PowerMac on Intel, the Beast Cuts Loose
If you have been paying attention lately you know that Apple has switched to Intel processors. Furthermore, in the next two years Macs will begin to come out with various Intel chips inside some of which are designed for speed (think iMac) and some for their low heat emmsions and frugal power consumption (think iBook). And of course several years down the road a new and improved PowerMac will roll out sporting Intel’s latest and greatest chipset. So, what will happen after that?
Think about it this way, as of right now the fastest PowerMac G5 compares favorably to the fastest Dell. Now, I realize that the Dell does somethings better (like games) and at the same time the G5 does some things better (like vector processing) but by and large the experience is the same for your average user. With me so far? So, as of right now things are almost tied with a top of the line Dell running XP being a little bit faster in several key areas. But what happens in two years?
What happens when Apple releases a PowerMac with the most powerful Intel chip it can find in it? Well, OS X will then seriously spank Windows in many instances. Now, I am sure that there will be some programs that will still work better/faster under Windows but that will be due to the fact that they are optimized for XP. But that advantage will eventually fade in time. Realistically Apple will be using chips that Intel has already designed. You should be able to go out and get the same chip in a Dell or an HP workstation. Hold on though, we aren’t done yet.
Now comes the really interesting part of this article. What happens when Apple tells Intel to go nuts, cut loose and design the best chip they can. Remember that with OS X to work with Intel doesn’t have to bend over backwards making their chips comply with old and outdated Windows specs. In short, what happens when Intel releases a chipset that won’t run Windows?
Think about what would happen if the newest PowerMac ran Intel chips that were so advanced that Microsoft’s clunky, outdated Windows Vista just couldn’t handle it? (and yes, based on the Longhorn fiasco I anticipate Vista still being the latest OS in 6 years or so) What happens is that Bill Gates freaks out and this is why:
Apple will license OS X when Intel makes a chipset that can’t run Windows
Imagine for a moment if the following happend: Apple release a PowerMac with an Intel chip that is so advanced it can’t run Windows. Other manufacturers want to sell the fastest chip so they include them in their machines as well. The only problem is that those machines no longer have an operating system. So, Jobs decides to license OS X to them. The net result is that Windows is completely shut out of the market. No one would want to run Windows because they would have to do it on outdated technology. Now the only hope for Microsoft to stay in the game is to put out an new OS that can take advantage of Intel’s best chips. But like I said before I think that with Longhorn/Vista we can be well assured that they will crash and burn long before creating an up to date OS.
I think that this is Apple’s long term plan, to finally cut MS off at the knees and try to reclaim the OS market. And do you know what, I think they can do it too.
del.icio.us
Digg
Facebook
Slashdot
StumbleUpon
Comments
It’s a brief analysis at the end there, but I think you may have nailed it…
My gut feel was that openning the floodgates to tier one PC manufacturer’s was somehow Apple’s plan in the future, and this makes sense if the performance delta between legacy x86 Intel chips and the new technology chips is vast.
Additionally, a Virtual PC like app or process, could cover the concern over running legacy Windows on these machines if the performace is sufficient.
Never a dull moment on planet Apple.
lb
Hmmm. Never thought about that. Thatd really cook Bill’s goose.
Apple is setting up a research group in the same building as Intel’s at Carnegie Mellon University. It’s the alma matter of Avie Tevanian, Apple’s head of software development and one of the designers of the Mach kernel for his PHD. Might this development give some added wieght to your thought?
Of course, you then have to ask the question would Intel be willing to develop a ‘custom’ chip specifically for a customer with such a small share of their market; after all, IBM wouldn’t do it for Apple, so why would Intel?
Unless they really, really want to break up their intimate link with Microsoft…
And who’d want to do that
You act like Intel is the only choice out there. Right now, AMD is just as good if not better in some areas as Intel. Windows will also be around as long as Apple makes you buy their hardware to run their OS.
very interesting speculation! if intel did make a nextgen mac-only processor, i’d think we could assume a few things…
1) it would debut shortly before vista. what better way to steal vista’s thunder and make it look even more like a cheap OSX knock-off?
2) we know rosetta does PPC to intel emulation. what if it can do more than this? what if it can do x86 to appleintel translation too. that would mean every copy of OSX 10.5 could essentially run basic windows apps under emulation AND also run older PPC apps too. that would kill
3) intel is free to do what it likes, because what is microsoft prepared to do? port windows to another processor? if intel wants options, working with apple gives them alternatives and will force microsoft to be more cooperative.
4) at first, apple would want any perceived or actual advantages to new hardware to be exclusive to the new macintel machines….but eventually there would be interesting possibilities for licensing. imagine a macintel inspiron or vaio! blesssed by intel. powered by OSX.
5) the possibility that fairplay apple DRM gets integrated into intel’s chip-based DRM. the apple HD movie store anyone?
There is something VERY wrong with the site, ALL comments are fully linked to the adserver, clicking the commentsbox also takes me there…
Very nice speculation, gave me a good chuckle. MS could still go back to creating software only & quit the OS-market. They were better back then when they did I think.
I think the point of Apple going Intel is twofold. First just to run on hardware that CTOs are familar with, gaining credibility by this fact alone (silly but true). But most important resonates with David’s thesis. Apple plans to show that moving OS X from platform to platform can be done with virtually no problems at all, establishing OS X as the clear development path into the future regardless of advances in processor technology. Intel wants to make (and sell) more advanced processors that abandon the x86 architecture. OS X can not only serve as the OS in these cases, but people will have faith that Apple can do this in way that doesn’t create much headache. If AMD jumps into the frey, all the better. Windows Vista becomes an interesting historical note. Microsoft comes out with Windows Lux (Linux Edition) to save some share of the market.
Start with a dash of questionable premise, add a heaping helping of wild speculation, then baste with a fantastical conclusion worthy of Tolkien, and you’ve got a steaming dish of Mac-fan article on your hands. Not very good for you but boy is it tasty.
Beeblebrox,
When ever you post a response to one of my articles all I can think about is Jonny Depp in Pirates of the Carribean:
Norrington: No additional shots nor powder. A compass that doesn’t point north. [unsheathes sword] And I half expected it to be made of wood. You are, without doubt, the worst pirate I’ve ever heard of.
Jack : Aye, but you have heard of me.
Any day now I expect us to have that conversation.
Beeblebrox : “You are, without doubt, the worst writer I’ve ever heard of.”
James : “Aye but you’ve heard of me.”
Anyway, as always thanks for the comment.
Microsoft will be shut out.
If the above comes to pass (man, what a day that would be!), Microsoft will not have *that* good of a friend in AMD, its next obvious choice.
Sun is in bed with AMD right now, and Sun is giving away its Solaris operating system optimized for AMD chips now. Also, AMD is a big name amongst the Linux weenies, no friends of Microsoft, either.
In a twist of irony, Microsoft stops selling operating system software (they will still make Office for Mac) and the XBox-360 becomes their compute platform, as well as game machine, toaster, dishwasher, whatever. And that is running IBM as we all know.
Interesting times we live in.
What makes you think that Intel is capable of designing a chip like that? Intel’s strength is process, not design, never design.
I mean, look at their track record:
4004 - this was a pocket calculator chip, complete with BCD registers and different word sizes for read-write and executable memory.
8008/8080 - Highly irregular design, slower than the competing 6502 and 6800 clock-for-clock, and beaten out by a clone… the Z80.
8086 - slow even when it came out, exotic “segment” design that made it easier to port CP/M software to it, which made it a marketing success but a technical disaster we’re STILL paying for.
iApx432 - CISCiest American chip ever, only beaten out by the Japanese TRON effort for the “missing the point” award. Total disaster.
80286 - Extended the address space of the 8086 with segment mapping, but that made segments even slower to use because you had to take a trip through the MMU every time you loaded a segment register, and they put the segment type bits at the wrong end of the segment word, so you couldn’t just lay down a bunch of segments every 64K and treat them like a big flat array.
80386 - Slower than the contemporary 680x0 models when it came out, had to switch to a subtly incompatible ISA to use the 32-bit addressing, and still short on registers.
i860 - About 10 years earlier than compiler technology could handle it, died.
i960 - Best processor, for the time and available technology, that Intel ever designed, so they promptly caponised it by removing the MMU and sold it to the embedded controls business who are still using it.
80486 - Intel builds a balls-to-the-wall processor they don’t have glitz up with cool features to make their marketeers happy, and wraps a translator for the 80x86 instruction set around this “RISC Core”.
80586/Pentium - It worked for the 80486, so they did it again, better. So does AMD.
IA64 - Itanium is a bomb, HP pulls them out of the water with Itanium 2… maybe too late.
XScale - Take DEC’s StrongARM, increase the clock speed, and end up with a processor that’s slower at 300 MHz than the StrongARM was at 206 MHz. On the upside, they’ve now got it up over 600 MHz, which is pretty good for an embedded controller like that…
About the only thing I could see that would let Intel do this would be to do what they did with the XScale… for them to get together with HP and resuscitate the Alpha. Call it “IAXP”, “IA” for “Intel Architecture”, “AXP” for “it sounds cool” (“AXP” was an early marketing name for the Alpha, it didn’t mean anything), and tell us it’s “Intel Architecture eXPress”. Give it an x86-64 to Alpha transcoder (from what I’ve heard, the P4 transcoder should be close enough), and a way for programs to switch to “IAXP mode” and “run directly on the RISC core”.
But they’d be more likely to put Itanium cores in there, and if you thing the G5 runs hot you ain’t seen nothin’ yet…
Although the prospect you lay out is most heartwarming for any serious Mac-zealot like me, there are IMHO several things wrong with your reasoning, the most serious error being the idea that Intel might actually develop a custom chip exclusevly for Apple.
If I remember correctly, Apple got a (more or less) competitive price for the relatively few PPC-chips they ordered from IBM and Moto because they actually owned part of the design, back from the old AIM-days. This price-advantadge doesn’t exists in their new relationship with Intel and will have to be compensated by the fact that they can order already designed off-the-shelf components at volume discounts (by agreeing to go exclusively with Intel-chips). Now, even IF Intel was willing to do it, how much would Apple have to pay for a custom-designed chip that (for a foreseeable time) NOBODY ELSE would use? Probably enough to send the next G6 PowerMac back into the $5000 price segment. And that’s where Aplle doesn’t want to be these days, right?
Facts:
- before WWDC, Steve J bragged about being contacted by three major PC manufacturers, begging Apple to port Mac OS X to PC hardware
- at WWDC, Intel transition was revealed, MS was snubbed for being slow on SW releases
- after WWDC, Michael Dell said he would be willing to sell MacIntels, given the chance
Probably true:
- Steve J still lusts for world domination on the OS market, even though he has publically said the the OS wars are over
- Steve J also understands that Apple cannot make all those computers needed for world domination
Speculation:
- Apple keeps the iBook/Powerbook, iMac and multiprocessing PowerMac lines, in order to keep a healthy revenue needed for further Mac OS X development
- Apple licenses single-processor (even with multi-core) PowerMacs to Dell, HP to gain CIO mindshare and Average Joe acceptance (2007)
- Apple’s market share goes double-digit (2008)
- Apple ends hardware manufacturing when market share > 25 % (2010)
- Windows market share < 10 % (2015)