it seemed obvious to me that he was being increadibly short sighted. These GPU's aren't restricted to doing real-time calculations...!
[quote]<strong>
Given that current x86 chips can barely feed the Radeon 9700...aren't we going to need a Powerlite CPU to help the Mac feed the GPU?
Or might Hyper/Rio help in this respect?
</strong><hr></blockquote>
Smarter programmers will make an even bigger difference. Faster CPUs and more bandwidth is good, but there is plenty of untapped potential because developers (by and large) aren't very savvy about how to use this kind of GPU hardware. There have been limitations in the OpenGL and D3D APIs as well, preventing the applications and drivers from performing at peak efficiency... but solutions to those problems are here now. Depending on the kinds of things being drawn, VRAM can be traded for bandwidth.
[quote]<strong>
The perception of current machines as 'stop gap'. Thought. New cards work on Agp x8? Firewire 2 should be ready to do early next year. Surely a San Fran' refresh is called for?
</strong><hr></blockquote>
Another way to look at it is that all machines are "stop gap" until the next architecture comes along. In this light these new machines are the ultimate (or penultimate if there is a significant revision early next year) development of the PowerMac G4.
AGP 8x might be nice, but really the current machines fully utilize their available memory bandwidth. DDR333 is 2.7 GB/sec (theoretical). AGP 4x is about 1 GB/sec. MPX 166 is about 1.3 GB/sec. Ethernet, ATA, PCI add up to at least 0.5 GB/sec. This adds up to roughly 2.7 GB/sec... so if you doubled the AGP bandwidth you would just be taking it away from something else. Stealing most of the I/O bandwidth is probably okay, but you don't want to start cutting into the MPX so AGP 8x in the current machines is bordering on overkill. In machines based on the coming IBM chip, however, there is 6.4 GB/sec of CPU bandwidth so AGP 8x will be feeling a little slow! (especially if DDR-II arrives on schedule)
FireWire2 is good for Apple to bring to market just to encourage industry adoption... but for the early consumers it isn't going to matter because Apple will get there first and there won't be anything to plug into it (that wouldn't also work with FW1).
I doubt the expected G4 refresh early next year will bring any surprises. Apple will be stacking the deck for its G5...
People seem to believe that the 1100 MHz G4 I wrote about above simply has a 100 MHz MPX bus (since 100x11 = 1100) instead of a new, faster bus (it can't be a 133 MHz bus or a 167 MHz bus if we go by the standard "multiplier" logic).
I can accept that. However, this leads to a second question. In another thread, I believe it was about 1.4 GHz G4's, Bigc wrote that the new 1.25 GHz G4 is basically the 1 GHz G4 with a faster bus since 167x7.5 = 1250 and 133x7.5 = 1000 (note: this is different than the "overclocking" argument going around). Does this work? Can you take a 1 GHz G4 with a 133 MHz bus, put a faster bus on it, and make the chip faster?
If this is true (although perhaps somewhat oversimplified), does this imply that, if Motorola can make a 1.1 GHz G4 with a 100 MHz bus (x11), then they could potentially make a 1.467 GHz G4 with a 133 MHz bus (x11) and a 1.833 GHz G4 with a 167 MHz bus (x11).
Basically, what I'm asking is: does that speed of the chip increase if you increase the speed of the bus? and can you make a chip faster just by putting a faster bus on it?
Another question I've had is: how hard would it be to design multiprocessor systems where the two processors are different?
I'm thinking that Apple might want to make a dual-processor system where one processor is the new, Power4-derived, desktop chip that IBM is working on (or even a Sahara G3), and the second processor is a pure-Altivec processor. How easily could this be done?
The advantage of doing this is that Apple protects its investment in Altivec code even if IBM or Motorola start using a non-Altivec compatible SIMD, stop supporting SIMDs altogether, or "cripple" Altivec by making it share registers inefficiently. I'm also thinking that Apple might have to rewrite parts of the OS/kernel to make this work, but that most programmers won't have to update their code to make their applications run on these new machines.
Maybe these "pure-Altivec" chips will be designed by Nvidia and fabbed by AMD*. These chips may also be the DSP/helper chips that I believe Moki and Dorsal were talking about a few months ago. Maybe Apple knew people would be suspicious when they saw these additional chips on their experimental motherboards so they said that they were specialized DSP chips.
*Sorry, I don't really believe in a super-secret Apple/Nvidia/AMD consortium (or that Apple is necessarily working on any of this), but I felt that I would throw that last part out there just to stir things up a little.
Yes...typically you increase the bus speed and depending on what your multiplier is the processors adjusts accordingly. That's the theory I believe.
As for Altivec the consensus seems to be IBM has every right an license to use Altivec(VMX). So we may find ourselves in a win/win scenario with IBM for the near future assuming the mantle for G5 production.
I don't buy the Apple/nvidia stuff either. Nvidia is in a dogfight with ATI right now and somehow I don't think they're devoting THAT many resources to Apple. I don't get the optimism over nForce. It's a good chipset but it's now JAW DROPPING nor does it obviate ones desire for discrete components IMO.
<strong>I can accept that. However, this leads to a second question. In another thread, I believe it was about 1.4 GHz G4's, Bigc wrote that the new 1.25 GHz G4 is basically the 1 GHz G4 with a faster bus since 167x7.5 = 1250 and 133x7.5 = 1000 (note: this is different than the "overclocking" argument going around). Does this work? Can you take a 1 GHz G4 with a 133 MHz bus, put a faster bus on it, and make the chip faster?
If this is true (although perhaps somewhat oversimplified), does this imply that, if Motorola can make a 1.1 GHz G4 with a 100 MHz bus (x11), then they could potentially make a 1.467 GHz G4 with a 133 MHz bus (x11) and a 1.833 GHz G4 with a 167 MHz bus (x11).
Basically, what I'm asking is: does that speed of the chip increase if you increase the speed of the bus? and can you make a chip faster just by putting a faster bus on it?</strong><hr></blockquote>
You can try to feed the processor whatever clock signal you want... but it may not run. These chips are tested by the manufacturer when they are produced and rated according to the speed that the individual chip can sustain... I imagine that Moto rates the chips in terms of both bus clock and internal clock.
The bus clock is fed into the chip and it has a clock multiplier built into it that has a bunch of different possible multipliers, one of which is selected by a register in the chip. Depending on the setup, either software selects one of these multipliers or it is based on a few external lines which are usually hardware wired high/low to get the right multiplier setting (some upgrade makers use a set of dip-switches to let the user change the multiplier). Setting a higher multiplier will cause the chip to run faster, but it may not operate correctly.
<strong>Sorry folks. No POWER5 until 2004. Motorola has also ceased G4 development putting Apple in a real pickle. Apple has two options:
Quad machines
or
Jump on the PowerQUICC III bandwagon.</strong><hr></blockquote>
Sorry I don't believe that at all. Things point to IBM aggresively persuing new markets with their shiny new Fab Toy. Motorola won't cease G4 development because the procs are used in other areas like Cisco routers and more.
I think we see late summer MWNY 03 for this new proc at the latest.
<strong>Sorry folks. No POWER5 until 2004. Motorola has also ceased G4 development putting Apple in a real pickle. Apple has two options:
Quad machines
or
Jump on the PowerQUICC III bandwagon.</strong><hr></blockquote>
I think I've mentioned this before, but here goes again.....
Yes. The POWER5 chip is supposedly due out in early 2004, with the "workstation" version due mid-year (if the Register can be believed).
Now, what does that mean for the POWER4lite being announced in Oct? Could IBM recoup their R&D in a year or less?
Somehow that makes me think the POWER4lite will be sooner rather than later....ie it's taped out and ready for sampling now.....and before flaming, I said sampling, not full production.
<strong>I think I've mentioned this before, but here goes again.....
Yes. The POWER5 chip is supposedly due out in early 2004, with the "workstation" version due mid-year (if the Register can be believed).
Now, what does that mean for the POWER4lite being announced in Oct? Could IBM recoup their R&D in a year or less?
Somehow that makes me think the POWER4lite will be sooner rather than later....ie it's taped out and ready for sampling now.....and before flaming, I said sampling, not full production. </strong><hr></blockquote>
Apple wouldn't be using the POWER5 in its machines anyway, unless it decided to get into ultra-high end computing.
Apple will be looking to use the new die-shrunk POWER4 (aka GP-UL), not the POWER5.
"Apple will be looking to use the new die-shrunk POWER4 (aka GP-UL), not the POWER5."
GP-UL.
Well, it'll be Jan' 2004 at the latest. I think Summer 2003 more likely. Especially as a Geforce 5 and FW2 should be with us by then. Nvidia is already using Hypertransport. AMD will ship their 64 bit solution by early next year. Do we really believe Apple will deliver their's a whole year later? I don't. I think Pinot will be 64 bit ready and a Summer 'G5' release. 'X' only. Improved Classic support? Only.
I think it's perfectly reasonable we'll get some 7470 or 7500 'G4' chips before then.
Maybe a 7470 or 7500 chip in Jan'. eg 7470 on proper DDR.
Summer could be 7500 on Rio?
Thing is. Are Apple going Hyper or Rio? Are IBM going Rio or Hyper. I would have thought Rio. That would indicate a 7500 on Rio being Moto's last chip to take us through to next Summer...and IBM early 2004. Annoying.
Do Moto' think they can get the G4 2 gig and higher? Maybe they can. But I say they can sodding keep it...because it'll take them a couple of years to get there from now...
Maybe that's Apple's plan for the consumer space. But they need a beefier 'performance' leveller...and an aggressive launch of the IBM chip is what the 'power'Mac doctor ordered. Sales indicate something needs to be done.
Jan' 2004 = IBM chip on Hypertransport?
What I'm confused about...I'd bet on the Powerlite coming a year, at the latest, after its annoucement at the Oct Microprocessor Forum. ie this time next year ahead of those 'storming' G4 1.25 gigs.
Will we really see a 7470 at 1.5 gig-ish on the same 'hack serve' board?
Where does this leave the 7500? If Apple are going Hypertransport then why bother using the 7500? Unless the IBM chip is on Rio...then maybe they'll squeeze the 7500 for next summer? Spring?
More to the point. The 7500 chip was removed from the Moto' website. I always thought this was supposed to the 'G5'. So maybe this is the chip that was 'cancelled' due toheat problems/or... and 'all' we'll get is a 7470 at 1.5 ish on 'hackserve' until IBM's G5 hits town next Summer.
The current set up looks like a 'penultimate' topping out of the G4 to me. The G4 was supposed to top out at its current speed. No?
...and it looks like it has topped out. Maybe they can eek one more revision out of it before it goes into the consumer line?
Apple's consumer line is already heavily using it.
I pray that if there is such a thing...that the 7470 G4 is the last revision before the 'Power'Mac returns...
I don't see why Apple can't continue to use Moto for the low-end machines like the iMac and laptops. After all, their designs do run cool, and I'm not sure there's room in the iMac base for the greater cooling that this new IBM chip could very easily require...
Okay, maybe I'm too optimistic, but why are many pushing the date so far out to get the IBM G5? Isn't it very possible that Apple realized long ago that Motorola would not be the most reliable supplier? Wouldn't Apple think twice about who should build the G5? Work could have started well over a year ago, maybe two. And with both Apple and nVidia as charter members of HyperTransport, wouldn't Apple be asking for such a bus on the G5? I believe samples of the G5 could have been undergoing tests for over six month already. I think it reasonable that IBM would delay talking about the G5, until it is about ready for production fab. I don't say that my opinion is right, but that it has just as good a chance of being right as those who expect a G5 at MWNY or later. I may be disappointed after October 15 or after MWSF, but I hope not.
To me, the new PowerMac case is a give away. It appears to be for higher power chips. It was introduced now for two reasons possibly. 1) It gives the impression that Apple is doing more to the PowerMac line now, when sales are hurting. 2) It makes for a quicker and easier introduction of the G5 when it is ready. There will be no new tooling for a new case. Sure, many would like to see some super nifty case for the G5, but I'll bet they will be happy just to get the G5 in today's case.
<strong>Okay, maybe I'm too optimistic, but why are many pushing the date so far out to get the IBM G5? Isn't it very possible that Apple realized long ago that Motorola would not be the most reliable supplier? Wouldn't Apple think twice about who should build the G5? Work could have started well over a year ago, maybe two. And with both Apple and nVidia as charter members of HyperTransport, wouldn't Apple be asking for such a bus on the G5? I believe samples of the G5 could have been undergoing tests for over six month already. I think it reasonable that IBM would delay talking about the G5, until it is about ready for production fab. I don't say that my opinion is right, but that it has just as good a chance of being right as those who expect a G5 at MWNY or later. I may be disappointed after October 15 or after MWSF, but I hope not. </strong><hr></blockquote>Just for reference, the Sahara was introduced at the Microprocessor Forum last October, and then got into iBooks 7 months later this past May. At the conference, they said the Sahara would begin sampling in January. So we'll have to see when they say this new Power4 begins sampling - that may give us an idea as to when Apple could use it.
Comments
<strong>
Heh. Autodesk may begin rethinking their non-Mac policy starting with Max?
Lemon Bon Bon</strong><hr></blockquote>
Even though they finally decide to make MAX away from being Windows exclusive, they will go Linux first.
<strong>
Graphic cards to do 3d rendering? As a CPU co processor?</strong><hr></blockquote>
The CPU will stay the "central" processor -- it has control and it tells the GPU what to do. The GPU is just much faster (at certain things). I had to laugh at the Pixar guy back when he made his statement two years ago... even then it seemed obvious to me that he was being increadibly short sighted. These GPU's aren't restricted to doing real-time calculations...!
[quote]<strong>
Given that current x86 chips can barely feed the Radeon 9700...aren't we going to need a Powerlite CPU to help the Mac feed the GPU?
Or might Hyper/Rio help in this respect?
</strong><hr></blockquote>
Smarter programmers will make an even bigger difference. Faster CPUs and more bandwidth is good, but there is plenty of untapped potential because developers (by and large) aren't very savvy about how to use this kind of GPU hardware. There have been limitations in the OpenGL and D3D APIs as well, preventing the applications and drivers from performing at peak efficiency... but solutions to those problems are here now. Depending on the kinds of things being drawn, VRAM can be traded for bandwidth.
[quote]<strong>
The perception of current machines as 'stop gap'. Thought. New cards work on Agp x8? Firewire 2 should be ready to do early next year. Surely a San Fran' refresh is called for?
</strong><hr></blockquote>
Another way to look at it is that all machines are "stop gap" until the next architecture comes along. In this light these new machines are the ultimate (or penultimate if there is a significant revision early next year) development of the PowerMac G4.
AGP 8x might be nice, but really the current machines fully utilize their available memory bandwidth. DDR333 is 2.7 GB/sec (theoretical). AGP 4x is about 1 GB/sec. MPX 166 is about 1.3 GB/sec. Ethernet, ATA, PCI add up to at least 0.5 GB/sec. This adds up to roughly 2.7 GB/sec... so if you doubled the AGP bandwidth you would just be taking it away from something else. Stealing most of the I/O bandwidth is probably okay, but you don't want to start cutting into the MPX so AGP 8x in the current machines is bordering on overkill. In machines based on the coming IBM chip, however, there is 6.4 GB/sec of CPU bandwidth so AGP 8x will be feeling a little slow! (especially if DDR-II arrives on schedule)
FireWire2 is good for Apple to bring to market just to encourage industry adoption... but for the early consumers it isn't going to matter because Apple will get there first and there won't be anything to plug into it (that wouldn't also work with FW1).
I doubt the expected G4 refresh early next year will bring any surprises. Apple will be stacking the deck for its G5...
[ 08-18-2002: Message edited by: Programmer ]</p>
IBM produces P5 for Apple.
Apple licences OSX for IBM.
I can accept that. However, this leads to a second question. In another thread, I believe it was about 1.4 GHz G4's, Bigc wrote that the new 1.25 GHz G4 is basically the 1 GHz G4 with a faster bus since 167x7.5 = 1250 and 133x7.5 = 1000 (note: this is different than the "overclocking" argument going around). Does this work? Can you take a 1 GHz G4 with a 133 MHz bus, put a faster bus on it, and make the chip faster?
If this is true (although perhaps somewhat oversimplified), does this imply that, if Motorola can make a 1.1 GHz G4 with a 100 MHz bus (x11), then they could potentially make a 1.467 GHz G4 with a 133 MHz bus (x11) and a 1.833 GHz G4 with a 167 MHz bus (x11).
Basically, what I'm asking is: does that speed of the chip increase if you increase the speed of the bus? and can you make a chip faster just by putting a faster bus on it?
[ 08-18-2002: Message edited by: cecil ]</p>
I'm thinking that Apple might want to make a dual-processor system where one processor is the new, Power4-derived, desktop chip that IBM is working on (or even a Sahara G3), and the second processor is a pure-Altivec processor. How easily could this be done?
The advantage of doing this is that Apple protects its investment in Altivec code even if IBM or Motorola start using a non-Altivec compatible SIMD, stop supporting SIMDs altogether, or "cripple" Altivec by making it share registers inefficiently. I'm also thinking that Apple might have to rewrite parts of the OS/kernel to make this work, but that most programmers won't have to update their code to make their applications run on these new machines.
Maybe these "pure-Altivec" chips will be designed by Nvidia and fabbed by AMD*. These chips may also be the DSP/helper chips that I believe Moki and Dorsal were talking about a few months ago. Maybe Apple knew people would be suspicious when they saw these additional chips on their experimental motherboards so they said that they were specialized DSP chips.
*Sorry, I don't really believe in a super-secret Apple/Nvidia/AMD consortium (or that Apple is necessarily working on any of this), but I felt that I would throw that last part out there just to stir things up a little.
[ 08-18-2002: Message edited by: cecil ]</p>
As for Altivec the consensus seems to be IBM has every right an license to use Altivec(VMX). So we may find ourselves in a win/win scenario with IBM for the near future assuming the mantle for G5 production.
I don't buy the Apple/nvidia stuff either. Nvidia is in a dogfight with ATI right now and somehow I don't think they're devoting THAT many resources to Apple. I don't get the optimism over nForce. It's a good chipset but it's now JAW DROPPING nor does it obviate ones desire for discrete components IMO.
<strong>I can accept that. However, this leads to a second question. In another thread, I believe it was about 1.4 GHz G4's, Bigc wrote that the new 1.25 GHz G4 is basically the 1 GHz G4 with a faster bus since 167x7.5 = 1250 and 133x7.5 = 1000 (note: this is different than the "overclocking" argument going around). Does this work? Can you take a 1 GHz G4 with a 133 MHz bus, put a faster bus on it, and make the chip faster?
If this is true (although perhaps somewhat oversimplified), does this imply that, if Motorola can make a 1.1 GHz G4 with a 100 MHz bus (x11), then they could potentially make a 1.467 GHz G4 with a 133 MHz bus (x11) and a 1.833 GHz G4 with a 167 MHz bus (x11).
Basically, what I'm asking is: does that speed of the chip increase if you increase the speed of the bus? and can you make a chip faster just by putting a faster bus on it?</strong><hr></blockquote>
You can try to feed the processor whatever clock signal you want... but it may not run. These chips are tested by the manufacturer when they are produced and rated according to the speed that the individual chip can sustain... I imagine that Moto rates the chips in terms of both bus clock and internal clock.
The bus clock is fed into the chip and it has a clock multiplier built into it that has a bunch of different possible multipliers, one of which is selected by a register in the chip. Depending on the setup, either software selects one of these multipliers or it is based on a few external lines which are usually hardware wired high/low to get the right multiplier setting (some upgrade makers use a set of dip-switches to let the user change the multiplier). Setting a higher multiplier will cause the chip to run faster, but it may not operate correctly.
[quote]Originally posted by Stecs:
<strong>Just a thought..
IBM produces P5 for Apple.
Apple licences OSX for IBM.</strong><hr></blockquote>
<strong>Just a thought..
IBM produces P5 for Apple.
Apple licences OSX for IBM.</strong><hr></blockquote>
once again, I think this is what is coming. i second the previous post.
Quad machines
or
Jump on the PowerQUICC III bandwagon.
<strong>Sorry folks. No POWER5 until 2004. Motorola has also ceased G4 development putting Apple in a real pickle. Apple has two options:
Quad machines
or
Jump on the PowerQUICC III bandwagon.</strong><hr></blockquote>
Sorry I don't believe that at all. Things point to IBM aggresively persuing new markets with their shiny new Fab Toy. Motorola won't cease G4 development because the procs are used in other areas like Cisco routers and more.
I think we see late summer MWNY 03 for this new proc at the latest.
<strong>Sorry folks. No POWER5 until 2004. Motorola has also ceased G4 development putting Apple in a real pickle. Apple has two options:
Quad machines
or
Jump on the PowerQUICC III bandwagon.</strong><hr></blockquote>
I think I've mentioned this before, but here goes again.....
Yes. The POWER5 chip is supposedly due out in early 2004, with the "workstation" version due mid-year (if the Register can be believed).
Now, what does that mean for the POWER4lite being announced in Oct? Could IBM recoup their R&D in a year or less?
Somehow that makes me think the POWER4lite will be sooner rather than later....ie it's taped out and ready for sampling now.....and before flaming, I said sampling, not full production.
<strong>I think I've mentioned this before, but here goes again.....
Yes. The POWER5 chip is supposedly due out in early 2004, with the "workstation" version due mid-year (if the Register can be believed).
Now, what does that mean for the POWER4lite being announced in Oct? Could IBM recoup their R&D in a year or less?
Somehow that makes me think the POWER4lite will be sooner rather than later....ie it's taped out and ready for sampling now.....and before flaming, I said sampling, not full production.
Apple wouldn't be using the POWER5 in its machines anyway, unless it decided to get into ultra-high end computing.
Apple will be looking to use the new die-shrunk POWER4 (aka GP-UL), not the POWER5.
<strong>
Apple wouldn't be using the POWER5 in its machines anyway, unless it decided to get into ultra-high end computing.
Apple will be looking to use the new die-shrunk POWER4 (aka GP-UL), not the POWER5.</strong><hr></blockquote>
Yep, that'd be roughly what I expect as well. But if the stories are to be believed, there's also a POWER5 desktop in the works, for 'bout Sept 2004.
Still makes me think that's an awfully short time for IBM to get the money back out.
GP-UL.
Well, it'll be Jan' 2004 at the latest. I think Summer 2003 more likely. Especially as a Geforce 5 and FW2 should be with us by then. Nvidia is already using Hypertransport. AMD will ship their 64 bit solution by early next year. Do we really believe Apple will deliver their's a whole year later? I don't. I think Pinot will be 64 bit ready and a Summer 'G5' release. 'X' only. Improved Classic support? Only.
I think it's perfectly reasonable we'll get some 7470 or 7500 'G4' chips before then.
Maybe a 7470 or 7500 chip in Jan'. eg 7470 on proper DDR.
Summer could be 7500 on Rio?
Thing is. Are Apple going Hyper or Rio? Are IBM going Rio or Hyper. I would have thought Rio. That would indicate a 7500 on Rio being Moto's last chip to take us through to next Summer...and IBM early 2004. Annoying.
Do Moto' think they can get the G4 2 gig and higher? Maybe they can. But I say they can sodding keep it...because it'll take them a couple of years to get there from now...
Maybe that's Apple's plan for the consumer space. But they need a beefier 'performance' leveller...and an aggressive launch of the IBM chip is what the 'power'Mac doctor ordered. Sales indicate something needs to be done.
Jan' 2004 = IBM chip on Hypertransport?
What I'm confused about...I'd bet on the Powerlite coming a year, at the latest, after its annoucement at the Oct Microprocessor Forum. ie this time next year ahead of those 'storming' G4 1.25 gigs.
Will we really see a 7470 at 1.5 gig-ish on the same 'hack serve' board?
Where does this leave the 7500? If Apple are going Hypertransport then why bother using the 7500? Unless the IBM chip is on Rio...then maybe they'll squeeze the 7500 for next summer? Spring?
More to the point. The 7500 chip was removed from the Moto' website. I always thought this was supposed to the 'G5'. So maybe this is the chip that was 'cancelled' due toheat problems/or... and 'all' we'll get is a 7470 at 1.5 ish on 'hackserve' until IBM's G5 hits town next Summer.
The current set up looks like a 'penultimate' topping out of the G4 to me. The G4 was supposed to top out at its current speed. No?
...and it looks like it has topped out. Maybe they can eek one more revision out of it before it goes into the consumer line?
Apple's consumer line is already heavily using it.
I pray that if there is such a thing...that the 7470 G4 is the last revision before the 'Power'Mac returns...
Lemon Bon Bon <img src="graemlins/hmmm.gif" border="0" alt="[Hmmm]" />
I did say 'if'
Lemon Bon Bon
(Anybody have any real evidence of mythical 7470/7500 G4s? Other than Register and MOSR rumours?)
To me, the new PowerMac case is a give away. It appears to be for higher power chips. It was introduced now for two reasons possibly. 1) It gives the impression that Apple is doing more to the PowerMac line now, when sales are hurting. 2) It makes for a quicker and easier introduction of the G5 when it is ready. There will be no new tooling for a new case. Sure, many would like to see some super nifty case for the G5, but I'll bet they will be happy just to get the G5 in today's case.
<strong>Okay, maybe I'm too optimistic, but why are many pushing the date so far out to get the IBM G5? Isn't it very possible that Apple realized long ago that Motorola would not be the most reliable supplier? Wouldn't Apple think twice about who should build the G5? Work could have started well over a year ago, maybe two. And with both Apple and nVidia as charter members of HyperTransport, wouldn't Apple be asking for such a bus on the G5? I believe samples of the G5 could have been undergoing tests for over six month already. I think it reasonable that IBM would delay talking about the G5, until it is about ready for production fab. I don't say that my opinion is right, but that it has just as good a chance of being right as those who expect a G5 at MWNY or later. I may be disappointed after October 15 or after MWSF, but I hope not. </strong><hr></blockquote>Just for reference, the Sahara was introduced at the Microprocessor Forum last October, and then got into iBooks 7 months later this past May. At the conference, they said the Sahara would begin sampling in January. So we'll have to see when they say this new Power4 begins sampling - that may give us an idea as to when Apple could use it.