Got a tip for us?
Let us know
Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.
AMD and Apple Rumors Part II
- Thread starter MacRumors
- Start date
- Sort by reaction score
Right and how has the PowerPC G4 changed significantly since Motorola introduced it?
Since the G4 introduction, we've seen Pentium III, Pentium 4, Athlon, Athlon XP, Athlon 64/Opteron. All with significant changes. P4 just got HyperThreading, for example; AMD going 64bit as another example (in desktop).
What, it'll take us ANOTHER year to come out with a decent PowerPC?
[edit] BTW, how different is it for Motorola and IBM to maintain PPC backwards compatibility in their lines (i.e. you can still run 601 code on your dual G4) than it is for Intel to maintain the x86 line.
Ahhh, we are fortunate to have bluecell. Someone who can take all of the potential issues with a switch in processors and simplify them to a single concept, performance. Through rational thought, the path is clear, we must switch to AMD chips. The huge increase in performance has been proven. There are no rational reasons for which we should use the 970. And the G4, it isn't suitable to power my digital watch. Developers, they are of no concern. They will do what Apple tells them, and they will do it in a timely manner.
Such clear insight into an issue is rarely seen. Long have I thought about what would be the proper solution for apple's predicament, but I just could not reach a decision as to whether or not many of the issues, previously posted in this thread, preclude the switch to AMD processors. But now I am certain. No thought of those issues is necessary. Indeed, we have all the information we need. Apple, IBM, and AMD have no additional insight into this matter. For no more is required. The case for a switch to AMD could not be any clearer if it were flashed in front of our faces 85 times a second.
To facilitate this wisdom reaching the apple decision making machine, I have created a new site. http://newappleceo.org. This site will be dedicated to convincing Steve Jobs that he should step down as CEO and hire bluecell as his replacement.
Join me macforums readers, we, the mac faithful, need this.
crackpip
Arg...
...the Itanium 2 costs over 1000 dollars per chip. It's a SERVER chip. Don't compare it to the 970. It competes with the Power4+ (which scores higher than it, btw, and is PowerPC)
Also, IBM only makes G3s for Apple, so they can't possibly have been "crippling" Apple.
As for OSX not being PowerPC optimized, "every pixel on the screen touches the Altivec unit" (Steve Jobs. Quote may not be exactly right). Altivec is PowerPC only.
About the 970 not beating the Athlon64, perhaps you haven't noticed, but NEITHER OF THEM ARE OUT YET. You're basing your comparisons off of wishful thinking and marketing numbers. In fact, SPEC numbers for the Athlon64 haven't been released, so I'm not sure how you're comparing the two processors (SPEC numbers were released for the Opteron, which is a SERVER CHIP).
Lastly, AMD just gave up (see my earlier post). They have stated that they are not going to compete with Intel on performance anymore. That leaves a choice between the P4 (32 bit, power hungry, no multiprocessing), the Itanium (incredibly expensive and power hungry), the PowerPC 970 (low power, fast, multiprocessing capable, massive company backing it, more versions planned), and something Moto may come out with (who knows).
Basically, most of the stuff you just said was either made up, incorrect, useless, or complete speculation. Go get some real information.
...the Itanium 2 costs over 1000 dollars per chip. It's a SERVER chip. Don't compare it to the 970. It competes with the Power4+ (which scores higher than it, btw, and is PowerPC)
Also, IBM only makes G3s for Apple, so they can't possibly have been "crippling" Apple.
As for OSX not being PowerPC optimized, "every pixel on the screen touches the Altivec unit" (Steve Jobs. Quote may not be exactly right). Altivec is PowerPC only.
About the 970 not beating the Athlon64, perhaps you haven't noticed, but NEITHER OF THEM ARE OUT YET. You're basing your comparisons off of wishful thinking and marketing numbers. In fact, SPEC numbers for the Athlon64 haven't been released, so I'm not sure how you're comparing the two processors (SPEC numbers were released for the Opteron, which is a SERVER CHIP).
Lastly, AMD just gave up (see my earlier post). They have stated that they are not going to compete with Intel on performance anymore. That leaves a choice between the P4 (32 bit, power hungry, no multiprocessing), the Itanium (incredibly expensive and power hungry), the PowerPC 970 (low power, fast, multiprocessing capable, massive company backing it, more versions planned), and something Moto may come out with (who knows).
Basically, most of the stuff you just said was either made up, incorrect, useless, or complete speculation. Go get some real information.
Apparently you didn't notice...
...but the change from G4 to G4+ was MAJOR redesign (lengthening the pipelines and adding vector dispatch units is very non-trivial). The P3 uses the same core (P6) as the Pentium Pro and Pentium 2. The only major redesigns in x86 have been the Athlon and the Pentium 4. Also, the 970 comes out in less than a year, AMD's 64 bit stuff is massively delayed, and HyperThreading actually hurts performance in many cases.
Maintaining backwards compatibility is similar, but x86 is about a decade older (at least) than PowerPC. 10 years from now, PowerPC will have similar problems to what x86 is having now.
Originally posted by MacCoaster
Right and how has the PowerPC G4 changed significantly since Motorola introduced it?
Since the G4 introduction, we've seen Pentium III, Pentium 4, Athlon, Athlon XP, Athlon 64/Opteron. All with significant changes. P4 just got HyperThreading, for example; AMD going 64bit as another example (in desktop).
What, it'll take us ANOTHER year to come out with a decent PowerPC?
[edit] BTW, how different is it for Motorola and IBM to maintain PPC backwards compatibility in their lines (i.e. you can still run 601 code on your dual G4) than it is for Intel to maintain the x86 line.
...but the change from G4 to G4+ was MAJOR redesign (lengthening the pipelines and adding vector dispatch units is very non-trivial). The P3 uses the same core (P6) as the Pentium Pro and Pentium 2. The only major redesigns in x86 have been the Athlon and the Pentium 4. Also, the 970 comes out in less than a year, AMD's 64 bit stuff is massively delayed, and HyperThreading actually hurts performance in many cases.
Maintaining backwards compatibility is similar, but x86 is about a decade older (at least) than PowerPC. 10 years from now, PowerPC will have similar problems to what x86 is having now.
Like someone else said - the P4 is the first real re-designe since the P-Pro. And, IBM tried and dumped HyperThreading a while ago in the Power series because the down sides of it were greater that it's up side.
Because the PowerPC is more than a decade newer than the x86.
Re: Arg...
No, everything I said has been FACTUAL. The PPC 970 won't go into production until late 2003, but the SPECs have been up for a while (check Architosh). I have seen the SPECs for the AMDs and they are here. AMD's processors will be out Q1 2003. AMD DID NOT GIVE UP. Check out the Hector Ruiz's keynote on C|NET. And about the marketing numbers-last time I checked Apple only had about 5%. You completely missed the boat.
Re: Re: Arg...
1) The SPEC numbers you linked to are for the Opteron which is not a desktop chip. Read your own links. If you can find Athlon64 SPEC numbers, please post them. I'm quite interested in how it's going to do.
2) Marketing numbers, in this context, means made up estimates of SPEC scores to keep investors happy, not market share.
3) AMD giving up: http://slashdot.org/articles/02/11/24/0556252.shtml?tid=142
Anything else?
1) The SPEC numbers you linked to are for the Opteron which is not a desktop chip. Read your own links. If you can find Athlon64 SPEC numbers, please post them. I'm quite interested in how it's going to do.
2) Marketing numbers, in this context, means made up estimates of SPEC scores to keep investors happy, not market share.
3) AMD giving up: http://slashdot.org/articles/02/11/24/0556252.shtml?tid=142
Anything else?
There's no way Apple will make its OS incompatible with G3 for at least the next 3 years. Not while Apple is still pushing G3-equipped iBooks and CRT iMacs to educational customers. I mean, they just caused enough disruption to their customer base by moving to a non-backward compatible OS X. There's no way they'll do anything similarly disruptive for a long, long time. Same reason why there's absolutely no chance that they're moving to the x86 platform.
Even if Apple started losing massive market share in the professional Graphics, Audio, Video, etc. markets that actually need t he horsepower, they'll go to extreme measures to stay with the PowerPC, since not doing so would be a marketing disaster. So at the high end they'd either have to pack in even more processors (Quad-CPU or even eight-way). Or maybe even start using the Power4 before the 970 comes out.
That being said, I actually feel pretty good about IBM and the 970. I don't understand why people are so pessimistic. Keep in mind, IBM has said that they're going to use the 970 themselves to make lower-end (relative to the Power4) Linux machines. a very different situtation from Motorola, where building CPUs is at best a distraction from their core business(es). I believe that IBM going to pull through with this, perhaps even ahead of schedule. I think at this time next year, IBM will be launching a two-pronged assault on Intel: trying to take a chunk of the low to mid-range Linux server market share (the existing high-end Power4 servers can already be configured with Suse Linux Enterprise), and helping Apple make inroads in desktop market.
Which is why its important to reiterate- (I think this is the third time I've said this)-- there is nothing that says any business between Apple and AMD has to be x86 based. AMD can easily fab other chips.
The other downside of Apple moving to an x86 chip at this time-- any one buying a machine with such a chip inside, would want to install Windows on it as well, due to the amount of software available written for x86s. There would be a few people with Intels wanting to use OS X -along with windows, red hat, whatever, but not as many that would merely not bother. It would certainly give developers a reason to stop coding for Mac OS X altogether, all you would do is give Intel boxes a unix with a decent GUI
Too much of Apples income is dependant on Hardware sales, not software. They simply can compete against a behemoth 20 times their size on OS sales alone.
Re: Running multiple OSes on x86
That is a good thing to know, but it still wasn't what I was saying, which is why I was rather short.
I hope that the question was even raised does not mean someone thinks that multiple OSes were somehow not possible. Their simultaneous use was not an issue. Even if you did have to reboot, people really don't mind rebooting to use another OS. They may be a lot less likely to go to the trouble however.
But I doubt anyone would care that much about developing for MacOS if it moved to an x86-- anyone with an x86 machine is going to want the option to boot into Windows, even if they want to stay in *nix all of the time, or even a *nix with a nice gui, like OS X. Or even running Windows on top of that *nix.
The point is, development would pretty much stop for OS X then, and every developer would stick with developing only for windows. Red Rover has come over, game over, fire the mac guys it all runs on intel now. OS X on an x86 would be running blue box [or whatever color 'windows' on the mach kernal bed is] , much more often than we are running Classic. And why run in such a mode, when you can be native in windows? If a better interface was all that people wanted, a lot more people would be using BeOS on x86.
AMD can do things for Apple that have nothing to do with having an x86 legacy on a chip- they give a damn about making, and can make, chips in volume and on time, which is more than you can say for Motorola. This doesn't make them a more obvious choice than IBM, but I say it is very possible for them to be a player.
That is a good thing to know, but it still wasn't what I was saying, which is why I was rather short.
I hope that the question was even raised does not mean someone thinks that multiple OSes were somehow not possible. Their simultaneous use was not an issue. Even if you did have to reboot, people really don't mind rebooting to use another OS. They may be a lot less likely to go to the trouble however.
But I doubt anyone would care that much about developing for MacOS if it moved to an x86-- anyone with an x86 machine is going to want the option to boot into Windows, even if they want to stay in *nix all of the time, or even a *nix with a nice gui, like OS X. Or even running Windows on top of that *nix.
The point is, development would pretty much stop for OS X then, and every developer would stick with developing only for windows. Red Rover has come over, game over, fire the mac guys it all runs on intel now. OS X on an x86 would be running blue box [or whatever color 'windows' on the mach kernal bed is] , much more often than we are running Classic. And why run in such a mode, when you can be native in windows? If a better interface was all that people wanted, a lot more people would be using BeOS on x86.
AMD can do things for Apple that have nothing to do with having an x86 legacy on a chip- they give a damn about making, and can make, chips in volume and on time, which is more than you can say for Motorola. This doesn't make them a more obvious choice than IBM, but I say it is very possible for them to be a player.