Submitted by Why didn’t Jobs, ever the maverick, opt for the scrappy challenger, Advanced Micro Devices, instead of the old-money establishment, Intel? The reason, industry analysts say, is that Jobs has a clear goal in mind: innovative designs. And such designs require the lowest-voltage chips, which IBM and Freescale were not going to make with the PowerPC chip core – and which AMD has not yet perfected.
Intel quoted Apple a better price than AMD, or just kicked something back to Jobs himself.
I doubt it. Apple sells more laptops than desktops. And Apple has a great profit margin with IBM and Motorola chips; so their profits on Intel chips should only get larger – Even with bad prices.
Apple needs a low power 64 bit chip that doesn’t get shown up by Intel’s offerings. IBM is trying for a low power G5, and TMK they sort of did it (with a nice clock down). But Intel is already shipping awesome 32 bit chips with plans for 64 bit ones.
Apple also needs chips for iPods and successors. And with the way iPod has gone so far, I think Apple will continue to ask for more power to put in more goodies: Video and the like.
AMD is just a high speed processor company. Their mobile stuff is a joke. Unless they’ve put out a non-toaster in the last year that I didn’t hear about?
Mark my words…the transition to Intels processors will be the biggest mistake in the history of the apple.
I mark them.
So when Apple goes under I’ll say: “Hey some anonymous said it!”
It’s based off of the PowerPC core, and is most certainly a lower voltage processor. As far as Intel goes, they certainly have plans for lower powered processors, but aside from the Pentium M, they don’t have a lot going for them at the moment. Sure, AMD hasn’t perfected low voltage yet, but Intel is just getting started. This sounds to me like more speculation. I don’t think we’ll know the real reason, until at least Apple releases a retail system based off of the Intel platfrom.
How many times must it be pointed out that Cell is mediocre as a general-purpose processor?
AMD has made a name for itself with super-fast machines, especially popular with gamers and bargain hunters, who value the couple hundred dollars you can often save by buying AMD-based PCs instead of Intel-powered ones.
A bit short through the corner.I’m afraid the author forgets all the dual-opterons and the dual-cores.So AMD isn’t popular in the server room?It’s not only the prize,but allso the superiour performance while consuming less watts.
My take,either AMD couldn’t agree on Apples childish demands or Intel paid to be sure Apple would go with them.
People, it comes down to this: Apple liked Intel’s roadmap more than they liked AMD’s. Yonah is perfect for the iBook and Mac Mini, Merom is perfect for the PowerBook, Merom/Conroe could appear in the iMac, and Conroe/Woodcrest in the PowerMac.
Meanwhile, AMD isn’t introducing any new designs next year: they’ve already said that they think their technology is so great that it can beat the designs Intel is debuting next year. We’ll see, but it wasn’t a mistake to go Intel, no matter how much hate some people around here have for the company.
Yeah, the new socket is not new technology for them (irony). Also the ability to run multiple OSes, which AFAIK they demonstrated first (sorry intel & ibm).
The fact is that their technology is superior, but I think it’s better to AMD that Apple bugs intel, not amd.
IBM’s updated roadmap shows that it can indeed keep up with Intel in terms of power.
The big issue with Apple is that they didn’t want to do their part to make it a *mutually* profitable business.
It’s the volume stupid.
Only Intel can provide the amount of chips so Apple can REALLY move a lot of product, AMD or IBM would take tens of years to ramp up even close to Intel if Apple could even provide enough orders.
Apple doesn’t have that kind of time, SJ isn’t getting younger either, it has to move fast to keep abreast of M$ endless imitation.
Steve has to everything he wished he could have done all his life in short order.
Apple when it was under SJ control the first time, they went down the PPC route, it was a better chip.
But things changed when SJ was away, the crap x86 chips took over and dominated.
Fast forward to today what does SJ have to work with?
A world of Win developers on x86 based chips.
So what can he do BUT go with the flow and control the direction of the computing industry instead of trying to divert it into a whole new direction?
Slide his OS right under the noses of x86 Win developers and make it easy to compile their code for Mac OS X. Smart move.
Look what already has happened in the short time he’s been back, he’s already assumed control and M$ is busy catching up and imitating.
Heck back a year ago Intel and the Win world said “What? 64 bit Dual processor workstations for consumers? Bahh they don’t need it”
Now look, Intel is going dual core 64 bit across their entire offerings.
Look at the early versions of Longhorn and the new Vista changes.
M$ is cracking the whip trying not to be left behind and look obsolete.
So SJ has his gameplan to make the computing world a better experience, it’s spilling over into the Win only world like it did from very on.
Computers are fun and exciting again, more power to him.
I look forward to the Mactels that will run most all Win and Linux based programs on Mac OS X along with it’s own.
How can one resist not buying a Mac like that?
Computers are fun and exciting again, more power to him.
WAHAHHAHA… sorry.. but they haven’t been “fun and exciting” in awhile.. I admit, they are headed that way again. but not there yet.
They ARE fun and exciting. Just install a C64 emulator and go fullscreen!
AMD runs about 12% of the chip market. Apple runs about 5% of the PC desktop market. AMD ships more units than Apple’s seen this decade, in a year.
AMD runs about 12% of the chip market. Apple runs about 5% of the PC desktop market. AMD ships more units than Apple’s seen this decade, in a year.
So, in other words, if Apple started using AMD chips, AMD would have to ramp up manufacturing from 12% of the market to 17%. That is a huge jump and probably way over their capacity. Sure, Apple could wait for them to ramp it up… or they could just go for Intel, who can supply the extra chips without even blinking.
I don’t think moving x86 to PPC is in any way the most difficult part of porting from Windows to OS X. The only part you have to do much with is going to be your asm, and your badly written c.
“Only Intel can provide the amount of chips so Apple can REALLY move a lot of product, AMD or IBM would take tens of years to ramp up even close to Intel if Apple could even provide enough orders.”
It’s certainly one big advantage of Intel versus anyone.
I agree but I think it’s not the only advantage:
– Intel has a bigger budget to invest in marketing to convince people that Intel products are the top, and Apple need the continuity with their costumers in saying: “I give you the best products”;
– Intel has a wide acceptance in businness customers, a sector that Apple has exploited only in some niches (creatives, graphic etc) and to *try to* enter in the rich market of office servers and clients Intel is a better name than AMD to use as a matra with old-minded corporate buyers;
– Intel has big money to guarantee research and developement and to guarantee continuity, it recently survived to the Netburst “Waterloo” and to Rambus “Caporetto” deploying totally alternative lines of products so Apple knows that can rely on a robust parthner that can recover from the worst and silliest errors having the money to invest in n alternative lines of r&d, that AMD would not probably be capable to do;
– apart from the advantages due to Intel size as biggest semiconductor seller of the planet, Intel has the bonus value of good alternative techs to x86: Xscale is growing as an interesting platform, Itanium has reached maturity, other specifical chip designs are I.P. of Intel; having good contracts for Xscale, since is yet very interesting for Apple as a more powerful processor for audio and video players, this alone can be a decisive point in favour of Intel!
– since Intel produces a wide variety of chips, it is capable to produce large volume of complete chipsets, from cpu to mobo with all it’s coprocessors and integrates, that means that is more rational to sign a contract with Intel to have a complete solution rather than with AMD+”someone else that produces high volumes of mobo for AMD”, expecially if you want to place some protection chip or some customization to make your solution not exploitable to build clones.
IHMO
AMD does not have an embedded solution.
Intel has XScale. By signing up with Intel, Apple gets cheap XScales. This means faster, cheaper, better iPods, iPhones, iPDAs and other assorted iCrap. And THAT, my friends, almost more than in computers, is where the big fat profit margin is for Apple these days. Why is nobody mentioning this?
i thought AMD had a mips chip line? alchemy?
The idea of using the XScale in an iPod makes me cringe. While the XScale most certainly consumes less power than any of Intel’s other processors, it’s most certainly not along the low power needs to keep the battery as long lasting as it is now. There is also no indication that Apple even plans on extending to the phone and/or PDA market. Also, AMD does have an embedded processor. It’s called the Geode, and I wouldn’t want to see it powering iPods either. AMD also makes flash memory. I don’t believe Intel does.
Laptops, AMD does not have anything to compete with the Centrino platform.
http://en.wikipedia.org/wiki/Turion
25W for Turion at 2.2GHz vs 27W for Pentium at 2.13GHz. Admittably, Turion doesn’t have anything in the 10W range like Intel’s LV and ULV chips, but I don’t see Apple using ULV chips anyhow. Considering that Turions are already 64-bit, as opposed to the 32-bit Dothans, performance is also a non-issue. Yonah is what it will come down to. Can a dual-core laptop chip work? Will dual 32-bit cores outperform a single 64-bit core? Of course, by then AMD will have introduced its Socket S Turions, and they’ll probably offer something new, too…
So, in conclusion, I doubt “Centrino”/Pentium M was the driving force behind the decision to go to Intel. I think their production capacity, larger product base (Xscale, desktop chips, lots of laptop chips) and secrets known only to Steve Jobs and the Intel execs were probably much more central to the issue
Yes it has: the Turion platform.
I believe that it comes down a single point of contention between Apple and IBM:
IBM asked Apple to pay for the R&D costs of future processor development.
This hinges on two points:
IBM dont really sell the 970 to anyone else.
IBM havent been able to reliably deliver on their promises.
Im sure that if IBM had stepped up and met or exceeded their performance commitments then Apple would have footed the bill for R&D. But there is no way Apple would put 100 million ( conjecture ) down on CPU development, when, at the end, IBM fails to fab in volume or on time.
In light of this Apple’s choice to change supplier is quite reasonable ( much as they did with Motorola ). However, this time there are no PPC manufacturers left. Look at the market place, no one else can run with AMD and Intel.
Choosing between the two is also pretty easy. Intel make huge profits, and AMD dont. Who would you partner with for the next 10 years?
Yeah, Intel’s road map. Thats proved to be really stable. They had no plans for 64 bit x86, then they changed their minds after seeing the Opterons profitability. They were relying on ever increasing processor speeds, then they kicked that to the curb in favor of multicore designs. While these are all sound decisions, it just goes to show that roadmaps are somewhat changing, and to arbitrarily decide to make a major architectural design change off of a processor company’s roadmap would be somewhat foolish.
Bear in mind that to architect a new core and bring it to market takes 4-5 years, so they must have had contigency plans for AMD64 if it took off – otherwise EMT64 chips would still be in development. I suspect there were business / political reasons for not publishing that bit of their roadmap – they were trying to promote Itanium at the same time.
The EM64T chips aren’t new cores, though. They’re 32-bit P4 cores with enhanced decoders and AGUs. There are no actual 64-bit integer pipelines in a Prescott.
> The EM64T chips aren’t new cores, though.
I guess they’re not actually 🙂 I shouldn’t have said that – I think I’ll modify my original statement; it’s not the core development time that showed they weren’t caught totally without plans, it’s the fact they’d already got license agreements with AMD.
Getting these things done admittedly wouldn’t take so long but it does seem to me that Intel had at least thought about 64-bit x86 and were waiting to see how the market took it. That said, it doesn’t sound like they made a particularly good job of it inititally…
> They’re 32-bit P4 cores with enhanced decoders and
> AGUs. There are no actual 64-bit integer pipelines
> in a Prescott.
Is that so? I hadn’t realised that – kind of glad I didn’t buy one! Do you have a link where I could read more? There are usually some whitepapers hanging about somewhere difficult to find 😉
Getting access to Intel’s highly optimized compiler technology and a bit of hand holding to enter the X86 market are good motivators to start with Intel. I think the real key thing that brought about the architecture change (that Jobs wont want to trumpet too loudly), is the DRM strategy that Intel brings them, how it will allow them to appease RIAA and MPA and protect their own OS.
It’s all about the little TPM chip. (sadly)
As an added bonus Intel could provide them with ARM processors to power future iPods if they so desire.
As much as I like AMD chips on the desktop, the 25 watt Turions (which are extraordinarily rare) tend to use more power than the comparable PM’s. That 25 watt figure is actually a max figure, not an average, and Intel’s average wattage tends to be lower than AMD’s. Also, AMD is still dependent on other system chipsets from nvidia et al that aren’t as power efficient as Intel’s system chipsets.
I see it as having a stable chip maker more than anything else. Both IBM and Motorola have screwed Apple in the past both in being able to deliver sufficient quantities, and getting the chips fast enough. If they went with AMD, they may get faster chips, but they run the risk of not getting enough of them. AMD doesn’t have near the capacity of Intel for sheer quantity. With Intel, the CPU is ‘good enough’, and they get enough of them. Apple can concentrate on the quality of the software and not worry so much about MHz myth, availability, etc.
They both are x86-64 now, and should be compatible, so if later down the line they want to use intel they can, can’t they ??
sorry i mean they can use AMD….
That Intel offers overall better performance per watt? You must be kidding me, no?!
That Intel is more innovative? You possibly haven’t heard of x64!
What about dual-cores? Who is leading this field?!
I think the point Jobs was making was the future (having presumably had rather extensive access to the roadmaps of the companies he considered) rather than the current state of play. I think Intel were first to *shipping* a dual core, although the AMD solution looks like it might be better thought-out overall. (and they were both beat to multicore by about 5 years by IBM’s POWER)
AMD had dual-core Opterons shipping LONG before Intel got their cobble-together chips out. Intel only managed to beat the dual-core Athlon64 meant for desktops to the market… and paid news agencies a pretty penny to scream that news at the top of their voices to make it look like Intel won the race.
Intel had dual core desktop chips shipped to OEMs before the dual core Opteron started shipping (though only just) [ http://www.serverwatch.com/news/article.php/3497241 ]. I remember it being a rather tight-run thing with the distinct look of Intel bringing their schedule forward, trying to get a publicity victory
AMD do great work to compete with Intel on technical innovation, performance and price but they still have limited resources.
Despite their *technical* strength, there’s the *logistical* problem of shipping enough CPUs. Do they have the fab capacity to reliably supply a vendor the size of Apple? (who, regardless of being a small share of the desktop market do ship more systems than most individual PC vendors)
To all you “AMD sells more machines than Apple” commentors
Apple obviously fully intends to sell more machines than AMD can possibly supply chips for and doesn’t want to be hobbled later on down the road.
Look at all the Apple Stores going up, they fully intend to move volume.
Doesn’t mean they will never use AMD chips, it’s just the game plan and the time restraints call for going with Intel for now.
If SJ was a young man I can see him making that decision to go with AMD because he would have the time, but time is not on his side anymore.
We can thank the scumbag John Scully for crippling computer innovation for the last 2O years.
Bill Gates is just a opportunist mongul control freak and doesn’t have a creative bone in his body.
The question isn’t why Apple went with Intel. That one’s easy: Intel made them a very good deal.
Ask yourself why Intel went with Apple. My guess is that Intel needs a technology showcase partner, and it’s worth giving away CPUs to get Apple.
Intel’s main customers resist innovation. Most sell commodity systems at the lowest possible price, and are clueless when it comes to new ideas. The history is one of same-only-faster, and that won’t create new markets. Look at how long it has taken to shed parallel ports and other legacy features. Look at the rather dull offerings in the tablet and media center market.
Intel needs a partner that can do decent industrial design, has vision, and isn’t held back by Microsoft’s inability to innovate. I can certainly picture Apple working closely with Intel to develop new products outside of the traditional computer market. That’s where Intel’s growth will be, and it won’t be provided by Dell or HP.
Ask yourself why Intel went with Apple
Exactly. And even more interesting would be why IBM or AMD didn’t.
Everybody is saying “Apple could’ve chosen AMD but didn’t”. Well, maybe AMD and IBM didn’t want/need to give Apple what they asked. Apple is not an easy client.
Intel has:
– Centrino/Pentium-M platform;
– Extended Firmware Interface;
– Xscale embedded platform.
Today Apple needs Freescale, IBM, PortalPlayer, Lucent, Broadcom, OpenFirmware, ecc. It’s “global economy”: one partner is ALWAYS BETTER than six (or more).
Today Apple needs Freescale, IBM, PortalPlayer, Lucent, Broadcom, OpenFirmware, ecc. It’s “global economy”: one partner is ALWAYS BETTER than six (or more).
Hah! That must be one of the worst ideas I’ve ever seen in the comment’s section. Relying on one partner to supply your entire product line is *stupid* and I don’t think Apple will do that.
It is stupid because if there is a problem inside Intel, Apple’s entire product line would feel it. If Intel makes a gross mistake, Apple will feel it everywhere. You need mutliple suppliers, it’s called spreading of risks.
Relying on only one supplier will come back to haunt you. I’m sure Apple isn’t that stupid. They’ll in the future look at AMDs offerings too, they have no choice. Or, VIA of course
.
“It is stupid because if there is a problem inside Intel, Apple’s entire product line would feel it. If Intel makes a gross mistake, Apple will feel it everywhere. You need mutliple suppliers, it’s called spreading of risks.”
I don’t agree.
1) It’s clever because you have more negotiating power with your parthner (i.e. “If you make good prices and good clauses for XScale for my ipod i’ll use your x86 cpus for my macs”);
2) Intel is a solid parthner: it doesn’t failed after big big mistakes like Netburst choice or Rambus affair; it does big money selling semiconductors and different lines of products with different targhets, so it’s really improbable a “general failure”;
3) if one out of the n little contractors fails to supply a component, you will fail to supply the entire products relying on it you are not spreading the risk but multiplying the risk by n: any failure of a parthner will be critical since you cannot deliver a product without any single chip or part and you cannot contract another partner in “no time” to start building the missing part! So a single solid parthner is better if you have a good contract that allows you to keep intellectual properties on co-works in order to give it to another parthner in case of (improbable) Intel’s failure; otherwise you will need n*2 parthners in order to have the possibility to replace on the fly one failing little parthner… with all problem of i.p. protection, fluctuations in the production that may cause non uniform supplies to the Apple’s production chain.
Apple has to face a feasible and fast migration if Intel fails (I have i.p. on co-works and I ask for new partnerships in change of those wanted specs, maybe without stopping the production a single day since I may sign contract with many little companies that were working for intel about my products) and if Intel doesn’t fail Apple will have a solid and easy to manage relationship (I’ll have uniform supplies, in time and volumes I need).
With many little parthners Apple will have troubles every time a single parthner fails or produce to less components, or retard the production, a single incident will impact on the whole product line.
Single source–all they way!
Thye hate choice.
Ok, I fail to see what lower power has to do with innovation. I can certainly see the interest in using it in smaller devices being a sensible choice…
But what the hell does innovation have to do with it?
while all your comments/opinions are valid, you forgot one.
apple is american company and so is intel and AMD is german so what do you think? what goes with what
please don’t remove this post though i can see that happening very soon. but hey its one important fact that everyone seems to go around of.