Actually when the G4 first came out it was the fastest chip around-by far-by a factor of two.</strong><hr></blockquote>
Okay, okay, I was being too vague. The speed increase to 500 MHz, that had to be retracted and changed to 400 MHz, is the early embarrassing event I was referring to. I didn't want to imply first, but simply some time ago. Not recent.
Might I suggest that it would be possible to pair (or more) an AMD processor with a VMX-compliant DSP processor, using one of the fast new bus protocols specifically developed for inter-chip communications? That would allow AltiVec code rewriting to be a fairly simple task.
Furthermore, it would set an Apple developed AMD-based computer (not a Mac, obviously!) apart from the IBM-compatible, Windows-compatible PC standard. I seriously believe that even by using standard components for most of architecture, Apple could produce a non-compatible x86 based model. A few central components (ASIC, and not just proprietary bootstrap ROMs), the very newest standards, and a very different architectural topology would do the trick.
The "Yellow Box" strategy was not viable - making Apple OS technology a mere subset in a Windows framework is an open invitation for MS to make borderline incompatibilities - but x86 hardware production is.
Of course, Classic would have to be abandoned and Carbon demoted in favor of Cocoa (which is better due to its object-code nature, rather than any cross-capability (though that, of course, is a secondary benefit of the first) - in other words, we are talking a last-ditch, white-of-the-eyes strategy.
And imagine the amount of crow it involves eating... basically a PR company's nightmare.
I can foresee only one possible situation under which we will see x86 macs and that is when x86 processors are so much faster than ppcs that they can emulate them perfectly and at the full speed of the fastest ppc available at that time.
Since that will never happen I think this conversation is pointless. PPCs have always been a more elegant, better designed architecture than Intel chips, and even the industry mostly agrees with this, with PPC derivatives being used in all kinds of embedded and otherwise applications (gamecube anyone), and I don't see ARM complaining that they can't get good performance out of RISC technology.
The only reason Intel chips are faster is because they have more $ behind them. There's nothing wrong with PPC and thye are still holding their own even if Moto aren't doing as well as they could be right now.
There is no way in hell that Apple is going to adopt x86 for a meager speed increase at the cost of 99.9% of existing software ceasing to function on the Mac AT ALL. There was enough fuss when Adobe/Microsoft took 12 months to carbonise their stuff and that's when we could still all use the original software under classic.
Apple would not survive a switch to Intel, the developers would throw their hands up in disgust and walk out on them.
<strong>Might I suggest that it would be possible to pair (or more) an AMD processor with a VMX-compliant DSP processor, using one of the fast new bus protocols specifically developed for inter-chip communications? That would allow AltiVec code rewriting to be a fairly simple task.
Furthermore, it would set an Apple developed AMD-based computer (not a Mac, obviously!) apart from the IBM-compatible, Windows-compatible PC standard.</strong><hr></blockquote>
If you're going to go reengineering AMD designs, why not strip off the x86 translation layer while you're at it and let that Alpha-inspired RISC core shine through?
You might even be able to put a (thinner) PPC translation layer over it and free up some room on the current die for those VMX units.
If you're going to go reengineering AMD designs, why not strip off the x86 translation layer while you're at it and let that Alpha-inspired RISC core shine through?
You might even be able to put a (thinner) PPC translation layer over it and free up some room on the current die for those VMX units. </strong><hr></blockquote>
I had no intention of reengineering AMD's design - I am aware that some people find it feasible to do what you are suggesting but I doubt that it would happen in the "real" world (read: be financially feasible). Apple is not a large enough customer to give AMD enough return for such an outlay.
My suggestion was to take "Hammer", or whatever other future design might be suitable AND seem to be easy to develop easily, and build a non-PC compatible architecture around it, using a blend of standard components, the best upcoming technologies, and some central customs developed ASICS. Hammer would be paired with a VMX-compatible SEPARATE DSP processor, the two coupled with HyperTransport (or another suitable inter- (not intra-) chip bus standard.
I think the main problem is perception. Apple is forced into a desperate move to adopt x86. Many people will perceive this a desperate move of a dying company. Your Apps won't work until they are recompiled and in all cases partially rewritten. Device drivers will need to be written from scratch. People will desert the Mac platform altogether. Remember when SGI tried to go to x86 partially. Flop. Sun stopped Solaris for x86. Apple would be foolish to go ahead with something so desperate.
I reiterate: if anything they need to invest in an inhouse PPC project and have it fabbed by any American or Taiwanese manufacturer that will do it for the least amount with the best technology process. x86 is not the answer. It's a question.
Just because everyone has heard of X86(intel) doesn't mean its automatically good. It's like saying that McD's is the best hamburger and that nothing could be better than VHS just because everyone uses it.
I also saw that article about Apple having an x86 compatible version of OS X but if they ported it to x86 then they could also port it to another platform.
Someone mentioned Apple would be better off doing their own chip and I certainly agree with that.
It would be impossible for Apple to develop their own chips. They would need a heck of a lot more money. Just think, they'd need plants to manufacture it and people to design it. The worst part would be the R&D. And this translates to higher prices for their computers to recoup the costs of manufacturing and R&D. This will further alienate users from buying Apple products and cause them to lose more money. If they ever do decide to make their own chips (God help them if they do) then sell all your stock because Apple is done.
It would be impossible for Apple to develop their own chips. They would need a heck of a lot more money. Just think, they'd need plants to manufacture it and people to design it. The worst part would be the R&D. And this translates to higher prices for their computers to recoup the costs of manufacturing and R&D. This will further alienate users from buying Apple products and cause them to lose more money. If they ever do decide to make their own chips (God help them if they do) then sell all your stock because Apple is done.</strong><hr></blockquote>
???
Why would Apple need plants to manufacture it? When so many companies out there do nothing but manufacture other people's chips. Hell, IBM just opened up a new modern plant just for this. And there are plenty of overseas companies that would love to get the work. that is not a problem. The effort would be in the design. But with a map already laid out (Book E) and partners that would invest time to help (IBM and Motorola), experience in chip design, and bright people that flock to Apple in general, Apple would not be in a bad position to take up a project such as this.
[quote] and partners that would invest time to help (IBM and Motorola), <hr></blockquote>
I am no chip salesman and I do not own a chip company. But I do know I would not help a competitor develop a new superior product that would compete with and replace mine.
I am no chip salesman and I do not own a chip company. But I do know I would not help a competitor develop a new superior product that would compete with and replace mine.
Comments
Motorola embarrassed Apple badly with performance of the early G4
<hr></blockquote>
Actually when the G4 first came out it was the fastest chip around-by far-by a factor of two.
And, of course, all that AltiVec code that Apple and Adobe and others have worked on suddenly becomes useless.
<hr></blockquote>
...and any assembly code that has been used in optimizations as well-a real pain in the ass to port.
<strong>
Actually when the G4 first came out it was the fastest chip around-by far-by a factor of two.</strong><hr></blockquote>
Okay, okay, I was being too vague. The speed increase to 500 MHz, that had to be retracted and changed to 400 MHz, is the early embarrassing event I was referring to. I didn't want to imply first, but simply some time ago. Not recent.
Furthermore, it would set an Apple developed AMD-based computer (not a Mac, obviously!) apart from the IBM-compatible, Windows-compatible PC standard. I seriously believe that even by using standard components for most of architecture, Apple could produce a non-compatible x86 based model. A few central components (ASIC, and not just proprietary bootstrap ROMs), the very newest standards, and a very different architectural topology would do the trick.
The "Yellow Box" strategy was not viable - making Apple OS technology a mere subset in a Windows framework is an open invitation for MS to make borderline incompatibilities - but x86 hardware production is.
Of course, Classic would have to be abandoned and Carbon demoted in favor of Cocoa (which is better due to its object-code nature, rather than any cross-capability (though that, of course, is a secondary benefit of the first) - in other words, we are talking a last-ditch, white-of-the-eyes strategy.
And imagine the amount of crow it involves eating... basically a PR company's nightmare.
Since that will never happen I think this conversation is pointless. PPCs have always been a more elegant, better designed architecture than Intel chips, and even the industry mostly agrees with this, with PPC derivatives being used in all kinds of embedded and otherwise applications (gamecube anyone), and I don't see ARM complaining that they can't get good performance out of RISC technology.
The only reason Intel chips are faster is because they have more $ behind them. There's nothing wrong with PPC and thye are still holding their own even if Moto aren't doing as well as they could be right now.
There is no way in hell that Apple is going to adopt x86 for a meager speed increase at the cost of 99.9% of existing software ceasing to function on the Mac AT ALL. There was enough fuss when Adobe/Microsoft took 12 months to carbonise their stuff and that's when we could still all use the original software under classic.
Apple would not survive a switch to Intel, the developers would throw their hands up in disgust and walk out on them.
Not gunn DU IT...
Wuhten be PRUdent...
[/Dana Carvey as George Sr.]
It's official: "OS X on x86" has become the single dumbest Mac topic EVAR. And people STILL don't get it! Econ 101 - take it, learn it, love it.
Next topic, please.
[ 09-03-2002: Message edited by: Moogs ]</p>
<strong>Might I suggest that it would be possible to pair (or more) an AMD processor with a VMX-compliant DSP processor, using one of the fast new bus protocols specifically developed for inter-chip communications? That would allow AltiVec code rewriting to be a fairly simple task.
Furthermore, it would set an Apple developed AMD-based computer (not a Mac, obviously!) apart from the IBM-compatible, Windows-compatible PC standard.</strong><hr></blockquote>
If you're going to go reengineering AMD designs, why not strip off the x86 translation layer while you're at it and let that Alpha-inspired RISC core shine through?
You might even be able to put a (thinner) PPC translation layer over it and free up some room on the current die for those VMX units.
<strong>
If you're going to go reengineering AMD designs, why not strip off the x86 translation layer while you're at it and let that Alpha-inspired RISC core shine through?
You might even be able to put a (thinner) PPC translation layer over it and free up some room on the current die for those VMX units.
I had no intention of reengineering AMD's design - I am aware that some people find it feasible to do what you are suggesting but I doubt that it would happen in the "real" world (read: be financially feasible). Apple is not a large enough customer to give AMD enough return for such an outlay.
My suggestion was to take "Hammer", or whatever other future design might be suitable AND seem to be easy to develop easily, and build a non-PC compatible architecture around it, using a blend of standard components, the best upcoming technologies, and some central customs developed ASICS. Hammer would be paired with a VMX-compatible SEPARATE DSP processor, the two coupled with HyperTransport (or another suitable inter- (not intra-) chip bus standard.
engpjp
<strong>Do you guys take your hotdogs with ketchup or mustard, or both, or do you prefer mayonnaise, or maybe even curry sauce?
And do you prefer weeners or porc sausages?
Soft bread or Parisette?
G-News</strong><hr></blockquote>
polish snausage with hot mustard and mayo makes for a great meal and a lot of gas afterwards. that's where threads like this come from. lots of gas.
I'm interested to know, how many people decided they would never buy a PPC Mac because it was leaving the beloved 680x0 behind?
As long as it runs Mac OS, I don't care what the chip is.
Unfortunately, there IS a connection.....
engpjp
I reiterate: if anything they need to invest in an inhouse PPC project and have it fabbed by any American or Taiwanese manufacturer that will do it for the least amount with the best technology process. x86 is not the answer. It's a question.
I also saw that article about Apple having an x86 compatible version of OS X but if they ported it to x86 then they could also port it to another platform.
Someone mentioned Apple would be better off doing their own chip and I certainly agree with that.
Alf
(didn't mean for it to sound like a rant)
[ 09-10-2002: Message edited by: alfredh ]</p>
<strong>Someone mentioned Apple would be better off doing their own chip and I certainly agree with that.
Alf
(didn't mean for it to sound like a rant)
[ 09-10-2002: Message edited by: alfredh ]</strong><hr></blockquote>
It would be impossible for Apple to develop their own chips. They would need a heck of a lot more money. Just think, they'd need plants to manufacture it and people to design it. The worst part would be the R&D. And this translates to higher prices for their computers to recoup the costs of manufacturing and R&D. This will further alienate users from buying Apple products and cause them to lose more money. If they ever do decide to make their own chips (God help them if they do) then sell all your stock because Apple is done.
<strong>
It would be impossible for Apple to develop their own chips. They would need a heck of a lot more money. Just think, they'd need plants to manufacture it and people to design it. The worst part would be the R&D. And this translates to higher prices for their computers to recoup the costs of manufacturing and R&D. This will further alienate users from buying Apple products and cause them to lose more money. If they ever do decide to make their own chips (God help them if they do) then sell all your stock because Apple is done.</strong><hr></blockquote>
???
Why would Apple need plants to manufacture it? When so many companies out there do nothing but manufacture other people's chips. Hell, IBM just opened up a new modern plant just for this. And there are plenty of overseas companies that would love to get the work. that is not a problem. The effort would be in the design. But with a map already laid out (Book E) and partners that would invest time to help (IBM and Motorola), experience in chip design, and bright people that flock to Apple in general, Apple would not be in a bad position to take up a project such as this.
I am no chip salesman and I do not own a chip company. But I do know I would not help a competitor develop a new superior product that would compete with and replace mine.
[ 09-10-2002: Message edited by: FlashGordon ]
[ 09-10-2002: Message edited by: FlashGordon ]</p>
<strong>
I am no chip salesman and I do not own a chip company. But I do know I would not help a competitor develop a new superior product that would compete with and replace mine.
[ 09-10-2002: Message edited by: FlashGordon ]
[ 09-10-2002: Message edited by: FlashGordon ]</strong><hr></blockquote>
Really? IBM and Mot didn't seem to think so when they shared Somerset. If the relationship benefits either one they have an incentive.
[quote]... the developers would throw their hands up in disgust and walk out on them.<hr></blockquote>
uh huh, and if the machines aren't fast enough, replace the word "developers" with the word "consumers".
neye