I'm skeptical...but then again, Apple is a member of the Hypertransport board with AMD and Nvidia, and it's been a member since 2001 (when they realized Motorola's G5 was a bust?).
What else could Apple have possibly been doing in the Hypertransport board besides developing a chipset in consort with Nvidia and AMD? Does anyone know or have any ideas?
I'm skeptical...but then again, Apple is a member of the Hypertransport board with AMD and Nvidia, and it's been a member since 2001 (when they realized Motorola's G5 was a bust?).
What else could Apple have possibly been doing in the Hypertransport board besides developing a chipset in consort with Nvidia and AMD? Does anyone know or have any ideas?
What else could Apple have possibly been doing in the Hypertransport board besides developing a chipset in consort with Nvidia and AMD? Does anyone know or have any ideas?
I don't know about the nVidia connection (although IBM will be producing chips for them soon, too), but I doubt that Apple would have been in the HT consortium all this time and NOT use HT.
As I've said before, I think most people will be quite happy when they see the support structure for the 970.
I wish I could reveal more, but hey, it's not the greatest time to be out of a job.\
True TO. I and others rather not know for a couple of more months then see you fired or worse.
Generally speaking though. IF nforce is for a "lowend" machine. I wonder how much money it saves and if this means that apple is really trying to reach a $1499 price point for the low end??
has anyone heard what price point PCs will be that have this chip set. (i know the mac version would be different) but are we talking $2200 machines or $1200 machines here...
High end solutions don't use integrated chipsets like the nForce. That sort of integration is a cost-cutting measure.
Cost-cutting measures and high-end solutions aren't mutually exclusive. There are two versions of the nForce 2 chipset, IGP and SPP. Only IGP has the integrated graphics processor. High-end chipsets like Intel's Canterwood integrate a whole lot too into their northbridges and southbridges...
Quote:
As far as I know, ATi doesn't design motherboards.
It's called ATi Radeon IGP and its about to get a whole lot bigger. nVidia doesn't build motherboards either,
Quote:
First, Apple has the option of getting boards from PC board makers -> lower production costs due to competition and volume, and the lack of a substantial initial design cost.
Apple already has its boards built by ECS, one of the biggest PC board makers out there. It was recently purchased by .... I forget who.
Quote:
Second, Apple has the option of using Opterons and their successors.
I really doubt this.
Quote:
Third, if true this would be the first ATX motherboard Apple has released in how long? That has a few implications as well, especially if Apple decides to let companies like Core Computer continue to exist...
Apple really hasn't ever released a pure ATX based product. Neither their PSUs or motherboards can be dropped in as ATX replacements without minor mods. The boards in the early G4s were pretty close to ATX spec.
I still want to know why Apple wouldn't use the same chipset for high and lowend Powermacs. They must have a meaty chipset lined up for the highend and the Xserves. Does that mean the Xserves will be the most powerful computers in the Apple lineup?
Using the Nvidia chipsets will hopefully drive costs down significantly by riding the PC commodity market gravy train. And maybe, just maybe, Powermacs will reflect that cost drop. I'd like to see Powermacs start at $999 instead of $1499 (I know, keep dreaming, right?)
I still want to know why Apple wouldn't use the same chipset for high and lowend Powermacs. They must have a meaty chipset lined up for the highend and the Xserves. Does that mean the Xserves will be the most powerful computers in the Apple lineup?
Using the Nvidia chipsets will hopefully drive costs down significantly by riding the PC commodity market gravy train. And maybe, just maybe, Powermacs will reflect that cost drop. I'd like to see Powermacs start at $999 instead of $1499 (I know, keep dreaming, right?)
If Apple is using the 970 then they wont be "PC commodity market gravy train", this will be for the Nvidia version that uses X86 or AMD chips. The 970 varients would cost more, just as ATI video cards cost more for Macs, lower volume to offset the cost of R&D. I doubt going with an Nvidia mobo would save Apple, or consumers much. But remember Apple is a member of the PowerPC consortium, so they could develop their own Hypertransport Mobo, or for more "cost-savings" co-develop it with IBM.
The whole idea of the UMA model that Apple has moved to is a cost savings measure, spreading the R&D cost of one chipset over as many products as possable. continuing with this model is Apples best hope for keeping product prices as low as possible, while maintaining the highest possible profit margins.
Using the Nvidia chipsets will hopefully drive costs down significantly by riding the PC commodity market gravy train. And maybe, just maybe, Powermacs will reflect that cost drop. I'd like to see Powermacs start at $999 instead of $1499 (I know, keep dreaming, right?)
Yes unfortunately it is a Dream. Apple has been riding the PC Gravy Train for years. When they dropped SCSI, ADB and other more expensive tech. Trust me Apple's motherboards are probably not more expensive than the typical ATX PC motherboard. With Apple's use of the same MB in multiple lines ...it's likely they're fetching a great price on their MBs. However apple charges 20 something % margin on some of their products and PCs could only dream about charging the same.
I realy don't see Apple using nForce. nForce is nice but one can purchase a KT400 board which is just as fast. nForce is somewhat overhyped.
Cost-cutting measures and high-end solutions aren't mutually exclusive. There are two versions of the nForce 2 chipset, IGP and SPP. Only IGP has the integrated graphics processor. High-end chipsets like Intel's Canterwood integrate a whole lot too into their northbridges and southbridges...
So, of course, do Apple's.
I was looking at the IGP, which has a level of integration that spells low(er) end. I was also going by the thread title, and assuming that the implication was that the IGP was what Apple would be interested in.
Quote:
It's called ATi Radeon IGP and its about to get a whole lot bigger. nVidia doesn't build motherboards either,
No, they design motherboards.
I didn't know about the ATi board. Thanks.
Quote:
Apple already has its boards built by ECS, one of the biggest PC board makers out there. It was recently purchased by .... I forget who.
Right, but the quantity built is whatever Apple asks for. If Apple uses the nVIDIA board, they get at least some of the economies of scale ofa product aimed at the much larger PC market.
Quote:
I really doubt [that Apple would use an Opteron-based board].
So do I. For that matter, I doubt Apple will use nVIDIA's board. But putting the speculation hat on and assuming they would, it's interesting to spin out the implications. That's one of the more radical and unlikely ones, but still.
Quote:
Apple really hasn't ever released a pure ATX based product. Neither their PSUs or motherboards can be dropped in as ATX replacements without minor mods. The boards in the early G4s were pretty close to ATX spec.
That's what I thought. Now, assuming they use this board, that raises some novel possibilities, doesn't it?
The trick would be making damn sure that the clone debacle didn't happen again, and that's a real trick. Personally, I consider it a deal-breaker, but I think it's interesting to poke around every once in a while and see if there's a way to pull it off.
Yes unfortunately it is a Dream. Apple has been riding the PC Gravy Train for years. When they dropped SCSI, ADB and other more expensive tech. Trust me Apple's motherboards are probably not more expensive than the typical ATX PC motherboard. With Apple's use of the same MB in multiple lines ...it's likely they're fetching a great price on their MBs. However apple charges 20 something % margin on some of their products and PCs could only dream about charging the same.
Dell's margin is a pretty consistent 20%.
Quote:
I realy don't see Apple using nForce. nForce is nice but one can purchase a KT400 board which is just as fast. nForce is somewhat overhyped.
nForce has one advantage, from Apple's point of view: There are lots of wierd little proprietary bits. If Apple tailored a hypothetical OS X to run on the nForce, it would not easily move to a different board.
However, I'm still not convinced that Apple couldn't do better with a board built around and for the PPC 970.
I still want to know why Apple wouldn't use the same chipset for high and lowend Powermacs. They must have a meaty chipset lined up for the highend and the Xserves. Does that mean the Xserves will be the most powerful computers in the Apple lineup?
Using the Nvidia chipsets will hopefully drive costs down significantly by riding the PC commodity market gravy train. And maybe, just maybe, Powermacs will reflect that cost drop. I'd like to see Powermacs start at $999 instead of $1499 (I know, keep dreaming, right?)
Quote:
Originally posted by keyboard12
Generally speaking though. IF nforce is for a "lowend" machine. I wonder how much money it saves and if this means that apple is really trying to reach a $1499 price point for the low end??
Wait a minute.
What does the title of this tread really say? If we ignore the credibility argument for a moment and blindly assume that the claim made in the title is true, then the Nvidia Nforce 3 will be the basis of lower end 970 Macs.
I've been reading this to mean lower end Power Macs. It doesn't say that. So, lower-end Macs, maybe even in the $1K ~ $1.4K range, would be using 970s and Nforce 3 chips.
No it isn't. The EV6 interface is on-chip in the Opteron. The nForce3 talks to the Opteron via HyperTransport. The Opteron has an onboard memory controller, the memory goes straight into the Opteron, not via the North Bridge.
Quote:
Unless Apple is about to start using the Athlon or Alpha, the nforce will be of absolutly no use to them.
I can see a significant advantage to using the nForce chipset. They already supply every interface Apple cares about on 1 piece of silicon. Apple create a HT <-> 970 FSB controller chip and get a proven design for a fraction of the cost that they could create, test, manufacture and integrate the same thing for.
This is the beauty of HT. Why should Apple have to design custon chipsets? What value do they actually provide? There are a handful of companies on the PC side that design great chipsets, the glue is HT to bind them together with smaller logic packages by Apple to add unique things, like the memory bridge. Let the other companies do the grunt work and Apple add value.
Of course the next step for the 970+ is to put that memory bridge onto the processor and just have an HT (or 2) connection(s) onto the motherboard.
I can see a significant advantage to using the nForce chipset. They already supply every interface Apple cares about on 1 piece of silicon. Apple create a HT <-> 970 FSB controller chip and get a proven design for a fraction of the cost that they could create, test, manufacture and integrate the same thing for.
This is the beauty of HT. [/B]
This is where I'm sitting. _If_ Apple is going anywhere near something like this, it's because the 970 will work on/with it. It's conceivable that Apple might do something _additional_ (offering AMD servers or something), but it would have to make sense to the 'mainline' macs. Someone else to split development costs with would rock. HT allows a lot of interesting things to be added to the mb with a heck of a lot less 'design'. You meet the spec - things interoperate.
Apple's current designs have the CPU on a daughtercard already, this would move the RAM on the card too...
What we know about the 970's FSB is remarkably similar to HyperTransport. I haven't compared them in detail but somebody should verify that it isn't a superset or variation of HyperTransport (i.e. HT w/ PPC specific bus snooping protocols added). Rumour is, after all, that it is the ApplePI and Apple is on the HyperTransport consortium. Even if it isn't, it is similar enough that getting nVidia to build a chipset to support it probably wouldn't be completely impossible. I'm not saying this is likely, but its not out of the question.
...Rumour is, after all, that it is the ApplePI and Apple is on the HyperTransport consortium.
Part of that is fact, not rumor.
Apple _is_ part of the Hypertransport Consortium. Wait for the logos on the right to cycle, look for the Apple. -> Fact.
Their old website listed Apple as one of the 'top tier' members too. (Five or six companies on a steering committee). I don't see this listed currently, so I can't prove it -> rumor.
But we still have no idea what "ApplePI" is, whether it exists, whether it is something fancy... or if it is just a renaming of something else. -> Everything involving ApplePI = rumor.
But we still have no idea what "ApplePI" is, whether it exists, whether it is something fancy... or if it is just a renaming of something else. -> Everything involving ApplePI = rumor.
Just because it is a rumor doesn't mean that it isn't 100% true - just that you can't point to widely available "facts" and convince yourself that it is true.
If everything I've heard about ApplePI is true, then I'd be giddy.
If _nothing_ I've heard about ApplePI is true then I'd be aghast.
But I don't see PR from AMD saying "Now Hypertransport will be called ApplePI on Apple machines." I don't see PR from IBM claiming that the bus on the 970 will be called ApplePI by Apple. I see nothing on the RapidIO sites/documentation... -> RUMOR.
The comments we have second- and third-hand are enticing. The main thrust of the rumors make sense "Apple has something similar/identical to HT or RIO that will be the next Apple bus." Ok, fine. IBM & MOT are both making/planning on making RIO chips, and are on the RIO steering comittee. Apple, nVidia, and AMD are on the HT members list. Those two things are facts - we can look them up.
But what's in _just_ a name? There's people running around arguing semantics to the point of "It depends on what your definition of 'is' is." Gimme a break. So the next Apple bus might be called ApplePI -> so what? I can call the MAXbus ApplePI if I like. But... if that means everything it _could_ mean, then that would rock.
That is, Apple going to a NUMA archetecture, developing greater than 4-way computers, having the "CPU daughtercard" socket their fond of being an _accepted standard_, potential for multiple GPUs per machine, potential to share motherboard costs with AMD... Things like this have been proposed as parts of "The ApplePI rumor". But anyone claiming Apple has a 4-way NUMA box with 2-pro-class graphics engines is spreading a rumor.
If you have inside info, great. If you don't think the ApplePI "rumor" is a rumor... then I need to start saving more assiduously, and I'll need to buy earplugs for Matsu's screams at the price of the decked-out version.
And I also have no doubt that there's a LOT of hardware inside Apple that's cool. And inexplicable. 32-way Sun servers in the Rhapsody era? Dell shipments? Areas labeled STTNG in the advanced research areas? But even functional prototypes seeded to developers have been scrapped before, so to move something from 'rumor' to 'planned' takes more that a physical box that works.
Comments
What else could Apple have possibly been doing in the Hypertransport board besides developing a chipset in consort with Nvidia and AMD? Does anyone know or have any ideas?
Originally posted by Gizzmonic
I'm skeptical...but then again, Apple is a member of the Hypertransport board with AMD and Nvidia, and it's been a member since 2001 (when they realized Motorola's G5 was a bust?).
What else could Apple have possibly been doing in the Hypertransport board besides developing a chipset in consort with Nvidia and AMD? Does anyone know or have any ideas?
In conjunction with IBM maybe...?
Originally posted by Gizzmonic
What else could Apple have possibly been doing in the Hypertransport board besides developing a chipset in consort with Nvidia and AMD? Does anyone know or have any ideas?
I don't know about the nVidia connection (although IBM will be producing chips for them soon, too), but I doubt that Apple would have been in the HT consortium all this time and NOT use HT.
As I've said before, I think most people will be quite happy when they see the support structure for the 970.
I wish I could reveal more, but hey, it's not the greatest time to be out of a job.
True TO. I and others rather not know for a couple of more months then see you fired or worse.
Generally speaking though. IF nforce is for a "lowend" machine. I wonder how much money it saves and if this means that apple is really trying to reach a $1499 price point for the low end??
has anyone heard what price point PCs will be that have this chip set. (i know the mac version would be different) but are we talking $2200 machines or $1200 machines here...
Originally posted by Amorph
High end solutions don't use integrated chipsets like the nForce. That sort of integration is a cost-cutting measure.
Cost-cutting measures and high-end solutions aren't mutually exclusive. There are two versions of the nForce 2 chipset, IGP and SPP. Only IGP has the integrated graphics processor. High-end chipsets like Intel's Canterwood integrate a whole lot too into their northbridges and southbridges...
As far as I know, ATi doesn't design motherboards.
It's called ATi Radeon IGP and its about to get a whole lot bigger. nVidia doesn't build motherboards either,
First, Apple has the option of getting boards from PC board makers -> lower production costs due to competition and volume, and the lack of a substantial initial design cost.
Apple already has its boards built by ECS, one of the biggest PC board makers out there. It was recently purchased by .... I forget who.
Second, Apple has the option of using Opterons and their successors.
I really doubt this.
Third, if true this would be the first ATX motherboard Apple has released in how long? That has a few implications as well, especially if Apple decides to let companies like Core Computer continue to exist...
Apple really hasn't ever released a pure ATX based product. Neither their PSUs or motherboards can be dropped in as ATX replacements without minor mods. The boards in the early G4s were pretty close to ATX spec.
Using the Nvidia chipsets will hopefully drive costs down significantly by riding the PC commodity market gravy train. And maybe, just maybe, Powermacs will reflect that cost drop. I'd like to see Powermacs start at $999 instead of $1499 (I know, keep dreaming, right?)
Originally posted by Gizzmonic
I still want to know why Apple wouldn't use the same chipset for high and lowend Powermacs. They must have a meaty chipset lined up for the highend and the Xserves. Does that mean the Xserves will be the most powerful computers in the Apple lineup?
Using the Nvidia chipsets will hopefully drive costs down significantly by riding the PC commodity market gravy train. And maybe, just maybe, Powermacs will reflect that cost drop. I'd like to see Powermacs start at $999 instead of $1499 (I know, keep dreaming, right?)
If Apple is using the 970 then they wont be "PC commodity market gravy train", this will be for the Nvidia version that uses X86 or AMD chips. The 970 varients would cost more, just as ATI video cards cost more for Macs, lower volume to offset the cost of R&D. I doubt going with an Nvidia mobo would save Apple, or consumers much. But remember Apple is a member of the PowerPC consortium, so they could develop their own Hypertransport Mobo, or for more "cost-savings" co-develop it with IBM.
The whole idea of the UMA model that Apple has moved to is a cost savings measure, spreading the R&D cost of one chipset over as many products as possable. continuing with this model is Apples best hope for keeping product prices as low as possible, while maintaining the highest possible profit margins.
Using the Nvidia chipsets will hopefully drive costs down significantly by riding the PC commodity market gravy train. And maybe, just maybe, Powermacs will reflect that cost drop. I'd like to see Powermacs start at $999 instead of $1499 (I know, keep dreaming, right?)
Yes unfortunately it is a Dream. Apple has been riding the PC Gravy Train for years. When they dropped SCSI, ADB and other more expensive tech. Trust me Apple's motherboards are probably not more expensive than the typical ATX PC motherboard. With Apple's use of the same MB in multiple lines ...it's likely they're fetching a great price on their MBs. However apple charges 20 something % margin on some of their products and PCs could only dream about charging the same.
I realy don't see Apple using nForce. nForce is nice but one can purchase a KT400 board which is just as fast. nForce is somewhat overhyped.
Originally posted by Eugene
Cost-cutting measures and high-end solutions aren't mutually exclusive. There are two versions of the nForce 2 chipset, IGP and SPP. Only IGP has the integrated graphics processor. High-end chipsets like Intel's Canterwood integrate a whole lot too into their northbridges and southbridges...
So, of course, do Apple's.
I was looking at the IGP, which has a level of integration that spells low(er) end. I was also going by the thread title, and assuming that the implication was that the IGP was what Apple would be interested in.
It's called ATi Radeon IGP and its about to get a whole lot bigger. nVidia doesn't build motherboards either,
No, they design motherboards.
I didn't know about the ATi board. Thanks.
Apple already has its boards built by ECS, one of the biggest PC board makers out there. It was recently purchased by .... I forget who.
Right, but the quantity built is whatever Apple asks for. If Apple uses the nVIDIA board, they get at least some of the economies of scale ofa product aimed at the much larger PC market.
I really doubt [that Apple would use an Opteron-based board].
So do I. For that matter, I doubt Apple will use nVIDIA's board. But putting the speculation hat on and assuming they would, it's interesting to spin out the implications. That's one of the more radical and unlikely ones, but still.
Apple really hasn't ever released a pure ATX based product. Neither their PSUs or motherboards can be dropped in as ATX replacements without minor mods. The boards in the early G4s were pretty close to ATX spec.
That's what I thought. Now, assuming they use this board, that raises some novel possibilities, doesn't it?
The trick would be making damn sure that the clone debacle didn't happen again, and that's a real trick. Personally, I consider it a deal-breaker, but I think it's interesting to poke around every once in a while and see if there's a way to pull it off.
Originally posted by hmurchison
Yes unfortunately it is a Dream. Apple has been riding the PC Gravy Train for years. When they dropped SCSI, ADB and other more expensive tech. Trust me Apple's motherboards are probably not more expensive than the typical ATX PC motherboard. With Apple's use of the same MB in multiple lines ...it's likely they're fetching a great price on their MBs. However apple charges 20 something % margin on some of their products and PCs could only dream about charging the same.
Dell's margin is a pretty consistent 20%.
I realy don't see Apple using nForce. nForce is nice but one can purchase a KT400 board which is just as fast. nForce is somewhat overhyped.
nForce has one advantage, from Apple's point of view: There are lots of wierd little proprietary bits. If Apple tailored a hypothetical OS X to run on the nForce, it would not easily move to a different board.
However, I'm still not convinced that Apple couldn't do better with a board built around and for the PPC 970.
Originally posted by Gizzmonic
I still want to know why Apple wouldn't use the same chipset for high and lowend Powermacs. They must have a meaty chipset lined up for the highend and the Xserves. Does that mean the Xserves will be the most powerful computers in the Apple lineup?
Using the Nvidia chipsets will hopefully drive costs down significantly by riding the PC commodity market gravy train. And maybe, just maybe, Powermacs will reflect that cost drop. I'd like to see Powermacs start at $999 instead of $1499 (I know, keep dreaming, right?)
Originally posted by keyboard12
Generally speaking though. IF nforce is for a "lowend" machine. I wonder how much money it saves and if this means that apple is really trying to reach a $1499 price point for the low end??
Wait a minute.
What does the title of this tread really say? If we ignore the credibility argument for a moment and blindly assume that the claim made in the title is true, then the Nvidia Nforce 3 will be the basis of lower end 970 Macs.
I've been reading this to mean lower end Power Macs. It doesn't say that. So, lower-end Macs, maybe even in the $1K ~ $1.4K range, would be using 970s and Nforce 3 chips.
If such a claim is at all true.
??
care to chime in producer?
at least clarify what you meant
Originally posted by Stagflation Steve
This is laughable, nforce is an EV-6 chipset.
No it isn't. The EV6 interface is on-chip in the Opteron. The nForce3 talks to the Opteron via HyperTransport. The Opteron has an onboard memory controller, the memory goes straight into the Opteron, not via the North Bridge.
Unless Apple is about to start using the Athlon or Alpha, the nforce will be of absolutly no use to them.
I can see a significant advantage to using the nForce chipset. They already supply every interface Apple cares about on 1 piece of silicon. Apple create a HT <-> 970 FSB controller chip and get a proven design for a fraction of the cost that they could create, test, manufacture and integrate the same thing for.
This is the beauty of HT. Why should Apple have to design custon chipsets? What value do they actually provide? There are a handful of companies on the PC side that design great chipsets, the glue is HT to bind them together with smaller logic packages by Apple to add unique things, like the memory bridge. Let the other companies do the grunt work and Apple add value.
Of course the next step for the 970+ is to put that memory bridge onto the processor and just have an HT (or 2) connection(s) onto the motherboard.
Originally posted by JRG
I can see a significant advantage to using the nForce chipset. They already supply every interface Apple cares about on 1 piece of silicon. Apple create a HT <-> 970 FSB controller chip and get a proven design for a fraction of the cost that they could create, test, manufacture and integrate the same thing for.
This is the beauty of HT. [/B]
This is where I'm sitting. _If_ Apple is going anywhere near something like this, it's because the 970 will work on/with it. It's conceivable that Apple might do something _additional_ (offering AMD servers or something), but it would have to make sense to the 'mainline' macs. Someone else to split development costs with would rock. HT allows a lot of interesting things to be added to the mb with a heck of a lot less 'design'. You meet the spec - things interoperate.
Apple's current designs have the CPU on a daughtercard already, this would move the RAM on the card too...
Originally posted by Programmer
...Rumour is, after all, that it is the ApplePI and Apple is on the HyperTransport consortium.
Part of that is fact, not rumor.
Apple _is_ part of the Hypertransport Consortium. Wait for the logos on the right to cycle, look for the Apple. -> Fact.
Their old website listed Apple as one of the 'top tier' members too. (Five or six companies on a steering committee). I don't see this listed currently, so I can't prove it -> rumor.
But we still have no idea what "ApplePI" is, whether it exists, whether it is something fancy... or if it is just a renaming of something else. -> Everything involving ApplePI = rumor.
Originally posted by Nevyn
But we still have no idea what "ApplePI" is, whether it exists, whether it is something fancy... or if it is just a renaming of something else. -> Everything involving ApplePI = rumor.
It is, huh?
Originally posted by Jonathan
It is, huh?
Well, yes.
Just because it is a rumor doesn't mean that it isn't 100% true - just that you can't point to widely available "facts" and convince yourself that it is true.
If everything I've heard about ApplePI is true, then I'd be giddy.
If _nothing_ I've heard about ApplePI is true then I'd be aghast.
But I don't see PR from AMD saying "Now Hypertransport will be called ApplePI on Apple machines." I don't see PR from IBM claiming that the bus on the 970 will be called ApplePI by Apple. I see nothing on the RapidIO sites/documentation... -> RUMOR.
The comments we have second- and third-hand are enticing. The main thrust of the rumors make sense "Apple has something similar/identical to HT or RIO that will be the next Apple bus." Ok, fine. IBM & MOT are both making/planning on making RIO chips, and are on the RIO steering comittee. Apple, nVidia, and AMD are on the HT members list. Those two things are facts - we can look them up.
But what's in _just_ a name? There's people running around arguing semantics to the point of "It depends on what your definition of 'is' is." Gimme a break. So the next Apple bus might be called ApplePI -> so what? I can call the MAXbus ApplePI if I like. But... if that means everything it _could_ mean, then that would rock.
That is, Apple going to a NUMA archetecture, developing greater than 4-way computers, having the "CPU daughtercard" socket their fond of being an _accepted standard_, potential for multiple GPUs per machine, potential to share motherboard costs with AMD... Things like this have been proposed as parts of "The ApplePI rumor". But anyone claiming Apple has a 4-way NUMA box with 2-pro-class graphics engines is spreading a rumor.
If you have inside info, great. If you don't think the ApplePI "rumor" is a rumor... then I need to start saving more assiduously, and I'll need to buy earplugs for Matsu's screams at the price of the decked-out version.
And I also have no doubt that there's a LOT of hardware inside Apple that's cool. And inexplicable. 32-way Sun servers in the Rhapsody era? Dell shipments? Areas labeled STTNG in the advanced research areas? But even functional prototypes seeded to developers have been scrapped before, so to move something from 'rumor' to 'planned' takes more that a physical box that works.