Intel-based Macs coming soon? - Page 3

I think you may be right about this. I just can't see Apple moving to x86 at this point. They've spent years touting the superiority of PPC. Then there are the programming and compiling issues. The obstacles are enormous and the whole plan doesn't see to have much benefit.

That is unless, as you said, Steve through a fit and went shopping. I will believe this one when I see it.

This could also be a giant ploy by Apple that has several effects:

1) Creates industry buzz before announcement

2) Makes CNet look like the hack job it is....years of negative Apple press? There's a solution for that and it's a nice unsubstantiated and perhaps totally inaccurate story. Apple would have to take pleasure at CNet losing all credibility and having to report on a Intel-destroying IBM chip at the same time, especially after years of biased reporting.

perhaps Apple are going to make a hybrid MacOS X PPC/ Wintel machine for the switchers.

There's no cost advantage to going with X86, so what is the advantage? Same dev cycle as the competition is all I can see, and that's only if Apple's got some swagger in their step and a big brass set...

No real performance advantages, no cost advantages, so why? What will an X86 Mac let me do that a current mac can't do? If Apple wanted to build cheap macs, they could do that with PPCs.

I'm not saying that this rumor is impossible, but it has to be thought about more carefully than what's going on here.

The future of chip design may very well be that a few large competent bidders build essentially modular/custom product to order for a host of segments.

I just don't see Apple on X86, or maybe even on PPC. The future may be that neither Windows or Apple run on anything currently recognizable as a standard platform -- the cost advantages of that have long since disappeared. There's so much competency now, and so much capacity (in a few places) that you may as well just order what you want as opposed to buy "off the shelf"

I've had a better idea.

The number of transistors on a 970 is +/- 50 million.

The number of transistors on a Pentium DualCore is over 200 million.

SO.

The Big announcement on Monday, is that IBM have not developed a dual core 970 at all

ITS GOING TO BE A QUAD CORE 970

970MP (MULTI Core) not 970DP (DualCore)

And it's going to be SMT, so each 970MP has 8 logical CPU's

and then Apple are going to ship dual processor machines,

and thus PowerMacs are going to have 16 processors.

Eat that Intel.

Because Apple wouldn't let anyone run OS X on non-Apple hardware, nor Windows on Apple hardware. Just slap a custom x86 chipset in the Mac and instantly Apple controls what runs on it, and where OS X can run.

Maybe Apple could even offer an x86 mobo with added functionality, that is only available on a Mac. Patent it and the platform might as well be PPC when it comes to compatibility.

I just love how people not affiliated with any of these software houses just 'know' how long a transistion would take. Unless you have access to their source code, you have no idea how long it will take. You can make assumptions, but any large software where speed matters is going to be partially written to use the underlying chip tech. And all of that would have to be re-written.


What does 'legally' mean. Apple would have to do some major voodoo to their OS to keep it from being hacked (remember Darwin already runs on Intel) and runnable on non-Apple macs. Hell, Apple can't even keep people from installing OS X on non-supported Macs now. Unless they're really careful, it will be hacked, and then you've got a huge hole where Apple's hardware sales (their lifeblood) used to be.


First, who says it would run Windows? Second, WHY would they switch to OS X, just to pay more for hardware that still will be incompatible with most of their software and devices (Intel support does not majically bring driver support for all those current Windows-only peripherals!). This has been tried by many people before. BeOS, OS/2. No one switched then. What makes you think that OS X will just march on in and people will be clamoring for it? Hell, when OS X came out, most people still booted into OS 9, because that's where the apps were. It took forever to get them to switch to OS X, as they were comfortable where they were. What makes you think Windows users are somehow more likely to switch because its on an Intel chip? Its just plain stupid thinking.


Yes, that's right, you'll pay a premium for the advantage of booting into windows? I don't think so. And if Apple ports their iLife suite to Windows, what's the point of OS X? Man, that's just a stupid comment. "Hey, I can get all these great apple apps for Windows, but, you know, I'm just going to spend $1500 on an Apple Mac/Windows dual-boot box and get iLife for OS X instead!" [Oh, and iLife will cost a lot more for Windows users than macs, since its used on the Mac to try to sell computers, not to make money]


But in the PC world, people don't spend more than they have to for a computer (with all that competition out there, you know). Macs are not going to sell anymore then they do now just because its got an intel chip on it.


Yeah, but would everyone? Or will they just say "screw it, you buy either the PPC version or an Intel version, you don't get both", like they started to do when we switched 10 years ago.

The only thing that keeps apple afloat is its loyal customers. This move would be just another long series of moves that just drives their loyal fanbase away. How many times do we as computer users need to keep updating our hardware and software because of Apple's transitions? First its System 1-6 to System 7. Then Switchover to PPC. Then switchover to OS X. Its getting expensive.

Meanwhile, PC users are using the same software and hardware platforms for the last 15 years. You don't have to worry that if you get a new computer, you're hardware all of a sudden is useless, you peripherals no longer work, your software needs to be updated. Laugh all you want, but more mac users will say "Enough, I'm switching to a 'stable' platform" rather than those who'll go "Cool, look, I can get a Mac with an intel processor and OS X. Can't do anything with it, but I can get one!"

One last thing. Some people seem to think making this announcement at WWDC makes sense. Yet, I don't. Follow. People are heading to WWDC to take seminars and such on developing for OS X/Tiger. They show up, all set knowing what they're going to be learning (or not going, because there's nothing compelling for them to show up this year). Then Apple drops a "we're going to Intel" bombshell.

Now you've got developers who aren't caring about the stuff they went their for, now they're stuck trying to find out as much as they can about developing for Intel. No matter what's being discussed, questions will be "How's this going to work with the CPU switchover" and stuff like that. Those who didn't show up are screwed beyond belief.

My thinking is this should've been announced months ago, so attendees could be prepared for what they're showing up for. [Of course, this is one reason Apple is so disliked by developers, they never tell them anything up front, as opposed to, oh, I don't know, pretty much everyone else, who've got roadmaps for as long as the eye can see]

Or is this just stupid thinking on my part?

I think just for the processors this would be a bad idea, but for an entire chip-set that would be great. I don't know how much money Apple had in the 'Apple chip' but it had to be a huge development cost. The reason for this chip was to add all of the interconnecting chip sets into one huge chip, firewire, PCI, the list goes on. Moving to Intel and suddenly Intel is doing this for in the normal chip-set that they ship, and Intel would be willing to work with Apple on hardware design. Apple would get integrated WiFi, PCI, USB, etc., economy of scale dictates that this is a path of least cost, if looking at the BIG picture. Yes CPUs cost about the same, but on the Intel side prices drop faster, whole chip-sets are designed for consumer PCs whether Apple wants it or not. For the last 20 years Intel has wanted Apple for a customer, Steve is friends with Andy Grove, the cost to Apple would be less especially for new designs. Look at USB Intel had it out on the market but MS would not write the drivers for it so no one would build devices that used it. Apple could see that with this they would get a forward looking technology not just something rehashed. Intel can see that they are getting a customer that is wanting to sell computers that are where the market is going later, an innovator. MS cannot do this because they are bogged down with security issues, longhorn, legacy, and a huge market the is fractured all over at best but each area needs attention. Most folks need a P4 just to run all of the spyware. Intel wants to add hardware to their MB designs but it is difficult if your largest customer will not support that technology because they have their own issues to deal with. With Apple Intel would get a customer that wants to innovate not just to be different but to add value to their customers computers, WiFi is a good example. Andy said that Apple could not beat the price/cost of having Intel as a partner. Whether it happens or not, there are some good business reasons for the change, for both companies.

I actually agree with your line of thinking. Apple going to Intel would be very upsetting, unless there is some sort of really amazing plan of action that none of us could even imagine.

We seem to be getting the same old tired arguments as to how good this switch would, without the understanding as to how difficult it would be.

To pretend that this would be simple is to be delusional.

For those who continue to think that this would simply require a re-compile "a day" I think someone here said. That's preposterous!

I beta test Photoshop and I can tell you that just upgrading the program for the SAME platform is difficult enough. Remember that Adobe and other writers of complex performance enhanced programs use the hand work to squeeze performance out of their software. Often, the versions of the code for different platforms has little in common, even if it's the same OS. If it's different OSes, it's even worse.

Even Linux isn't truly portable-and that's on the SAME platform.

The only software that's basically portable is Posix compliant. Some Unix variations are. The rest aren't. Apple isn't as yet, but is working towards it. If that ever happened it would be easier. But not now. Unless, of course, that Intel virtualization chip comes out.

As for the arguments about consumers caring about the chip, it doesn't matter, it's the developers who matter.

The most likely thing (it seems to me) is that there are multiple underlying "facts" that have gotten "crossed" to form this rumor and that Apple is indeed planning something around "Intel chips" (whatever that might mean)...but perhaps a new, different kind of device.

I think we need to stop thinking about Apple as a "computer" company. It seems clear that they have. They haven't (publically) called themselves "Apple Computer" for a while (it is just "Apple")...and they have diversified (software, iPod, etc.) and are probably looking for even more ways to diversify their product portfolio. As they do this kind of thing, they have many options (remember that when they went from Apple II (6502 chip) to Lisa/Mac they changed chip architectures). Remember also that for iPod, they didn't write the OS themselves (it is no variation of anything they have written themselves) or used an existing chip architecture (ARM vs. anything else).

This could be the basis of something completely new...but with a "bridge" (some scaled back version of OS X) to the "past". Maybe this rumored "tablet" thing fits in here. Maybe there is some new device with a ported/scaled-back OS X, using some "Intel chip"...who knows.

About the PPC chips themselves;

The 2.7GHz is pretty much at a par with the Opterons and Xenons. A little better here, a little worse there.

As we don't know the progress of the GX series, we can't say what's going to happen over the next few months. I'm surprised that the MP chips based on the GX haven't come out already. But remember that both Intel's and AMD's chips haven't increased in speed for months, and are pretty much stuck where they are until 65nm comes out. Despite the claims of some in the industry, it's not known if 65nm will be ready on time, or if the problems have been solved. The interesting thing here is that the problems occuring in 90nm are much worse in 65nm. What I've been thinking about is if they've *solved* the problems for 65nm, they should be able to provide the same benefits to 90nm. So why haven't we seen an improvement here?

If the GX is somewhat on track, and we see it later this year, it would move Apple beyond x86, as it's supposed to come in at 3GHZ. Remember x86 duals are running slower than the single units. Even if a dual PPC came out it would perform better than x86 units. Also if IBM finally upped the L2 cache to at least 1MB, that would make a big difference.

If IBM can't supply this however, then there could be trouble.

AMD's chips have been getting a steady speed increase. The next chip in the FX series is expected to be a 2.8ghz.

I'm not sure what dual core chips your looking at the the Athlon 64 X2's are running on par with the single core chips...

I would think a move to Intel, the way it's described, would be a hard sell to developers--a "phased" release would mean that any developers that care about performance would have to release TWO versions of their app--one for PPC, one for Intel. Not only that, they'd have to buy new computers to test with. It would be hard to believe that the emulation would be decent. If I remember correctly, the new PPC (601s) were faster than the fastest 68K chips, so with emulation the 68K code was only slightly slower on those PPCs. It was a lot faster with "native" PPC code. However, from all the app comparisions, an optimized G5 app is generally as fast as the high end Intel app. So, I'd be surprised if Apple could demonstrate any (or much of a) performance gain. And given that Intel has been increasing its performance slower than IBM, the last few years, I don't see how Apple could promise anything spectacular in the futurre, either.
So all in all, I don't see much gain for all that pain. IBM still seems to have some "legs", although it has stumbled in the 90nm transition. I don't see that Intel has done any better. There would have to be some "carrot" or stick that no one's heard about--either Intel has something quite hidden up its sleeve or IBM is dropping support of Apple to focucs on the gaming (low end) and server (high end) business. Going with Intel seems to make no sense, given what is known

Hence are Rumors...................

That's funny, I pretty clearly said they were sophisticated. Internally their chips are obviously leading the industry. The x86 ISA at this point has so many modes, weird encoding nuances, register limitations, limited cache controls, etc.... worst of all is the number of variations. It is really a whole set of related ISAs.


Did I say it was? Intel owns IA-64, they built the 860/960 in the early 90s (if I recall the numbers correctly -- that was long ago), they own Xscale, and didn't they end up with Alpha? Or was that just some of the Alpha tech & talent? Intel's whole story is a lot more than just x86, but x86 does dominate what people know about them and their income.


Its also not wanting to get buried here in technical argument that has been raging for 20 years. The only real answer we'd ever get to this argument would be if Intel actual went ahead and did this and (for the first time) we saw the impact of different ISAs on exactly the same hardware. That's interesting enough that I think they ought to do that just to get an answer.


What will keep Apple from switching is the pain inherent in the transition, and the lack of enough upside to overcome the pain.

SSE3 does have a few inherent limitations in its ISA that make it inferior to VMX, its not just implementation. A key one is that the lack of registers in vectorized code is a major issue, and evidence of this can be seen in the adoption of 128 registers by XBox360 and the Cell SPEs. As for the importance of SIMD in the future... well, we'll just have to wait and see. Until then we'll each have our own opinions.


I didn't say that they would use the P4, did I? And look, the rumor points to a late 2006 transition. That's a year from now, and we don't know precisely when they started. Hmmm... that's some time. And Intel has some new cores coming. Hmmm... that's a new chip.


So what's this, are you agreeing with me?

I hear what people are saying about diversifying their product portfolio, that the iPod doesn't have the same OS or chip architechture, etc. And I heartily agree with those people.

Could this be the solution for the long rumored and now ripe Apple iHome Media Center Solution?

Based on Intel processors to keep down price and meet what will be a drastic demand, Apple can finally conquer the living room. Set top box much like a Tivo, WiMax, huge hard drive, H.264, wirelessly stream music, photos, and movies from the iTunes Music/Video Store, HD ready, content on demand, content from your computer, home videos and DRM home movies...

It could be intense.

You got that right. If this news is true, a dual-2.3 PM will be my last Mac. I refuse to pay Apple prices for Intel junk. (Why switch to Intel now, just as they are losing in every way to AMD, and the hot new video games are using PPC? It's like deliberately picking the slowest horse. Unfortunately, that's what makes it so believable. )

Edit: Forgot to mention that, yes, Intel does have the rights to manufacture Alpha, and then there's XScale (formerly known as the Acorn RISC Machine). There's the 860 RISC chip, and 960 and 8051 and other embedded chips. And of course Itanium. Alpha is an intriguing possibility, but not worth putting Apple's customers through yet another transition.

Here's my 2 cents. Maybe Intel is providing a custom OEM daughter card for Apple that will allow the Mac to have the PPC chip and the x86 chip in the same box, with lo level software that recognizes the difference. Now imagine being able to run Windows Programs NATIVELY on the Mac. Insert a PC install disc and it "just works". Developers have a choice... Mac or PC. Launch MS Access or some other MS only program, and it just runs... albeit with some ugly interface leftovers for the program. This becomes the final incentive for PC users to switch to Mac software when they see how the Ugly MS counterparts look. Mac users get the best of both worlds. Virtual PC is history. Your Mac is simply THE computer. Software platforms become a thing of the past. All the Windows people could start using Macs and not lose their software investment. Switchers in the droves. Intel makes more money, MS makes less and less. Dell can jump in a lake.

The oerating system is 100% Mac with custom software directing the CPU calls to the right CPU.

Thoughts??

I have to admit i have not had the chance to read through the entire thread yet so this may have already been brought up, but one of the main reasons apple may be announcing this so early, much in contrast to their normal style, is because they will need help from outside companies to have software ready for it. What if, on top of that, one of the goals of this transition is to address the uneducated cries that apple lacks software variety? They could be trying to have a whole range of typically pc only software available upon release. But seriously, I do think they would need to give a heads up to all of the third-party software companies to give them time to prep for it, and the likelyhood of them being able to control the mouths of that many people is not very likely. So their plan: announce it now to build hype and keep people thinking about where apple is going instead of where longhorn is supposedly going to go.

Just my two cents, i'm not taking a side as to whether or not I belive it, though I have to admit, i would welcome it, since my Imac G3 several years ago I have not been able to afford a new mac, and even with the release of the mac mini, i'm still working on a home built x86 system running Debian. If i could just plop down a copy of OSX on this comp I would die a happy man!

Well, how would switching to Intel CPUs get software developers to develop for OS X. Even taking the "there's nothing to a recompile" folks at their word, they'd still have to completely write the software, something they haven't had the urge to do yet (its not a recompile of windows software, they'd still have to code it to Cocoa/Carbon). I don't see how this announcement would somehow drive Intuit, say, to making an OS X version of Quicken on par with the Windows version.

And this whole "OS X vs. Longhorn" thing is just nothing. Most people don't care about longhorn as much as they don't care about OS X. They just need their email and browser to work. And they aren't going to switch OSes just to get where they already are. Hell, no one used Next when it was on PCs, why would OS X be any different.

This would still be an Apple machine. Apple could always use AMD chips if they saw a need, if this would be a port to x86 itself and not something else. I wouldn't worry there. Apple wouldn't do it if they didn't think it would be better.

Still think it won't happen though.

This is always possible. Apple and others supplied PC cards before. If they could get the price right this time, it might make sense.

Don't think about dying just yet then. If Apple does this it won't be an Intel mobo box. You can be quaranteed that it will be exclusively Apple's Mobo, ROM, etc. Nobody will be dropping Windows on it, and you won't be able to drop X into another box.

Remember, Apple is a hardware company. They are not going to lose almost all of their sales.

This year Apple might break $1 billion in software sales for the first time. That includes sales of OS X. Apple's total sales this year might reach $15 billion.

Oh yes you would. It'd require a bit of effort from the MOL guys, but it will be very doable.

You pretty clearly said a lot of things - the general gist of which is plain uncomplimentary to x86. Then this rumour comes up and all of a sudden the x86 cores are considered "sophisticated". Ahem. Is that egg all over...


You raised the possibility of Intel sticking a PPC decoder in an x86. I was quoting a /. comment which mentioned Intel fabbing a PPC for Apple. That's different. And its more plausible.

Look, I am just enjoying the sight of you lot squirming. Even if the switch does not come to pass, which is likely, reading people's rationalizations, backpedalling, disbelief, etc has been hugely entertaining. The creative attempts at explaining away the Switch have been hilarious.

Just let this be a lesson. In future, before badtalking your competition, consider the fact that you may end up in bed with them. You know, like switching

Reasons I don't see this happening...

Mac x86 wouldn't happen because of all the work done on Altivec. Everything Apple has done to speed up their apps for Altivec (and that's a lot of stuff) would just be thrown away, along with the problems with emulating older apps, etc. It would kill the Mac and all its software. Not to mention, why would anyone see a premium on Macs with Intel procs in them when Dell sells PC's with Intel's in for sigificantly less? Big marketing hurdle there. At least now they can say PPC blows away Intel (whether it's true or not).

Intel making PPC I can't see being a reality simply because we haven't heard anything about it. Everything we've heard is Apple to x86. Not one person outside of rumor forums has suggested Intel making PPC chips. If they were working on it, you'd think there would at least be something leaked. Apple would have had to keep their own mouths shut about it, and keep Intel's shut as well. I can't see something as big as this happening without anyone hearing a word about it (we've at least had wind of a lot of rumors before they were announced recently, this is bigger than almost any of them. You think it wouldn't slip out?).

I have a question about a potential move to Intel chips.

Does this mean that future Apple computers will have DRM in their chips?

one company might shed light to all this TRANSITIVE. http://www.transitive.com/technology.htm. Thats all im gonna say about this matter
and if this is correct ( apple moving to intel ) its this company thats gonna be front in center on this matter.

Yes, I hate the x86 ISA and think it sucks. As you pointed out, however, there is a difference between the architecture and the chip. Intel and AMD have built very sophisticated cores in order to make x86 perform well. In my opinion sticking a decent ISA on them will make them perform even better (and will save a lot of pain for Apple and its developers). I don't see anything inconsistent about my position. I do see that you're trying your best to interpret what I say so that it looks inconsistent.


Okay, which PPC are they going to fab? The one IBM owns, or the one Motorola owns? Neither of which get top marks for performance, and neither of which has much of a future if the original designers don't pursue it. If it is just a fab arrangement, why wait until late 2006? I don't know a lot about the fab process, but I was under the impression that fabs aren't fully "compatible" due to process differences so its not just a simple matter of sending the design to Intel to be fabbed -- it would have to be re-cast on Intel's process and would become essentially a new chip. Might even be more work that just replacing the instruction decoder (which is a relatively small %age of the chip in the latest designs).


No squirming going on here -- just trying to make sense of a very old rumor that seems to have some weight behind it this time. I haven't rationalized or backpedalled, but I do look upon the simple "X on x86" with disbelief because it doesn't seem the sensible thing for Apple to do.


Tell Apple that. If the "X on x86" is the real path Steve has chosen, he is going to have some fancy footwork to do.

Check it out here on how many platforms Linux runs currently, and then make such statements.

Okay, kids, things are getting very interesting indeed:

Scobelizer





Either Scobel's dead on, or he's gonna be laughed out of the blogosphere. I can't see him getting laughed out of the blogosphere. My old news instincts seem to suggest that, as wacky as it sounds, we could see Mac OS X on x86.


Also, according to Ars Technica, Apple has offered a Czech programmer a job, saying they were "Seeking ACPI & BIOS person."


Things are heating up. Monday I'll look back at this post and wince.

In case anyone has not noticed:

The Appleinsider news page now says that this report has now been confirmed by The Wall Street Journal. Still no word as to whether it will be Intel x86 or a new Intel PPC.

well if it does happen its probbly best they take down this webpage

http://www.apple.com/powermac/performance/

or they sure will look stupid

"website shows how crappy intel processor are"

I'm still very skeptical on this one. I think Thinksecret would have dug something up on this, but their website they say their sources can't confirm this. With their track record I would think they would have confirmation.

x86 doesn't make much sense. I can see an alliance with Intel on something else, maybe a chipset or something? Could Intel fab an Apple designed chip, essentially like what IBM is doing for a lot of people now?

My sources at IBM have said nothing of Apple dumping IBM. I do know that Apple has 970MPs in their posession and the run at 3GHZ at least with 1MB of L2 per core. Apple had been unhappy with supplies earlier, but are fairly happy now - especially with multi-core G5s on the way.

Could this be an intentional bad leak? Is this Intel's people saying this, thus why Thinksecret has nothing, because Apple isn't doing anything with Intel? Is it a cover up for a 970MP unveiling, especially if those MPs are too hit sooner rather than later?

It actually doesn't say anything about Intel it says Pentium. I'd say the chances are high that future Macs will not be running a Pentium chip as we know it... but rather a brand new Intel PPC chip designed specifically for Apple.

My guess is as good as anyones, But...

Would the Wall Street Journal really go out on a limb and confirm this report if they were not confident they had a reputable source?

Why would he have to be "laughed out of the blogosphere"? 95% of the "news" reported this day and age IS bunk news. Nothing is clear and concise anymre and no one backs anything up either. Most news today, whether it's world news, or technological news, is based on rumor and speculation. If Scobel is wrong, which I'm sure he probably is, he'll be laughed at breifly and fade away back to doing his yellow reporting like all the other news reporters and news agencies out there. The only person I've ever heard of actually getting punished for doing bunk news is Dan Rather and his cronies. Heck, most of the reporting done on ZDNet is laughed at even by the M$ crowd so I never take any of these rumors and speculations seriously. Most people here can use some simple logic and deduce that Apple moving all the way over to x86 would destroy the company and all the software companies that write OSX software. Of course it's plausible but it's not economically or politically probable whatsoever. Not to mention is Jobs was wanting to do this to expand the user base then we can all kiss our virus-free mac OS's goodbye. 50% MS and 50% OSX = 50% virus development for MS 50% virus development for OSX. Also, just because another newspaper "confirms" what another paper has reported doesn't make all that more true. It probably just means they have the same "source" or even worse are just jumping on the rumormill bandwagon.

This is from the Appleinsider new page:

UPDATE: The Wall Street Journal on Saturday confirmed CNET's report, stating that an industry executive "familiar with the matter" verified the schedule outlined in the story.