If Apple is going to spring a x86-based Mac on us developers, they will have to wait a VERY LONG TIME before 3rd party software and hardware is ported. . .
</strong><hr></blockquote>
Read the whole post by 'strobe' again if you forgot the gist of it. Most who want a x86 Mac think a year is too long to wait for it. Well, an x86 Mac could take longer when you consider how long it would take to get most of the applications we need. Not all developers are going to salute, and say, "Yes sir, Mr. Jobs. We will get all our products ported as soon as possible." As soon as possible might be three to six months, and most will take longer. Some applications may not get ported at all. The threat to say, "It ain't worth it," could be very real.
Then too, developers would need to provide both a PPC and x86 version for many years. Mac users who are not happy about buying all new applications may put off purchasing new hardware for quite a while. Gee, just what Apple needs. Someone used the word "suicide," and I think that comes close to the effect such a move would have on Apple right now.
Read the whole post by 'strobe' again if you forgot the gist of it. Most who want a x86 Mac think a year is too long to wait for it. Well, an x86 Mac could take longer when you consider how long it would take to get most of the applications we need. Not all developers are going to salute, and say, "Yes sir, Mr. Jobs. We will get all our products ported as soon as possible." As soon as possible might be three to six months, and most will take longer. Some applications may not get ported at all. The threat to say, "It ain't worth it," could be very real.
Then too, developers would need to provide both a PPC and x86 version for many years. Mac users who are not happy about buying all new applications may put off purchasing new hardware for quite a while. Gee, just what Apple needs. Someone used the word "suicide," and I think that comes close to the effect such a move would have on Apple right now.
I find it highly unlikely ( for reasons already stated ) that Apple would swich to a x86 or a PC AMD chip. But, could it be Apple and AMD have developed something together?
<strong>What's my point? IBM may be more like Apple in the sense that it wants to use these chips in high end workstations (good). They claim scalability. But how can its future be guaranteed? In three years what if IBM's linux workstations are a flop and they decide to suspend the 970s/future PPC chips. The PPC's future is not as certain as chips made by AMD and Intel.</strong><hr></blockquote>
A couple of things:
1) AMD is not in rock-solid shape. They've been living from chip-intro to chip-intro... and barely making it. Getting a guaranteed customer (like Apple) might help, but that's only if Apple could commit to 3+ million AMD's a year. (Dropping ppc entirely, and NOW) I'm not sure I'd want Apple to be reliant upon buying chips from Intel - ever. I'd expect Apple to pay more per 10000 chips from Intel than any other computer manufacturer - just from years of bad blood. Just like Dell gets a sweetheart deal. Piss the supplier off - pay the price. Witness Motorola's lackluster performance ever since clones were killed off (which irritated Mot). Nothing _so_ bad Apple could expect to successfully sue Mot for breach of contract perhaps... but still pretty dismal.
2) There are ppc 601's and ppc604's (not to mention the Power3's and Power4's which also support the ppc instruction set) in a LOT of other portions of IBM's business. It isn't just 'Some linux on the workstation initiative', they're interested in replacing aging chip lines in AS400s, RS6000s, their SP lines and the big custom supercomputers. IBM also competes in the embedded space.... Precisely which chips they'll use where isn't something I can guess but there's plenty of places for IBM to reuse Power-4 or Power-5 research -> just doesn't seem like there's any way IBM will drop this line until they have something provably better. (Which would be better than fine with us too).
Edit added:
On the other hand, an Apple XRAID powered by AMD, or perhaps an announcement about Apple & HyperTransport (which is/was driven by AMD) wouldn't completely 'switch' but would line up with available rumors.
<strong>What about apple just moving the XServe to AMD? There are no peripherals to worry about (hardly)...</strong><hr></blockquote>
Could open the door to "hacks" of the Xserve OSX running on equally-specced PCs. cheap and non-Apple ones. a bunch of G3s, or Crusoes would be a more likely "switch" for cheaper servers.
AMD CPUs, no, but what about AMD designing/producing some of the other chips on the mb. People are wondering where Apple will get its new front side chips from. AMD has experience in this area and is Apple's Hypertransport partner and so this might make more sense.
Would AMD make support chips for Apple's next generation of machines?
Whilst not a reply to any single post on what the heck's going on in Las Vegas, I thought I'd just contribute my $0.02.
It makes no sense for Apple to make a switch to IA-64 or any other Intel-derived hardware base: Whilst 970 is a year or more away, it does at least have binary compatibility with the rest of the PPC family from the 601 (G1) onwards. A move to a different architecture will cause wholesale confusion not just in the developer community but also in the marketplace, where any Apple purchases will go on hold for 18 months whilst the consumer (personal or corporate) tries to establish what an appropriate direction.
Also, do people really believe that an announcement of that kind would be made at an event controlled by AMD? If so, where can I buy the chemicals that are so obviously being abused by the supporters of that view.
In reality, the most outlandish theory that could be promoted is that AMD has chosen to hedge its bets and has chosen/been accepted to join the AIM consortium.
Why?
Maybe AMD has been convinced by Apple that Moto is no longer considered a viable supply partner for 970 and beyond.
There are some who don't even believe that Moto is capable as acting as a supply partner for G4, but I think that's a little extreme given the fact that they're still trying to squeeze G4+ up to 2GHz which maybe all we've got to look forward to for the next year. That said, any G5 from Moto is now looking like embedded-use only and it's fair to say that Moto seems to have no long-term strategic interest in the desktop/server CPU market, which means that the consortium starts to look a bit lightweight.
Maybe Apple and IBM have both agreed that Apple needs a fallback/joint-supplier for 970 in case East Fishkill needs to be refocussed for 'Cell' and other purposes.
If 'Cell' takes off, Apple will need a security blanket of a fab plant that can be focussed on Apple's requirements. Duron/Opteron/etc. will only ever be a niche market in the WinXXX marketplace, and AMD can never dedicate the same kind of dollars to creating an Intel-like level of brand recognition or loyalty.
AMD develop another revenue stream, Apple get a security blanket, and - if the deal is structured correctly -
the AIM (or should that be AIMA) consortium get HyperTransport and any other desirable AMD technologies at a strategic design level.
This is probably quite important to AMD! It has no long-term stability in the WinXXX marketplace and needs to find a safety net in case the marketshare gulf with Intel widens unexpectedly.
Sure, everyone has been focussing on the fact that there are more Apple faces at Comdex than people were expecting, but would we even recognise the relevant IBM or Moto VPs who would want to be present to welcome AMD to the family - they may be there but we can't see them in their dull blue/brown suits.
Comments
<strong>
If Apple is going to spring a x86-based Mac on us developers, they will have to wait a VERY LONG TIME before 3rd party software and hardware is ported. . .
</strong><hr></blockquote>
Read the whole post by 'strobe' again if you forgot the gist of it. Most who want a x86 Mac think a year is too long to wait for it. Well, an x86 Mac could take longer when you consider how long it would take to get most of the applications we need. Not all developers are going to salute, and say, "Yes sir, Mr. Jobs. We will get all our products ported as soon as possible." As soon as possible might be three to six months, and most will take longer. Some applications may not get ported at all. The threat to say, "It ain't worth it," could be very real.
Then too, developers would need to provide both a PPC and x86 version for many years. Mac users who are not happy about buying all new applications may put off purchasing new hardware for quite a while. Gee, just what Apple needs. Someone used the word "suicide," and I think that comes close to the effect such a move would have on Apple right now.
[ 11-19-2002: Message edited by: snoopy ]</p>
<strong>
Read the whole post by 'strobe' again if you forgot the gist of it. Most who want a x86 Mac think a year is too long to wait for it. Well, an x86 Mac could take longer when you consider how long it would take to get most of the applications we need. Not all developers are going to salute, and say, "Yes sir, Mr. Jobs. We will get all our products ported as soon as possible." As soon as possible might be three to six months, and most will take longer. Some applications may not get ported at all. The threat to say, "It ain't worth it," could be very real.
Then too, developers would need to provide both a PPC and x86 version for many years. Mac users who are not happy about buying all new applications may put off purchasing new hardware for quite a while. Gee, just what Apple needs. Someone used the word "suicide," and I think that comes close to the effect such a move would have on Apple right now.
[ 11-19-2002: Message edited by: snoopy ]</strong><hr></blockquote>
Finally, somebody else that gets it. OS X + x86 = No more Macs for anybody. Thanks, snoop.
They apparently fell through. Why should AMD produce a desktop PowerPC? They have enough problems keeping up with Intel.
By the way, PowerPC is as much a standard as x86. What good would x86 bring to the Mac platform that a future PowerPC won't?
Barto
<strong>What's my point? IBM may be more like Apple in the sense that it wants to use these chips in high end workstations (good). They claim scalability. But how can its future be guaranteed? In three years what if IBM's linux workstations are a flop and they decide to suspend the 970s/future PPC chips. The PPC's future is not as certain as chips made by AMD and Intel.</strong><hr></blockquote>
A couple of things:
1) AMD is not in rock-solid shape. They've been living from chip-intro to chip-intro... and barely making it. Getting a guaranteed customer (like Apple) might help, but that's only if Apple could commit to 3+ million AMD's a year. (Dropping ppc entirely, and NOW) I'm not sure I'd want Apple to be reliant upon buying chips from Intel - ever. I'd expect Apple to pay more per 10000 chips from Intel than any other computer manufacturer - just from years of bad blood. Just like Dell gets a sweetheart deal. Piss the supplier off - pay the price. Witness Motorola's lackluster performance ever since clones were killed off (which irritated Mot). Nothing _so_ bad Apple could expect to successfully sue Mot for breach of contract perhaps... but still pretty dismal.
2) There are ppc 601's and ppc604's (not to mention the Power3's and Power4's which also support the ppc instruction set) in a LOT of other portions of IBM's business. It isn't just 'Some linux on the workstation initiative', they're interested in replacing aging chip lines in AS400s, RS6000s, their SP lines and the big custom supercomputers. IBM also competes in the embedded space.... Precisely which chips they'll use where isn't something I can guess but there's plenty of places for IBM to reuse Power-4 or Power-5 research -> just doesn't seem like there's any way IBM will drop this line until they have something provably better. (Which would be better than fine with us too).
Edit added:
On the other hand, an Apple XRAID powered by AMD, or perhaps an announcement about Apple & HyperTransport (which is/was driven by AMD) wouldn't completely 'switch' but would line up with available rumors.
[ 11-19-2002: Message edited by: Nevyn ]</p>
<strong>What about apple just moving the XServe to AMD? There are no peripherals to worry about (hardly)...</strong><hr></blockquote>
Could open the door to "hacks" of the Xserve OSX running on equally-specced PCs. cheap and non-Apple ones. a bunch of G3s, or Crusoes would be a more likely "switch" for cheaper servers.
<strong>
Barto, great job, glad to be of inspiration. I have a new desktop pic.
Would AMD make support chips for Apple's next generation of machines?
It makes no sense for Apple to make a switch to IA-64 or any other Intel-derived hardware base: Whilst 970 is a year or more away, it does at least have binary compatibility with the rest of the PPC family from the 601 (G1) onwards. A move to a different architecture will cause wholesale confusion not just in the developer community but also in the marketplace, where any Apple purchases will go on hold for 18 months whilst the consumer (personal or corporate) tries to establish what an appropriate direction.
Also, do people really believe that an announcement of that kind would be made at an event controlled by AMD? If so, where can I buy the chemicals that are so obviously being abused by the supporters of that view.
In reality, the most outlandish theory that could be promoted is that AMD has chosen to hedge its bets and has chosen/been accepted to join the AIM consortium.
Why?
Maybe AMD has been convinced by Apple that Moto is no longer considered a viable supply partner for 970 and beyond.
There are some who don't even believe that Moto is capable as acting as a supply partner for G4, but I think that's a little extreme given the fact that they're still trying to squeeze G4+ up to 2GHz which maybe all we've got to look forward to for the next year. That said, any G5 from Moto is now looking like embedded-use only and it's fair to say that Moto seems to have no long-term strategic interest in the desktop/server CPU market, which means that the consortium starts to look a bit lightweight.
Maybe Apple and IBM have both agreed that Apple needs a fallback/joint-supplier for 970 in case East Fishkill needs to be refocussed for 'Cell' and other purposes.
If 'Cell' takes off, Apple will need a security blanket of a fab plant that can be focussed on Apple's requirements. Duron/Opteron/etc. will only ever be a niche market in the WinXXX marketplace, and AMD can never dedicate the same kind of dollars to creating an Intel-like level of brand recognition or loyalty.
AMD develop another revenue stream, Apple get a security blanket, and - if the deal is structured correctly -
the AIM (or should that be AIMA) consortium get HyperTransport and any other desirable AMD technologies at a strategic design level.
This is probably quite important to AMD! It has no long-term stability in the WinXXX marketplace and needs to find a safety net in case the marketshare gulf with Intel widens unexpectedly.
Sure, everyone has been focussing on the fact that there are more Apple faces at Comdex than people were expecting, but would we even recognise the relevant IBM or Moto VPs who would want to be present to welcome AMD to the family - they may be there but we can't see them in their dull blue/brown suits.